University of Groningen

Changing face-to-face communicationDiggelen, Wouter van

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2011

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Diggelen, W. V. (2011). Changing face-to-face communication: collaborative tools to support small-groupdiscussions in the classroom. Groningen: University of Groningen, SOM research school.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 31-01-2020

Changing Face-to-Face Communication:Collaborative Tools to Support Small-group

Discussions in the Classroom

Wouter van Diggelen

Publisher: University of Groningen

Groningen, The Netherlands

Printed by: Ipskamp Drukkers B.V.

Enschede, The Netherlands

ISBN: 978-90-367-4820-9 (Book)

978-90-367-4821-6 (Electronic version)

© Copyright 2011 by Wouter van Diggelen

All rights reserved. No part of this publication may be reproduced, storedin a retrieval system of any nature, or transmitted in any form or by anymeans, electronic, mechanical, now known or hereafter invented, includingphotocopying or recording, without prior written permission of the author.

RIJKSUNIVERSITEIT GRONINGEN

Changing Face-to-Face Communication:Collaborative Tools to Support Small-group

Discussions in the Classroom

Proefschrift

ter verkrijging van het doctoraat in deEconomie en Bedrijfskunde

aan de Rijksuniversiteit Groningenop gezag van de

Rector Magnificus, dr. E. Sterken,in het openbaar te verdedigen op

donderdag 31 maart 2011om 14.45 uur

door

Wouter van Diggelen

geboren op 4 juni 1964te Wijk en Aalburg

Promotor: Prof. dr. H.G. Sol

Beoordelingscommissie: Prof. dr. A. Boonstra

Prof. dr. G. Kanselaar

Prof. dr. ir. A. Verbraeck

For Agnes, Karlijn,

and David

Table of Contents

Preface......................................................................................................................vii

1 Introduction.........................................................................................................11.1 Communication and Learning in the Classroom................................................11.2 Group Discussions in the Classroom..................................................................31.3 Research Problem................................................................................................41.4 Research Objective..............................................................................................51.5 Questions that Guide the Research...................................................................101.6 Background: The DUNES and the LEAD project...........................................131.7 Outline.............................................................................................................15

2 Research Approach: Understanding through Design...............................192.1 Action Research................................................................................................202.2 Design Science..................................................................................................232.3 Design-based Research......................................................................................272.4 The Practice of Design......................................................................................302.5 Way of Working................................................................................................32

3 Social Reality Explained: Systems and Functions.....................................393.1 Way of Thinking...............................................................................................403.2 Systems Thinking.............................................................................................423.3 Collaborative Learning: Three Levels of Complexity........................................453.4 Understanding Groups: A Functional Perspective.............................................493.5 Summary..........................................................................................................51

4 Group Discussions: Learning as Co-Construction...................................534.1 Constructivist Views towards Learning.............................................................544.2 Collaborative Learning in the Classroom..........................................................584.3 A Constructive Dialogue..................................................................................62

5 The Learning Environment of the Classroom..........................................655.1 Two Learning Methods.....................................................................................655.2 The Classroom Learning Environment.............................................................685.3 The Teacher......................................................................................................695.4 Media................................................................................................................785.5 Other Learners..................................................................................................805.6 Summary..........................................................................................................82

6 Changing Communication: Designing Tools for Collaborative Learning...................................................................................83

6.1 Ineffective Communication Patterns................................................................846.2 The Envisioned Direction of Change...............................................................876.3 Design Guidelines.............................................................................................906.4 Functional Design..........................................................................................1036.5 CoFFEE: The Software System.......................................................................1086.6 Summary........................................................................................................1136.7 Research..........................................................................................................115

7 The Organization of Computer-mediated Communication................1177.1 Problem Analysis: Parallel Access as Floor-control Mechanism.......................1177.2 Study I: The Organization of Computer-mediated Communication.............1197.3 Study II: A Refinement of the Design............................................................1267.4 Conclusion.....................................................................................................131

8 Computer-Mediated Communication and Task Performance............1338.1 Computer Support for Task-related Communication.....................................1348.2 The Learning Environment............................................................................1368.3 Analyzis I: The Content of the Computer-mediated Interactions...................1398.4 Analysis II: The sequence of Task-related Interactions....................................144

9 A Diagnostic Reasoning Discussion with the Support of CoFFEE...1539.1 Diagnostic Reasoning in the Classroom.........................................................1549.2 Method...........................................................................................................1569.3 Instructional Design.......................................................................................1659.4 Analysis II: The Minimal Requirements for a Group Discussion....................1729.5 Analysis III: Group Communication..............................................................176

10 Conclusions.....................................................................................................19110.1 Mediated Face-to-Face Communication.......................................................19210.2 The Design-based Research Approach..........................................................199

References................................................................................................................ 203Appendix A A Scenario for the Threaded-discussion tool....................................... 219Appendix B A Functional Description of the Graphical Tool..................................225Appendix C CoFFEE: A Brief Description and Download Information.................. 231Summary.................................................................................................................. 233Samenvatting............................................................................................................243CV............................................................................................................................253

Preface

The final words of this thesis, which are written in this preface, gave cause for reflection. When I look at the manuscript, I sometimes feel like a cognitive scientist who believes that it all comes from the individual mind. However, these moments of inward esteem are brief. Quickly, I return to the social-constructivist camp that states that knowledge and insight emerge from interactions with others. With this in mind, it is more than appropriate to acknowledge my collaboration with various people.

....... First of all, I would like to thank Henk Sol for his patience, guidance and support. This thesis would not have been finished if I did not have Henk on the background who kept me on track. I am also grateful to him for his constructive feedback. Thanks also to Albert Boonstra, Gellof Kanselaar and Alexander Verbraeck for their valuable comments during the final stage of the PhD project.

The research that I describe in this thesis has been carried out within the context of two European projects. I learned a lot from the people who participated in these projects, intellectually as well as culturally. For this I would like express my gratitude to Reuma de Groot, Baruch Schwatz, Raul Drachman, Josef Börding, Angi Voss, Ahmet Ocakli, Marije van Amelsvoort, Jerry Andriessen, Michiel Klønhammer, Noam Knoller, Annie Corbel, Gregory Dyke, Steven Collins, Jean-Jacques Girardot, Kristine Lund, Michael Baker, Grégory Six, Francois-Xavier Bernard, Steven Collins, Charles Crook, Claire O'Malley, Giulia Gelmini, Marie Buda, Delfina Malandrino, Beatrice Ligorio, Ilaria Manno, Luca Tateo, Rosario De Chiara, Giuseppina Palmieri, Raffaella Grieco, Riccardo Prinzi, Furio Belgiorno, Ugo Erra, Shaaron Ainsworth, Vittorio Scarano and Maarten Overdijk.

I would like to thank Vittorio and Shaaron for their feedback on the design documents that lay at the basis of chapter 6. I am greatly indebted to Vittorio and his fantastic team who developed CoFFEE, the award winning software whose genesis I discuss in this thesis. I also like to thank Astrid Broeker with whom I carried out the study that is described in chapter 9. Her enthusiasm and persistence as a teacher and researcher made that study to a success. I would also like to express my thanks to Dirk and Anja who gave me a place to work en provided me with necessary facilities for editing and printing.

vii

Preface

Furthermore, I would like to thank my colleagues at the department of Educational Sciences of the Utrecht University and my colleagues at the Systems engineering group of Delft University of Technology. They were the essential elements of a fine working environment.

I had the pleasure to work with Maarten, from whom I learned a lot. Our discussions were of vital importance for the work that I describe in this thesis. As Maarten already mentioned in his thesis, together we have achieved a lot. I am glad to say that it was not only the destination but also the travel that made our collaboration a wonderful experience.

A PhD project is not only a matter of the mind. I would not have come this far without the support from those who know me well. In the thesis, I typify that kind of support as social-emotional communication that has to do with recognizing the uniqueness and value of the people around you. This support can be expressed in many ways like a simple word, a smile, a pat on the back or a glance of recognition. Various people close to me gave me that kind of support. It is impossible to express in a few sentences what that meant for me. I just hope that these people – some of them were already mentioned before – will recognize themselves in the final words of this preface. To them I express my gratitude and appreciation.

Wouter van Diggelen

Rijswijk, February 2011

viii

1 Introduction

You understand, what I tell you here is just a summary; for in the state of real life things are loaded with much more artifice: one day

there is hope, the next day disconsolation.

Joâo Guimarâes Rosa1

1.1 Communication and Learning in the ClassroomThe classroom is a place for learning where students come together to be taught in a formal way. The classroom is primarily a communicative environment. Students’ participation in the communicative practices in the classroom affects their achievement (Green, 1983). The view of what constitutes effective classroom communication has changed in the final decade of the 20th century (Cazden & Beck, 2003). Traditionally, classroom communication takes place between the teacher on the one hand and the students on the other, with the teacher as the pivot of the communicative activities. The teacher has the initiative and directs the flow of communication. Lectures are a good example of a teacher-centered communication pattern with their emphasis on the transfer of pre-structured knowledge from the teacher towards the student. In these cases, the teacher controls what is learned and how. Knowledge is objectified and learning consists of transferring that knowledge from outside to within the learner (Driscoll, 1994). It reflects a view that knowledge is simply to “be listened to” (Bruner, 1996).

Constructivist approaches have challenged the objectivist view towards knowledge and learning. These approaches do not see learners as passive recipients of knowledge. Rather, learners are actively involved in constructing their own understanding of things

1 Joâo Guimarâes Rosa (1956), The Devil to Pay in the Backlands.

1

Chapter 1

(Campbell, 2002). Learners construct useful and viable knowledge when they engage in meaningful experiences during which they have to draw upon their own concepts, judgments and actions to make sense of these experiences. A number of constructivist accounts introduce a particular view of knowledge that emphasizes its social nature. These Social-constructivist accounts stress the construing aspect of communication and state that knowledge constructions take place within everyday discourse between people in interaction (Burr, 1995). Learning from this perspective can be seen as a continuing effort to improve on existing knowledge through an engagement in a discourse that advances mutual understanding (Bereiter, Scardamalia, Cassells & Hewitt, 1997).

Social-constructivist learning methods have gained their place in the classroom. These methods take the interacting student as the starting point for learning with the aim to create learning environments that stimulate meaningful interactions, i.e. meaningful interactions between the learner and relevant learning resources like the teacher, fellow students or various kinds of learning materials. In this thesis, we describe how such a learning environment may look like for students who learn together in small groups.

New Learning

Quite recently, the Dutch education system went through an education reform that aimed to introduce new forms of learning that are based on a Social-constructivist epistemology. This educational reform is known as New Learning (Simons, van der Linden & Duffy, 2000) and can be described by the following characteristics (Teurlings, Wolput, Vermeulen, 2006, Bronneman-Helmers, 2007):

– autonomous learning,

– learning in meaningful and authentic settings,

– collaborative learning, and

– a different kind of relationship between the teacher and students.

New Learning already gained its acceptance in higher education where lectures are just one way to teach students, besides methods such as project work, problem-based learning and collaborative learning (CHEPS scenarios, 2004). More recently, these alternative learning methods have also been introduced in secondary education. It

2

Introduction

seems that these methods gradually “dropped down” from higher education towards preparatory and vocational schools. These new forms of learning better prepare students for continuing education or the working practice. They provide the students with practical experience with everyday work situations where people are responsible for their own work and where they have to collaborate with others.

The new learning methods require different kinds of teaching and learning skills. The teachers is not only the expert who makes knowledge accessible for the students, the teacher is also an effective facilitator who creates meaningful learning experiences. For students the situation also changes. Students must have the skills and the motives to become learners who are largely responsible for their own learning processes. Furthermore, they should be effective collaborators who can learn with and from each other. For example, when students collaborate they must have the appropriate skills so that their communication becomes effective and conductive for learning (Sfard & Kieran, 2001). These changes require different kinds of learning environments, i.e. learning environments that fosters meaningful interactions between students.

1.2 Group Discussions in the ClassroomSocial-constructivist approaches have a different view with regard to the communication patterns that can be associated with learning. Communication does not proceed in a one-way direction from an expert teacher towards a novice learner. Instead, students are seen as an important source for knowledge. Communication between students is a valuable vehicle for learning. Active participation in the communicative practice of the peer group makes the students aware of their own thinking in relation to what others say. Students may even be better able to address their fellow students' cognitive system because their concept of things is more closely related.

Collaborative learning explicitly exploits the interactions between students as a means for learning. Collaborative learning – as it is presented in this thesis – takes its form as a problem-solving discussion that is directed towards the exploration of a particular topic or the resolution of a problem. A problem-solving discussion consists of one or more meetings between a small group of students who communicate with each other, often face-to-face, in order to achieve one or more goals such as increased understanding, the coordination of an activity, or the solution of a shared problem (Galanes & Adams, 2007). In an educational setting the goal of solving a problem may

3

Chapter 1

not come first, rather increased understanding is the main reason for introducing problem-solving discussions in the classroom. When students participate in a discussion, they develop authentic solutions for complex problems, and, while doing so, acquire useful knowledge of theories and concepts (Chernobilsky, Nagarajan and Hmelo-Silver, 2005). Through discussions, students generate and evaluate evidence to confirm or enhance their understanding (Hogan, Nastasi & Pressley, 2000).

A Constructive Dialogue

The group discussions that are the object of the research take place in the classroom where students are in physical proximity and interact directly with each other. Several small groups populate the classroom, which makes direct supervision by the teacher impracticable. As a consequence, the classroom communication shifts from mainly teacher-centered towards more diverse patterns of communication where the students often take the lead. From a Social-constructivist perspective, the communication should meet a number of requirements. Collaborative learning requires active and equal participation of all group members. However, active and equal participation is not sufficient. The communication between students must be coherent, meaningful and oriented at the construction of shared knowledge. We use the term constructive dialogue to indicate that students have to share their knowledge with the group and elaborate on what is shared. During a constructive dialogue, students collaboratively explore their mutual learning experiences. Differences in understanding give rise to further explorations with the aim to come to a shared understanding. During these collaborative explorations students communicate, defend, prove and justify their ideas (Twomey Fosnot, 1996). They explain their reasoning, listen to each other, learn from, and even argue with their peers (Cazden & Beck, 2003).

1.3 Research ProblemBringing students together in small groups does not imply that they will actually engage in a constructive dialogue that promotes learning. Communication can be counter-productive; ineffective communication may prevent students to discuss a topic thoroughly with the risk that the group does not achieve the intended learning outcomes. When we look at these groups from a functional perspective, we can state that groups do not always enter into a constructive dialogue and that their performance varies. Several studies into collaborative learning observed that not all groups were able

4

Introduction

to carry out a productive discussion (e.g. Gillies, 2004; Barron, 2003; Kneser & Ploetzner, 2001; Hogan, Nastasi & Pressley, 2000; Keefer, Zeitz & Resnick, 2000). These observations provide the rationale for the research that will be discussed in this thesis. It results in the following research problem:

groups who communicate face-to-face do not always display communication patterns that can be associated with a constructive dialogue; sometimes they perform less than expected and do not achieve their learning objectives.

The functional perspective states that the performance differences have to do with the communication patterns that can be observed during a group discussion. Collaborative learning requires particular kinds of communication patterns, which trigger learning mechanisms that lead to an increased understanding. It is however no guarantee that the expected patterns actually do occur (Dillenbourg, 1990).

Sfard and Kieran (2001) concluded that the merits of collaborative learning cannot be taken for granted due to ineffective communication patterns. These ineffective patterns inhibit the free expression of ideas and the further exploration of these ideas by the group. Interruptions are a good example of an ineffective communication pattern (Stein & Albro, 2001). Frequent interruptions by a dominant group member could hamper the process of collaboration and learning. Other group members have less opportunities to talk so that they cannot share their ideas freely with the group. It means that the group does not fully capitalize on the knowledge and skills available.

1.4 Research ObjectiveInformation- and Communication Technologies (ICT) offer new opportunities for learning by providing easy access to information and better openings for communication. Learning that is supported by ICT is usually specified by its technological origin like e-learning, online learning or web-based learning. These terminologies describe ICT as the means to bring learning resources closer to the learner. These resources can be situated everywhere and learners can access them from different places. They include general information, teaching material, instructions, teachers or a community of learners.

E-learning, for example, refers to the use of technology to deliver content and instructions to individuals to aid their learning. Various technologies can be used for e-learning, like multimedia to display content that is brought to the individual through

5

Chapter 1

the Internet or through storage devices as the compact disc. Online learning and web-based learning, in contrast, are confined to the Internet; the computer is the medium to access learning resources available on the World Wide Web.

We opt for the term networked learning because it contains a view on technology-enhanced learning that can be characterized by connectivity. The term connectivity refers to a general function of ICT that applies to a variety of learning situations. ICT can be used to promote connections: between one student and other students, between students and tutors, or between a learning community and its learning resources (Goodyear, Banks, Hodgson & McConnell, 2004). Networked learning stresses that new technologies enable learners to interact more fluently with the relevant resources in their surrounding.

Although different terminologies are used, there is a large degree of consensus about the role of the technology. Its main purpose is to overcome the time and space constraints associated with traditional learning situations like the classroom. These constraints have to do with the limited availability of information and people in relation to a particular place where learning takes place. Learning facilitated by ICT typically focuses on those situations where there is a perceived lack of face-to-face contact (Holmes & Gardner, 2006). The Internet gives the learner easy access to learning resources that are located elsewhere. There are no time zones; location and distance are no longer an issue (Anderson, 2008). Furthermore, learning is no longer confined to the classroom or the workplace; every place can be a place for learning, as long as there are computers or smart devices available. It is clear that these technologies have a direct added value because they broaden the scope of interactions so that a variety of learning resources becomes accessible that are difficult to reach otherwise. New learning environments emerge that could not exist without ICT and it seems that when the boundaries of time and space fade away, learning finally reaches its full growth.

Facilitating Face-to-Face Communication

In this thesis, we question the partial view that networked learning is all about crossing the physical and temporal boundaries of the learner. Networked learning, in our view, is not only about expanding interactions, it also applies to those situations where there is a need for better interaction. Not because more communication is needed but

6

Introduction

because existing communication is not effective and may even hamper learning. From this perspective, networked learning suits a wide variety of situations – face-to-face and distance – as long as it improves connectivity. A networked-learning environment may have virtual properties, as is the case of online learning, but it could also be a real environment where people, tools and materials exist only in a particular location in time and space.

Computer Support for Collaborative Learning

Van Diggelen and Overdijk (2007) make a distinction between three types of computer support for collaborative learning, or more specific, for supporting group discussions (Figure 1.1). These situations take the face-to-face discussion as the point of reference.

The first situation (upper right corner of Figure 1.1) was already mentioned in the previous paragraph. It refers to collaborative learning situations where the computer connects students who are dispersed in time and/or space. The majority of networked-learning environments focuses on this kind of support where all the communication is mediated by a tool. For many researchers, this represents the archetypal networked-learning environment.

7

Figure 1.1: Three situations of computer support for group discussion.

Chapter 1

Online learning usually reflects an attitude of “more support is better”. A rich information flow between the students is seen as a guarantee for collaboration and learning on a distance. Research into online learning indicates that it is extremely difficult to facilitate the full range of group interactions by collaborative technologies. Computer-mediated interactions are often restricted to those interactions that mirror the cognitive processes in a group (Kreijns, Kirschner and Jochems, 2003). An enrichment of the information flow may improve online collaborative learning: for example, students may use multiple tools simultaneously to enrich their communication, or they may use an awareness tool that gives detailed information about their performance. The aim of these tools is to broaden the range of interactions or to make the context of communication more meaningful. They try to reproduce the richness of face-to-face communication. Still, it remains unclear if online collaboration can and should mirror its face-to-face counterpart. Research into computer supported collaborative work seems to indicate otherwise (Olson & Olson, 2000; Kiesler & Cummings, 2002).

Computer Support for Face-to-face Collaboration

The other two situations of Figure 1.1 (situation 2 and 3) have a fundamentally different orientation. They take the existing face-to-face context as the starting point and support learning activities of students who are co-located. The two situations have two distinctive features: 1) the students are in the same room in close proximity, and 2) the students communicate face-to-face.

Computer support for face-to-face settings has already been applied, not so much to mediate communication, but more as an information source that could be used during a discussion (Stahl, Koschmann & Suthers, 2006). Computer support in the classroom mainly concerns: 1) “single-display groupware” (Stewart, Bederson & Druin, 1998) where students collaborate with the support of a single, shared computer screen, 2) communicative tools like hand-held devices that support the interaction between the tutor and the students (e.g. Ratto, Shapiro, Minh Truong & Griswold, 2003) but that do not support collaboration between students, 3) collaborative technologies that support the manipulations of visual objects or models in a shared workspace rather than the interactions between students (e.g. Sugimoto, Hosoi & Hashizume, 2004), or 4) the use of new technologies like interactive computer desks (e.g. Cheng, Chang, Deng and Chan, 2005).

8

Introduction

A shift from online towards face-to-face collaborative learning brings along a shift in the kind of interactions that are mediated. The starting point for support differs: students can already communicate without the support of computers. This observation draws the attention to those interactions that can be facilitated by the collaborative technology and that improve learning. It seems that “less but specific support” is the leading principle.

Single-display groupware (situation 2, lower right corner of Figure 1.1) supports students by giving them feedback about certain actions that they perform with the use of the computer. It represents a situation where students are co-located and work with a stand-alone computer application. That application typically models a problem situation that the students have to investigate. Students can manipulate the model and receive feedback about their intervention by running a simulation. This feedback triggers a discussion between students about the consequence of their actions. This form of support has received some investment in terms of research (e.g. Roschelle & Teasley, 1995; Munneke, Amelsvoort & Andriessen, 2003).

The third situation (lower left corner of Figure 1.1) represents a collaborative learning situation where students communicate orally and simultaneously use the computer to communicate. Their interactions will be distributed between the two modes of communication, i.e. an oral and a computer mediated part. The combination of verbal and computer-mediated communication is largely ignored by the educational community. It marks a shift in focus from feedback on actions as a means to advance learning like in the case of single-display groupware towards facilitating certain kinds of communication patterns that stimulate group learning.

Bringing Networked-learning to the Classroom

The research that is discussed in this thesis focuses on the third situation of Figure 1.1. It brings networked learning to the classroom to support a group of students who are co-located and interact face-to-face. We describe the development of a networked-learning environment that intervenes in the existing learning situation with the aim to improve collaboration and learning. This intervention has a rather technical character: a collaborative tool is introduced in the classroom that mediates part of the communication. However, the guidelines that form the basis of the intervention refer to the dynamics of small-groups who work together on a learning task.

9

Chapter 1

The objective of the research is to develop two collaborative tools that mediate part of the communication during a small-group discussion. These tools stimulate the occurrence of certain communication patterns that make the desired learning behaviors more likely. Students use – in addition to their verbal, face-to-face interaction – the computer to communicate (Figure 1.2). Intuitively, the situation that is displayed in Figure 1.2 seems an odd setting for learning. Why should the students use the computer to communicate when they can easily talk verbally? We expect that improvements do occur because the collaborative tool addresses some problems that can be traced back to verbal communication. The collaborative tool changes the nature of communication so that the groups are better able to carry out a constructive dialogue. The tool offers a number of additional structural features that guide the group towards the attainment of their learning goals.

1.5 Questions that Guide the ResearchThe objective of the research is to develop two collaborative tools that mediate part of the group communication with the aim to improve collaboration and learning. It is expected that the collaborative tools neutralize the ineffective communication patterns that can be observed during a small-group discussion. They offer the group an alternative medium to organize their communication. They do so by means of structural features that regulate the communicative exchanges in a way that better matches with the criteria of a constructive dialogue.

10

Figure 1.2: The envisioned networked-learning environment.

Introduction

Ineffective Communication Patterns

A communication pattern can be regarded as a recurrent sequence of related communicative acts; it describes the regular flow of communication on a micro level. A pattern entails factors such as the distribution of acts among persons and over time, and the medium used (McGrath, 1984; McGrath, 1991). To improve group learning, we need to understand how the communication is patterned. The communication patterns must be compared to a standard for learning achievement. In our case, the standard is based on the criteria for a constructive dialogue. This brings us to the first research question:

Q1: What are the criteria for a constructive dialogue in terms of requisite patterns of communication?

To answer this question, we will examine research into collaborative learning that relates group communication to learning outcomes. The answer provides us with a set of criteria for understanding small-group discussions. These criteria will be formulated in “media-neutral” terms so that they can be used to study verbal as well as computer-mediated communication. This in accordance with Briggs (2004) suggestion that a theoretical model for the design of collaboration technology must be technology free.

The criteria for a constructive dialogue help us to the identify ineffective communication patterns that inhibit learning during a verbal face-to-face discussion. The second research question aims to identify these patterns:

Q2: Which communication patterns during a verbal face-to-face discussion prevent group members from carrying out a constructive dialogue?

Ineffective communication patterns refer to observable group activities. They can be considered as symptoms for problems that hamper group performance. The causes of these problems will be addressed by the following research question.

Structures that Organize the Communication

The communication patterns that can be observed on the level of the group are closely linked with individual behaviors. For example, interruptions as a dysfunctional communication pattern can be traced back to dominant behavior of a group member who wants to control the discussion. Dominant behavior is a manifestation of a person’s “dominance-submissiveness” trait (Cattell, 1965) that accounts for regularities

11

Chapter 1

in behavior. This does not mean that a person who can be characterized as dominant always displays dominant behavior during a discussion. The actual manifestation of dominant behavior depends on the processes and structures that organize individual actions into group behavior. These processes and structures are situated outside the individual on the level of the group or the larger classroom environment. For example, if a classroom culture is oriented towards competition then the atmosphere will be hostile, while students in a cooperative classroom exhibit more friendly behavior (Johnson, 1980).

The manifestation of individual behaviors depends upon the way the groups couple individual actions into a sequence of collective behavior. Groups apply various structures to organize individual actions into a coherent and meaningful whole. These structures refer to the underlying patterns of stable relationships among group members and include components like roles, group norms or authority (Forsyth, 1983). In this thesis, we focus on structures that can be associated with a particular medium for communication. These features of a medium prescribe the content and sequence of communication (Cushman & Kovacic, 1994). They determine how the group members express their ideas and how they organize these individual communicative acts into a coherent and meaningful whole. Turn taking is a good example of a structural feature that can be associated with a specific medium. Turn taking is a type of sequential organization because it concerns the relative ordering of speakers (Schegloff, 2007). Interruptions, for example, are a deliberate violation of the turn-taking principle that states that only one participant talks at the same time. Interruptions that disturb the ongoing discussion are possible because verbal exchanges are organized according to the principle of turn-taking (Sacks, Schegloff & Jefferson, 1974). This example brings us to the third research question:

Q3: How do the structural features of the medium relate to the ineffective communication patterns?

Linking Computer-mediated Communication to Performance

In this thesis, we make a distinction between two kinds of media: 1) a human medium for verbal face-to-face communication, and 2) a digital medium to support computer-mediated communication. The former can be associated with the existing learning situation while the latter refers to the envisioned learning situation.

12

Introduction

The collaborative tools provide the groups with an alternative medium that has different structural features to organize the communicative exchanges. It is expected that these alternative structures change the communication and improve the collaboration and learning of groups. The structures of the digital medium stimulate the occurrence of certain communication patterns that better match with the criteria for a constructive dialogue. The challenge for the research is to identify these structures and to implement them in the collaborative tool. It results in the fourth research question that lies at the basis of the design of the two collaborative tools:

Q4: How do the envisioned collaborative tools change the group communication for the better?

Briggs (2004) stated that theory-driven design of collaborative tools should be guided by two types of research questions: a scientific and an engineering one. The scientific questions address issues of group performance. The first three research questions (Q1, Q2, and Q3) fall into this category. The engineering research questions follow from the scientific research questions. Characteristic for these questions is their reference to the technology. Their aim of these questions is to find out how the use of a technology can improve group performance. The fourth research question (Q4) belongs to this category.

1.6 Background: The DUNES and the LEAD projectThe research activities that are discussed in this thesis have been carried out within the context of two European funded projects: the DUNES project2 and the LEAD project3. The DUNES project focused on argumentative discussions. The project delivered a software system – the Digalo – that enables groups to represent their arguments into a diagram. Initially, the DUNES project was about online learning. We introduced the software in the classroom to support small groups of students who met face-to-face. This shift in orientation laid the foundation for the LEAD project. The LEAD project explicitly focused on face-to-face learning situations where students are co-located and communicate directly with each other.

2 The DUNES project (IST-2001-34153) was funded by the European Commission under the IST Programme of Framework V.

3 The LEAD project (IST-2005-028027) was funded by the European Commission under the IST Programme of Framework VI.

13

Chapter 1

The starting point for the LEAD project was the observation that tomorrow’s learning will still take place in schools where learners and teachers meet face-to-face (CHEPS scenarios, 2004). These face-to-face learning situations are largely ignored by much of the current research into computer-supported collaborative learning. The LEAD project focused on those collaborative learning situation where a group sits together to discuss a topic orally. The LEAD project assumed that these collaborative situations could be improved with the appropriate collaborative technologies (van Diggelen & Scarano, 2006; van Diggelen & Overdijk, 2006). The project delivered two software applications:

1. CoFFEE(Cooperative Face-to-Face Educational Environment), and

2. Tatiana (Trace Analysis Tool for Interactions Analysis).

CoFFEE

One of the objectives of the LEAD project was to develop and evaluate a networked-learning environment to support problem-solving discussions in the classroom. CoFFEE – the networked-learning environment that was delivered by the project – consists of various tools to support a range of collaborative learning activities. CoFFEE makes a distinction between four types of tools (van Diggelen, 2008):

1. collaborative tools to support a particular collaborative learning activity during a group discussion,

2. shared tools that enable students to share information relevant for the task or relevant for their performance as a group,

3. personal tools for private use, and

4. communication tools that allow teachers and students to communicate with each other.

In this thesis, we focus on the development of two collaborative tools that are part of the CoFFEE software system.

Tatiana

Students’ ongoing actions within CoFFEE are recorded and stored in log files as interaction data. The software system “Tatiana” supports the researcher who have to

14

Introduction

identify, analyze and represent meaningful patterns of interactions from these log files. We used Tatiana to analyze the data of students who worked together with the support of CoFFEE. However, the design of Tatiana is not a topic in this thesis.

1.7 OutlineThe research activities that are discussed in this thesis can be situated at the interface between Educational sciences, Social Psychology, and Information Sciences. The object of the research – small-group discussions in the classroom – is a topic that has received a lot of attention in educational research. Research into collaborative learning provided us with valuable insights into small-group discussions and the kind of problems that students face. A further analysis of these communication problems relies on social psychological insights about group behavior. It tells us more about the functions of communication in terms of the needs, goals, problems and challenges that groups must satisfy or overcome to be effective (Poole, 1999). The results of that analysis serve as input for the design of two collaborative tools that regulate the communication so that certain learning behavior are more likely to occur. The design of the collaborative tools can be placed in a tradition of research – which has its origin in Information Sciences – that aims to develop and evaluate technologies for collaboration (Keen & Sol, 2008).

The multidisciplinary character of the research is visible in the various chapters: some have a strong educational flavor, while other chapters discuss group dynamics issues. A common theme throughout the thesis is communication. Communication is considered as essential for learning to occur in groups, communication processes and structures are a central topic for the study of groups, and communication is one of those human activities that is supported by ICT.

Methodology

The thesis starts with a methodological discussion that covers chapters 2 and 3. There is a need to justify the methodological choices that we made because we do not follow mainstream research where inferences are based on experimental manipulation and independent observations. Instead, we deliberately intervene in an existing real-life learning situation to create the phenomenon under study – networked learning – with the aim to generate new insights about the relationship between communication, collaboration and learning.

15

Chapter 1

Chapter 2 presents the approach that guides the research activities. The research approach employs the design process as a means to generate scientific insights. An artifact – in our case a collaborative tool – will be intentionally designed to intervene in an existing learning situation with the aim to create specific outcomes. The design intervention is based on clear expectations of how the artifact changes the existing situation for the better. These expectations can be conceived as hypotheses that are made applicable for evaluation through the artifact.

We argue that a design-based research approach is related to a distinct social ontology, and hence with a particular practical theory. In chapter 3, we present that philosophical framework. Collaborative learning will be approached from a Systems perspective. This perspective opens up explanations of groups in terms of functions. A functional explanation differentiates between more and less successful groups. It provides us with an explanatory framework that matches with the design-based research approach, i.e. functionalists often design interventions that make desired patterns of communication more likely (Hollinghead, Wittenbaum, Paulus, Hirokawa, Ancona, Peterson, Jehn & Yoon, 2005).

Theory

The second part of the thesis defines small-group face-to-face discussions in the classroom. This part covers three chapters. It starts with a discussion about the communication patterns that are associated with a constructive dialogue. Next, the focus is on the learning environment of the classroom within which the research takes place. Finally, we discuss the design of the collaborative tools that are used in the classroom with the aim to create the proper conditions for a constructive dialogue.

Chapter 4 addresses the epistemological assumptions of Social-constructivist views that conceptualize group learning as a continuous process of collaborative knowledge construction. The co-construction of knowledge requires specific communication patterns. These communication patterns are presented by four criteria for a constructive dialogue. These criteria serve as a reference for the subsequent research activities.

Chapter 5 discusses the learning environment within which we carried out the research. The research that is described in this thesis deals with real-life problems and it aims to formulate authentic solutions for these problems. These solutions are, on its turn, implemented and evaluated in naturalistic learning setting. It means that we have

16

Introduction

little opportunities to examine distinct relationships within a controlled environment. Instead, we need to deal with the situation as a whole or, to put it more formally, as a system. To understand the design interventions, it is therefore necessary to know the nature of the learning environment within which the research takes place.

In chapter 6, we discuss the design of the two collaborative tools. The criteria for a constructive dialogue that have been formulated in chapter 4 served as a reference for the design. They will be used to identify the verbal communication patterns that inhibit learning. An analysis of these ineffective communication patterns focuses on the underlying structural features that organize the verbal exchanges into a coherent and meaningful whole. It acknowledges that a crucial aspect of collaborative learning is the way students manage to coordinate and adjust their actions (Erkens, 2004). The outcomes of the analysis serve as input for the formulation of the design guidelines.

Chapter 6 discusses eight design guidelines that lie at the basis of two collaborative tools. These guidelines contain clear expectations of how the tools mediate and regulate the communication so that certain learning behaviors are more likely to occur. The eight guidelines will be translated into a number of services that describe the behavior of the tool from an user's point of view. These descriptions serve as input for the software development process. It results in two collaborative tools that are presented at the end of chapter 6.

Research

The design guidelines that are discussed in chapter 6 were developed during successive research cycles that lead to a number of adaptations of the initial tool design. The studies that are discussed in third part of the thesis reflect the iterative character of the research. In chapter 7 and 8 we discuss two studies that aim to “strengthen” the tool design. The outcomes of these studies led to a number of adaptations of the design. These improvements had to do with the way the collaborative tools, as a medium for communication, organize individual actions into collective group behavior. They reflect an increased understanding of computer-mediated communication in face-to-face situations.

Chapter 7 presents a study that investigates the first basic property of the collaborative tools, i.e. parallel access as floor control mechanism. Parallel access allows users to access the shared digital workspace simultaneously. Users can put forward their contributions without any delay or interruptions. The study that is presented in

17

Chapter 1

chapter 7 examines the communication problems that the groups experience when their interactions are based on parallel access. These problems had to do with the coordination of the joint actions in the digital workspace. Problems of this kind had to be solved before we could study the learning effects of the collaborative tools.

Chapter 8 elaborates on the second property of the collaborative tools: an orientation towards effective task-performances. It focuses on two distinct communication patterns – coherence and elaborations – that can be associated with effective task performance. The study that is discussed in chapter 8 analyzes the task-related interactions that can be observed in the shared digital workspace of the collaborative tool. The analysis makes a distinction between the content and the sequence of the computer-mediated actions and interactions.

Chapter 9 presents a large study that evaluates the two collaborative tools that are part of the newly developed CoFFEE software system. The aim of the study is to examine if the introduction of CoFFEE led to a more constructive dialogue. Chapter 9 starts with a discussion of the instructional strategy that makes the collaborative tool suitable for the specific context of use. The instructional strategy describes the sequence of learning activities, the expected outcomes, and the collaborative tools that the groups use. Next, we examine if the communication that is supported by the collaborative tools fulfills the minimal requirements of a group discussion. Finally, we compare the communication patterns of groups who only communicate verbally with groups who use CoFFEE in addition to their verbal communication.

Finally, in chapter 10, we draw some conclusions with regard to the research that has been described in this thesis. This chapter reflects the three-part structure of this thesis. We reflect on the design-based research approach, the design of the networked-learning environment and the effects of that environment on small-group group discussions in the classroom.

18

2 Research Approach: Understanding

through Design

We might compare the animal’s successful solution to an expectation and hence to a hypothesis or a theory.

Karl Popper1

The research that is discussed in this thesis follows a problem-solving paradigm that considers the design process as a means to advance scientific understanding. The aim of the design is to intervene in an existing learning situation with the aim to improve that situation. In our case, the design intervention affects the communication between students who meet face-to-face to discuss a problem. One of the most noticeable outcomes of the design process is a concrete artifact, in our case a collaborative tool to support face-to-face discussions. Another result of that process – less tangible but of equal importance – is increased understanding, or what Winograd and Flores (1986) referred to as “the interaction between understanding and creation”.

When designers engage in a design process, they become more knowledgeable about the situation that is addressed by the design. They begin with some initial understanding of the problems that the design aims to solve. These initial conceptions are formulated as expectations that usually stem from designers’ knowledge and experience gained during previous design projects. From there they move towards a deeper understanding of the situation: insights in the problems and possible solutions begin to take shape when the designers take into account the subtleties of the design situation. Initial expectations are continuously evaluated and refined so that they better fit the designers’ perception of the situation. This process of progressive understanding

1 The citation comes from a talk given by Karl Popper on the North German Radio (NDR) on March 7, 1972. The talk has been published in Popper, K. (1999), All life is problem solving.

19

Chapter 2

applies for the professional designer but it also holds true for the design-based researcher. Exactly this process of “understanding through design” comprises the essence of the approach that shapes the research activities that are discussed in this thesis.

In this chapter, we present a research approach that utilizes the design process as a means to generate new knowledge. The chapter starts with a discussion of Action Research as one of the first strategies within the Social sciences that considers action, or more precisely real-life problem solving, as a valuable means for research. Next, we present two recent lines of thought from the fields of Information sciences and Educational sciences that give a central place to design as a genuine research activity. We explicitly choose these two disciplines because the research that is discussed in this thesis lies at the intersection of the two: the research intervenes in an existing learning situation and comes up with a technical solution that has its roots in the Information sciences. The first line of thought considers Design science (March & Smith, 1995) as a valuable method for Information systems research, while the second line of thought from the Educational sciences sees Design-based research as a means to integrate the design process with research (Edelson, 2002). The two lines of thought provide the proper ground for a design-based research approach. Before we address that approach, we introduce the concept of design patterns. This concept provides us with a practical orientation towards the design-based research. Design patterns aim to establish a proper fit between the design and the context of use. Furthermore, the method adopts a specific language that makes knowledge about the design process explicit.

2.1 Action ResearchAction research is one of the first research strategies within the Social sciences that combines action with research. The origin of the approach can be traced back, among others, to the work of the psychologist Kurt Lewin who is known for his dynamic field theory (Lewin, 1936). The dynamic field theory states that human behavior is the resultant of the totality of facts that act upon an individual or a social system. These facts form a social field of interdependent forces that are in a quasi-stationary balance. These forces have direction, distance, strength, and a point of application (Lewin, 2009). Together they determine the actual behavior of human beings or of the social system they are part of. Lewin applied the dynamic field theory to study the process of social change. For Lewin (1947) any actual social change should come from inside the

20

Research Approach: Understandingthrough Design

social system, triggered by people who are part of that system. The scientific method of fact-finding plays an important role in bringing about that social change. It generate valuable insights about the forces that act upon the system.

Action research holds a clear position about the role of the researcher. The researcher, who maps the forces that act upon the system, is not an outside observer. Lewin opposed the idea of a researcher who studies a social system as an independent observer. He states that a social system is difficult to understand from the outside, it can only be known when the researcher interacts with the people that are part of the system and looks beyond the observable to map the underlying facts. Furthermore, Lewin argues that the researcher who wants to understand a social system should exercise influence on the field of forces that determines its behavior (Boon, 1989). It means that knowledge about the social system can only be obtained through active participation of the researcher in the practices directed at improving the system.

Lewin identified two aspects of a social system that make it difficult for an outside observer to understand and predict what will actually happen. These aspects are: 1) a distinction between the “subjective” and “objective” elements of a social system, and 2) a distinction between perception and action. A social system, according to Lewin, is based on divergent perspectives and subjectivity. A person or a group perceives its own situation and the situation of others in a subjective way, while other groups or persons in that same system may perceive that situation differently. The resultant of these subjective perceptions predicts the actual or objective actions. Any change in the social system must take these two aspects – subjectivity versus objectivity and perception versus action – into account. Going from perception to action and from subjective to objective and back again, are no arbitrary demands of a scientific methodology but they mirror a basic property of social life (Lewin, 1947).

It is evident why action research uses real-life problem solving as a research method. Action as the basis for research can be associated with the way that social reality is viewed: dynamic field theory requires that the researcher is involved in activities that aim to change a social system. It makes clear that a methodological framework is closely related to the “way of thinking” (Sol, 1982). In chapter 3, we further elaborate on this issue. In that chapter, we relate an ontology with a methodological framework and a basic theoretical position, i.e. systems thinking, design-based research and the functional perspective.

21

Chapter 2

A Cycle of Related Activities

Action research is based on the active involvement of the researcher in the social system that is the object of change. It combines action with research that is carried out during a process of successive activities. Kurt Lewin (1947) defined action research as “a spiral of steps, each of which is composed of a circle of planning, action and fact-finding about the result of the action”. Various researchers from different scientific disciplines have elaborated on this cycle (see e.g. Checkland & Howell, 2007; Reason & Bradbury, 2006; Carr & Kemnis, 1986; Susman & Evered, 1978). Susman and Evered (1978), for example, came up with an action-research cycle that consists of five steps that resembles Dewey’s (1933) description of reflective thinking (Figure 2.1). The steps are closely linked; they consist of diagnoses, action planning, action taking, evaluation and the specification of learning. This cyclic character of the research is also an essential aspect of the approach that is followed in this thesis. It emphasizes that research findings are never definite.

22

Figure 2.1: The cyclical process of action research (Susman & Evered, 1978).

Research Approach: Understandingthrough Design

The Primacy of Action

It is not our intention to discuss action research into detail. What is important is that action research focuses on interventions in naturalistic settings. The research interventions or actions exert their influence on the forces that together determine the behavior of the social system. These interventions aim to improve the performance of the system. At the same time, they provide an opportunity for learning (see Figure 2.1). Knowledge development is closely linked with action. Knowledge generated through research is more practical of nature than theoretical because understanding emerges when practical problems are solved. Action research goes beyond the notion that theory informs practice, to a recognition that theory can and should be generated through practice (Brydon-Millar, Greenwood & Maguire, 2003). Conceptual models emerge from the data, rather than being tested against that data (Kock, 2003).

Action research, in our view, seems to ignore the “re-framing” character of theories. Theories shed a different light on practical problems while hypotheses that stem from these theories lead to new solutions for existing problems. In the next two paragraphs, we turn our attention towards to two more recent research strategies that stresses the interrelatedness of practice and theory.

2.2 Design ScienceThe aim of Design science is to improve organizational practice by developing novel technological solutions, models or methods. The root of Design science can be traced back to the work of Herbert Simon (1996) who made a distinction between two worlds – a natural world and an artificial one – that are studied by different scientific disciplines. While the natural sciences study natural objects, the artificial sciences study artificial objects created by humans. Simon (1996) stated that the boundaries of the artificial sciences are set by the objects of study:

1. Artificial objects are synthesized by human beings although it is not always that deliberately.

2. Artificial objects may resemble the appearances of natural things while lacking, in one or many respects, the reality of the latter.

3. Artificial objects can be characterized in terms of functions, goals and adaptation.

4. Artificial objects are often discussed, particularly when they are being designed, in prescriptive as well as descriptive terms.

23

Chapter 2

Artificial objects encompass a broad domain. They do not only include material objects but also comprise cultural artifacts like language that mediates human behavior.

Simon’s notion of the sciences of the artificial does not limit scientific method to the testing of a theory through research. The design process that is concerned with the creation of an artifact can also be used to generate new insights. March and Smith (1995) spoke in that context of Design Science as an approach that produces and applies knowledge of tasks and situations to create effective artifacts. Hevner, March, Park and Ram (2004) stated that research involves the creation and evaluation of information-technology artifacts intended to solve relevant problems. They formulated seven guidelines for Design Science in Information Systems Research (Table 2.1).

The guidelines of Table 2.1 guarantee that the design process becomes an opportunity for research. First, the design process should fulfill the requirements of effective problem solving. It addresses relevant problems and comes up with valuable solutions that are thoroughly evaluated. Secondly, the design process should be relevant for those who are affected by the design. It means that the research activities address

24

Guideline Description

Design as an artifact

Design science research must produce a viable artifact in the form of a construct, a model, a method, or an instantiation.

Problem relevance

The objective of design science research is to develop technology-based solutions for important and relevant business problems.

Design evaluation

The utility, quality, and efficacy of a design artifact must be rigorously demonstrated via well-executed evaluations methods.

Research contributions

Effective design science research must provide clear and verifiable contributions in the areas of the design artifact, design foundations, and /or design methodologies.

Research rigor Design research relies upon the application of rigorous methods in both the construction and evaluation of the design artifact.

Design as a search process

The search for an effective artifact requires utilizing available means to reach desired ends while satisfying laws in the problem environment.

Communication of research

Design science research must be presented effectively both to technology-oriented as well as management-oriented audience.

Table 2.1: Design science research guidelines (Hevner, March, Park & Ram, 2004).

Research Approach: Understandingthrough Design

business needs (Hevner & March, 2003). Thirdly, the research requires the application of rigorous methods in both the construction and evaluation of the designed artifact (Hevner et al., 2004). The design must be firmly grounded in a sound conceptual framework, while the evaluation of the design should be based on appropriate methods.

Theory plays a rather modest role in Design science, at least compared to those research strategies that empirically test theories. The design of an artifact is based on theoretical insights. However, existing theories are often insufficient. Researchers also rely on experience, creativity, intuition, and general problem-solving methods (Hevner & March, 2003).

The Inductive Hypothetical Research Strategy

The inductive-hypothetical research cycle (Sol, 1982) can be characterized as an early instantiation of Design science. The inductive-hypothetical cycle utilizes the problem-solving process as a means for research. The strategy takes real-life situations as the starting point for research. Characteristic for the inductive-hypothetical cycle are the various steps between the identification of problems and the formulation of solutions. These steps further conceptualize the problem situation. Sol (1982) argued that a problem needs to be conceptualized and specified before it can be solved. It means that each step of the inductive-hypothetical cycle leads to the construction of specific model types that deepen the researcher's understanding of the situation.

The inductive-hypothetical model cycle (Figure 2.2) consists of five steps: initiation, abstraction, theory formulation, implementation, and evaluation (Churchman, 1971; Bosman, 1977; Sol, 1982; de Jong, 1992). The model recognizes

25

Figure 2.2: The inductive-hypothetical problem solving cycle.

Chapter 2

that the researcher enters a design situation with a set of preconceptions or a “way of thinking” that determines the way a problem is conceptualized. The researcher perceives the current situation through a lens (Keen & Sol, 2008) that distinguishes between essential aspects of the problem situation. That lens reflects a basic choice with regard to “how and what to see in the world”. It guides the construction of a descriptive empirical model of the current situation that – just like a theory – is an interpretation of reality (Morgan, 1986). For example, in our case we look at collaborative learning from a functional perspective that states that the performance depends on how well communication functions within the context of a group to satisfy requisite conditions for successful learning (Waldeck, Shepard, Teitelbaum, Farra & Seibold, 2002).

The descriptive empirical model guides the subsequent activity of abstraction that leads to a descriptive conceptual model that describes the mechanisms that explains essential aspects of the problem situation. A description of these mechanisms is important for the subsequent research activities that mark a transition between descriptive towards prescriptive models. It stresses that finding mechanisms does not only satisfies the yearning for understanding, but also satisfies the need for control (Bunge, 2004).

The next activity of theory formulation characterizes a change in orientation from problem definition towards problem solving, that is from descriptive towards prescriptive models. The aim of this activity is to find appropriate solutions for the problems that are conceptualized and specified during the previous two steps. The alternative solutions are worked out in ‘to be’ models (Janssen, 2001). They consist of a set of design guidelines that represent alternative solutions. The next activity of implementation makes the guidelines operational, which results in a prescriptive empirical model such as an artifact or a set of actions that aims to change the current situation.

Linking Practice with Theory: Solution Finding

The inductive-hypothetical cycle can be typified by a number of transitions from descriptive towards prescriptive, from empirical towards conceptual and vice versa. These transitions link practice with theory. Practical problems are conceptualized and specified during the first two steps. The outcome of these two steps – a description of the mechanisms that explain the problem – gives rise to the actual problem-solving

26

Research Approach: Understandingthrough Design

activities. The transition from a descriptive towards a prescriptive model can be characterized as theory formulation. Theories set the direction for the solutions that are laid down in the prescriptive models. The concept of “theory” is used in a general sense and indicates an explicit and elaborated set of solutions for the original problem statement (van Meel, 1994).

In the next paragraph, we further elaborate on the “theory formulation” activity of Figure 2.2. We discuss a research strategy – Design-based research – that stems from the Educational sciences. This strategy considers the transition from descriptive towards prescriptive as the formulation of a set of hypotheses that clearly state how the situation will change when the solution is implemented.

2.3 Design-based ResearchDesign science is associated with the field of Information Sciences but the research strategy has its counterpart in Educational sciences. Design-based research (Brown, 1992; Collins, 1992; Gravemeijer, 1994; Hoadley, 2002) is a research approach within the Educational sciences that integrates the design process as part of general research approach.

Design-based research marks a shift in focus from the study of individual cognition towards the study of cognition in context. It can be contrasted with the experimental approach that is illustrative for cognitive studies (Table 2.2). The cognitive perspective focuses on distinguishable aspects of human cognition. It decomposes the complexity of human cognition into identifiable processes that are carefully studied by manipulating a number of conditions within the environment (Greeno, 1998). More recently, researchers began to study “learning in the wider sense” (Brown, 1992). They began to pay attention to the learning processes as they emerge in real-life settings such as the classroom. This cast doubt on the traditional methods of research: Paradigms that simply examine learning in terms of isolated variables within the laboratory or other impoverished contexts of participation will necessarily lead to an incomplete understanding of their relevance in more realistic settings (Barab & Squire, 2004; Brown, 1992). This change in research focus and method is closely related with the development of situated and constructivist learning theories that take into account the social and cultural context within which learning takes place. Situated

27

Chapter 2

and constructivist perspectives make room for alternative research strategies that study learning in naturalistic settings.

28

Category Experimentation Design-based research

Location of research Conducted in laboratory settings

Occurs in the buzzing, blooming confusion of real-life settings where most learning actually occurs

Complexity of variables

Frequently involves a single or a couple of dependent variables

Involves multiple dependent variables

Focus of research Focuses on identifying a few variables and holding them constant

Focuses on characterizing the situation in all its complexity, much of which is not know a priori

Unfolding of procedures

Uses fixed procedures Involves flexible design revision in which there is a tentative initial set that is revised depending on their success in practice

Amount of social interaction

Isolates learners to control interaction

Frequently involves complex social interactions with participants sharing ideas, distracting each other, and so on

Characterizing the findings

Focuses on testing hypothesis Focuses on the development of models that characterizes the design in practice

Role of participants The researcher makes all the decision

Involves different stakeholders who bring in their differing expertise into producing and analyzing the design

Table 2.2: Comparing psychological experiments and design-based research methods (Collins, Joseph & Bielaczyc, 2004; Barab & Squire, 2004).

Research Approach: Understandingthrough Design

The Role of Design and Research

Design-based research aims to improve a learning practice; it employs the design process to generate scientific insights. The approach combines educational research with theory-driven design in order to produce new theories, artifacts, and practices that account for and potentially impact learning and teaching in naturalistic settings (Barab & Squire, 2004).

Theory-driven Design

The design-based research approach does not merely study empirically what works. The creation of new learning activities is critical, not only to evaluate hypotheses but also to develop models and theories. The design-based research approach starts with the analysis of problems in actual learning settings. The problem analysis characterizes the goals, needs, or opportunities that a design intends to address, together with the challenges, constraints, and opportunities presented by the learning context (Edelson, 2002). The problem analysis gives rise to the formulation of an initial set of hypotheses and principles that guide the subsequent design process. Educational designs are implemented with a hypothesized learning process and the means of supporting it in mind (Cobb, Confrey, DiSessa, Lehrer and Schauble, 2003). Importantly, these hypotheses and principles are not detailed enough to determine every design decision (Edelson, 2002). That is because the hypotheses do not only reflect theory, they also take into account the complexity of the learning setting. Such a real-life setting cannot be fully know beforehand. The researcher must be able to make additional decisions when the educational design is introduced in the classroom. Design-based research can be characterized by continuous testing and refinement of hypotheses that makes the design more appropriate for its context of use. Here, a crucial difference emerges between Design science that was discussed in the previous paragraph and Design-based research. The former stresses understanding before design while the latter emphasize understanding through design.

Educational Research

Research within design-based approach generates insights about the effects of the design. It must account for how the design functions in authentic learning settings. These accounts are opportunities to advance the researchers' understanding of teaching, learning or the educational systems. Through a parallel and retrospective

29

Chapter 2

process of reflection upon the design and its outcomes, the design researchers elaborate upon their initial hypotheses and principles, refining, adding and discarding them and gradually knit together a coherent theory that reflects their understanding of the design experience (Edelson, 2002). This iterative design process reflects progress in understanding of the learning situation.

2.4 The Practice of DesignDesign science and design-based research suggest several general directions for research: the research must deal with real-life problems that matter and it should reflect characteristics of effective problem solving. We further elaborate on these requirements by discussing the concept of design patterns.

A design pattern describes a solution for a recurrent problem that takes into account the subtleties of the context in which the problem occurs and the solution is implemented. The concept originates from architecture and was first introduced by Christopher Alexander as a description of a problem which occurs over and over again. A design pattern describes the core of the solution to that problem in such a way that you can use that solution a million times over, without ever doing it the same way twice (Alexander, Ishikawa, Silverstein, Jacobson, Fiksdahl-King & Angel, 1977). The notion of design patterns is an elaborated concept that originally addressed the issue of fit between form and context in architectural design. Architecture is a domain that has a long history of design practice that evolved from “unselfconscious” design, embedded in tradition towards a “self-conscious” design that is based on formal models and methods (Alexander, 1964). The notion of design patterns is a response to the latter.

In unselfconscious cultures, there is a natural fit between the design and the context of use. Designers operate in an architectural tradition that is guided by a set of rules and norms attached to the building habits of that culture. Design and use cannot be separated because those who live in the buildings are the same as those who design and built them. The unity of the two roles leads to fluent adjustments of the design when conditions in the context change. Changes in the context that result in failures in the design – e.g. a shortage of certain building material – are directly incorporated into the design. Necessary adaptations of the design occur at the micro level but they are anchored at the level of society because they are incorporated in the traditional building habits. Tradition prevents any radical change in the design, and at the same time, reinforces proven adaptations. Consequently, unselfconscious cultures produce

30

Research Approach: Understandingthrough Design

clear design patterns that own their existence to the good fit that they provide between form and context.

Architectural design in modern, self-conscious cultures is faced with more complex design problems. These problems require formal methods that consider the design process in a rational manner. Alexander (1964) discussed the drawbacks of these rational methods: they lead to a detachment of the design process from its context of use. Formal methods separate the problem analysis from the design and the design from use. These methods decompose a complex problem into separate parts that are solved independently, at the expense of an integrated view. Furthermore, formal methods rely on expert designers while users hardly have anything to say. This sets the design process even further apart from its context of use.

Alexander (1964) stated that formal design methods result in a loss of fit between form and context. This misfit can be overcome by the notion of design patterns and the use of a pattern language to express the essence of the design. Patterns represent the organized ways in which people use the material arrangements and the problems they encounter in the course of use (Crabtree, Hemmings & Rodden, 2002).

An Ecological Approach

The notion of design patterns has been applied outside architecture in disciplines like software design, interaction design and educational design. There is the danger that the meaning and use of the concept may change when it is applied to other domains, with the risk that the richness of the underlying method may be neglected (Goodyear, 2004). The notion of design patterns encompasses a fundamental view towards the design process. Characteristic for design patterns is the ecological approach that deals with the complex relations between elements of the system and its surrounding context. An ecological approach integrates all the relevant aspects of the design process. It reunifies analysis and synthesis, form and context, and the different actors involved in the design, which results in a product that is optimally adapted to the micro-structures of local conditions and constraints (Lea, 1994).

Pattern Language

Alexander (1979) asserted that the complex structure of a design pattern has to be made explicit. Which brings us to another characteristic of design patterns; they are expressed in a formal language. Design patterns must be made explicit, so that they can

31

Chapter 2

be shared and reused. The research approach of this thesis incorporates Alexander’s (1979) three-part rule that defines a pattern as a relationship between a certain context, a problem, and a solution. This is seen as a useful means for sharing knowledge about the design. Often, a pattern language is used to communicate the knowledge of the expert who is perfectly aware of the subtle relationships between context, problem and solution. Not surprisingly, design patterns are considered a good candidate within ethnography to give form to the design (Crabtree, 2003).

2.5 Way of WorkingThe research approach that frames the research has been developed within the context of the LEAD project (van Diggelen & Overdijk, 2006). It took into account the limited time available that we had to develop and evaluate novel solutions for existing problems. The project team stared from scratch had to come up with a working prototype within 18 months.

The approach was inspired by the inductive-hypothetical model cycle (Sol, 1982) that has been adapted to an educational context. The inductive-hypothetical model can be characterized by understanding before design. It places great emphasis on the development of empirical models that provide a proper description of the problem. We just lacked the time to develop extensive empirical models that are grounded in several cases. The initiation phase of the project took 6 months after which we moved towards the phase of conceptualization. The problems that guided the research came from educational practice, but our understanding of these problems evolved during the project. In that sense, we were inspired by the design-based research approach that stresses understanding through design.

There was the danger that we might address the wrong problems or that the problems were not properly described. Two strategies were chosen to counteract the danger of insufficient validation. First, we relate the practical problems that were identified at the start of the project to earlier research. In that sense, we use the scientific community as a community of practice and searched for confirmation within that community. Secondly, the design process consisted of an iterative cycle of design, evaluation, and refinement of the design. This iterative cycle reflects the notion of co-evolution: understanding of the problems and the solutions co-evolves through a design that is repeatedly framed by practice.

32

Research Approach: Understandingthrough Design

Our research approach incorporates the notion of Design-based research that considers the formulation of hypotheses is essential for design-based research. These hypotheses contain clear expectations of how the design creates specific learning behaviors. They mark the transition from descriptive towards prescriptive models. Subsequently, the research that evaluates the design aims to test and refine the hypotheses. In our case, this is done through micro-level descriptions of communication patterns that are associated with group learning.

The research approach that framed the research activities consisted of five steps (Figure 2.3):

1. initiation,

2. abstraction,

3. theory formulation,

4. implementation, and

5. evaluation.

33

Figure 2.3: Research activities and outcomes.

Chapter 2

Initiation: Problem and Context

The research started with a sense that an existing learning situation can be improved. This notion gave rise to an analysis of the human activity system – that is, face-to-face discussions in the classroom – with the aim to describe the recurrent problems that affect the learning achievements. The descriptions of the problems were based on empirical observations as well as on insights obtained from studies that are reported in literature. We looked at these problems from a functional perspective that states that group performance depends on how well communication functions. In our case, we focused on the verbal communication patterns that can be observed during a face-to-face discussion. First, we defined what constitutes a productive group discussion and identified the communication patterns that inhibit groups from learning effectively. During this step we answered the first two research questions (Q1 and Q2) that aim to identify the criteria for a constructive dialogue and the ineffective communication patterns that hamper collaboration and learning.

The initiation phase of the research is described in chapter 4 and 5 where we give a more detailed description of the research context. Chapter 4 provides us with a evaluative framework for analyzing group discussions in the classroom. This framework helped us to interpret the problems that groups may experience during a discussion. In chapter 5 we present some thoughts about learning that we came across in the educational practice. Two dominant views – direct instruction and collaborative learning are explained. Chapter 5 also gives an overview of the various aspects of the learning environment that may affect the research and design activities. The initiation phase ends at the beginning of chapter 6 where we describe a problematic communication pattern – interpersonal dominance – that gave rise to the research.

Abstraction: Underlying Mechanisms

The recurrent nature of the problem implies that it goes beyond a particular situation, in other words, the problem can be described in terms that are more general and still relates to a specific observation. The activity of “abstraction” aims to describe the problems in general terms by describing the mechanisms that cause these problems.

The functions-mechanisms relation can be characterized as a one-to-many relation (Bunge, 2004). A function of a social system can be realized through different mechanisms. For example, the sharing of knowledge during a group discussion is

34

Research Approach: Understandingthrough Design

usually realized through verbal exchanges. As we will argue in this thesis, groups could also share knowledge with the support of a collaborative tool that mediates their communication. This kind of communication is based on different kinds of mechanisms that performs the same function, i.e. the sharing of knowledge.

The communication patterns that were identified during the previous step are described in terms of the underlying mechanisms that organize group talk into a meaningful and coherent whole. This description will answer the third research question (Q3) that provides us with deeper insight in the causes of the dysfunctional communication patterns.

Abstraction does not only deepens the description of the problems identified during the previous step, it also broadens our view of the problems that come into play. A second problematic communication pattern – product blocking – is described at the beginning of chapter 6. This chapter forms the core of this thesis. It describes the three activities of abstraction, theory formulation and implementation.

Theory Formulation: Hypotheses and Guidelines

Theory formulation marks a transition from descriptive towards prescriptive models. The prescriptive models describe the envisioned learning practice. This is done through a process of interpretation that draws upon theoretical insights but also on the creativity, intuition, and general problem solving methods of those involved in the design (Hevner & March, 2003). The underlying mechanisms that were identified during the previous step set the direction of change. These mechanisms can be considered as the causes that need to be addressed if any actually change will occur.

The activity of theory formulation results in a number of design guidelines that describe how the envisioned tools affect the current practice for the better. These guidelines are grounded in a theoretical body of knowledge with regard to collaborative learning, group dynamics and computer-mediated communication. They contain a set of hypotheses that set the direction of change that the design will bring about.

General and Specific Guidelines

The collaborative tools that are part of the networked-learning environment must be used effectively in different kinds of learning situations. These situations differ considerably: for example, the group size may vary, the educational level can range from secondary to academic education, and various collaborative learning activities are

35

Chapter 2

supported. In sum, the collaborative tools remain more or less the same while their usage differs from situation to situation. This seems at odds with the need of a good fit between the tools and a specific context of use. To solve this issue we made a distinction between two sets of guidelines: general guidelines and specific guidelines.

General guidelines apply to a variety of collaborative learning situations. They have to do with the fundamental change that the tools will bring about. These guidelines describe the basic properties of the collaborative tools. It means that any adaptation of a general guidelines creates a fundamental different kind of support.

Specific guidelines make the general guidelines applicable to a specific context. These guidelines preserve the basic properties of the tools from situation to situation. They ensure that the intended effects associated with the basic properties actually do occur in a specific learning situation.

The specific guidelines are suitable for micro-adaptations; they can be adapted to the specific context of use. Not every guideline can be adapted that easily without affecting the basic properties of the tools. Micro-adaptations aim to solve practical issues that slightly vary from situation to situation. The specific guidelines, in our case, relate to the instructional strategy that accompanied the introduction of the tools.

Implementation: Functional Models

Implementation is a natural continuation of the previous activity of theory formulation. The essence of the tool design is to translate the design guidelines that were formulated during the previous step into a functional model and subsequently to a concrete product. The functional model prescribes the functions and appearances of the envisioned collaborative tools. It refers to an object, often a sketch, model or a set of instructions that is a preliminary stage in the process that leads to the finished product (Raizman, 2003). At a more abstract level, the tool design consists of a set of guidelines that lay down essential aspects of the collaborative tools. At a more tangible or empirical level, these guidelines are represented as mock-ups, sketches of the user-interface and scenarios that describe the envisioned use of the tools.

Evaluation: A Cycle of Research

The next step – evaluation – addresses the fourth research question (Q4) that evaluates the collaborative tools that were developed during the previous step. The research

36

Research Approach: Understandingthrough Design

activities that will be carried out during this step must account for when, how and why the tools function in practice. Through research, we draw conclusions about the hypotheses that form the basis of the design guidelines. Comparing the predictions of the design guidelines with observed data enables the researcher not only to test the design, but also to pinpoint problematic aspects of the design (Poole, McPhee & Canary, 2002). If necessary, the hypotheses, design guidelines, and subsequently the collaborative tools will be adapted.

The activity of Evaluation was addressed by a number of studies that are presented in chapter 7, 8, and 9. These three chapters reflect the cyclic character of the research whereby each study triggers a process of reflection upon the design and its outcomes. After each study we evaluated the design guidelines and gradually constructed together a coherent model for computer-mediated communication during face-to-face discussions.

Communities of Practice

Alexander (1964) stressed the importance of an integrated approach that combines all the relevant aspects of the design process, such as thinking and doing, design and usage, solutions and problems, and system and context. Usually, this kind of integration requires considerable application of expert knowledge, rooted in the rich set of experiences of the individual designer or in a history of best practices. Both conditions were lacking in our situation. We did not have “proven solutions” (Haberman, 2006) because we focused on a relatively new situation of mediated face-to-face discussions the classroom. Instead of referring to knowledge and experience of the individual designer, we based the design on scientific knowledge and insights. We used the knowledge that is available in different communities of practice (Wenger, 1998), in our case the scientific communities that deal with collaborative learning, the study of groups, and computer-supported collaborative work. The theories and insights that are shared by these communities provide us with the theories and models of how networked learning affects collaboration and learning in the co-located setting of the classroom.

Iterative Design Cycle

Insights in the problem, solution and the context – the three elements of a design pattern – are generated during different steps. The problem description results in the

37

Chapter 2

first element: the problem. The set of empirical verifiable hypotheses and design guidelines refer to the second element of the design pattern: the solution. Insights obtained from the research activities lead to the third element: the context of use. Understanding of the three elements co-evolves when the researcher progresses towards a solution that actually works.

The notion of co-evolution implies that the different elements cannot be set apart; their mutual relationships must be taken into account. This requires an iterative design cycle where a solution is evaluated in relation to 1) the problem that it aims to solve, and 2) the context within which problems emerge and solutions are implemented.

Each phase of evaluation, in our case, led to subsequent design cycles where context, problem and solution were redefined because understanding about them has increased. So gradually a more coherent set of design guidelines appeared that reflects progress in our understanding of networked learning in face-to-face situations.

38

3 Social Reality Explained: Systems and

Functions

Nor would we explain the power of water to extinguish fire by deriving it from the powers of its constituents, for oxygen and

hydrogen are highly inflammable.

Andrew Sayer1

When we study a particular scientific problem, we first need to specify what social reality is and how to begin to explain it (Archer, 1995). Therefore, before we enter into an exploration of the research topic, we discuss the ontological claims and the methodological statements that frame the research. We follow the sociologist Margaret Archer (1995) declared that a social ontology, an explanatory framework, and a practical theory constitute each other and should therefore correspond. In this thesis, the complexity of social reality is approached from a Systems perspective. Systems thinking is just one way to get a hold over the complexity of the social reality. It acknowledges that the intricacy of everyday life cannot be captured in one, unifying framework. A Systems perspective describes reality on various levels of complexity, each with its own distinct properties.

Social reality from a Systems perspective consists of functional entities or systems that are related. This makes it possible to come up with functional explanations. In our case, it allows us to describe the performance of groups in terms of the effectiveness of the communication processes. A basic premise of the functional perspective is that performance depends on how well communication satisfies the requisite conditions for successful learning (Waldeck, Shepard, Teitelbaum, Farra & Seibold, 2002). The functional perspective is in accordance with the Design-based research strategy that was discussed in chapter 2. A basic premise of design-based research is to create particular

1 Sayer, A. (1992). Method in Social Science: A Realist Approach.

39

Chapter 3

forms of learning in naturalistic learning settings. A functional perspective provides the explanatory framework for such an approach. Functionalist often design interventions to control and regulate interactions with the aim to create desired group behaviors (Hollingshead, Wittenbaum, Paulus, Hirokawa, Ancona, Peterson, Jehn & Yoon, 2005).

3.1 Way of ThinkingAn ontology implies a way of thinking that pervades how we understand the world in general (Sol, 1982; Morgan, 1986). It contains a number of basic choices with regard to “how and what to see in the world”. These choices differentiate on fundamental aspects of social reality and lay down the possible explanations of the social phenomenon under study. The ontology dictates what kind of explanatory mechanisms a theory contains.

The dominant ontology within a scientific discipline may vary: reductionism, for example, is the leading ontology to study the natural world. It describes physical phenomena in terms of a limited number of particles, like atoms and molecules, which are responsible for “nearly all the properties of matter that have shaped the world around us” (Beiser, 1981). For social phenomena, the picture is more complex: for example, within a scientific discipline like Sociology there is a vivid debate about the social ontology that should govern social explanations (see e.g. Bunge, 1979; Archer, 1995).

Reductionism

As mentioned before, we approach social reality from a systems perspective. A Systems perspective can be contrasted with a Reductionist and a Process ontology. Social reality from a Reductionist ontology is broken down to a limited number of entities that behave in a regular manner. Explanations that can be associated with Reductionism decompose social phenomena to the individual: every possible object is either an individual or a collection of individuals (Bunge, 2000). For many years, theories of collaborative learning tended to focus on how individuals function in a group (Dillenbourg, Baker, Blaye & O'Malley, 1995). These theories conceptualize the processes and outcomes of collaborative learning in individualistic terms. Research within this tradition, was mainly concerned with how individuals process information, and not so much with the

40

Social Reality Explained: Systems andFunctions

dynamic nature of group processes.

Reductionism entails a basic problem of ontological nature: it denies the social dimension of human performance (Archer, 1995). It neglects that human beings create social structures that have an existence of their own. It is found wanting for making no room for social relations or emergent properties (Bunge, 1979). Group interactions and social structures that are illustrative for collaborative activities serve only as a background for individual activity rather than as a focus of research in itself (Dillenbourg, et al., 1995).

Process Ontology

A Process ontology considers human activity as the quintessence of social reality: individuals and social structures can only be conceptualized by looking at the activities they play a part in. The two have no meaning outside the activity. A Process ontology regards the individual and the environment as inseparable; it holds that only processes that bring these two together are real. Individuals and the sociocultural environment within which they are situated can only be studied by looking at human activities.

Sociocultural theories of learning can be typified by a process ontology (Packer & Goicoechea, 2000; Sawyer, 2002). The basic unit of analysis of these theories is the social activity (Dillenbourg et al, 1996). Learners are viewed as coming into contact with, and creating, their surroundings as well as themselves through the actions in which they engage (Wertsch, 1991). Sociocultural theories stress that human thinking only exists in relation to the social practice and the cultural artifacts that frame human thought.

A Process ontology brings about methodological problems. A true Process ontology is holistic: it views the individual as inseparable from the sociocultural environment and it does not distinguish between distinct processes (Sawyer, 2002). Because a Process ontology examines a single process in the present tense, issues surrounding the relative independence, causal influence and temporal precedence of the components have been eliminated (Archer, 1995). It means that for those who are engaged in educational design it is almost impossible to make any inference about the effects of the design because an intervention has no meaning outside the practice or activity. Instructions, tools or tasks receive their meaning when they are used in practice and they cannot be set apart from that practice. This makes it almost impossible to come up with an intended design.

41

Chapter 3

3.2 Systems ThinkingWe adopt a Systems perspective to study collaborative learning. Systems thinking can be traced back to the work of Ludwig von Bertalanffy who developed a general framework to study phenomena in the natural world. He introduced the concept of an open system that maintains itself in exchange of materials with the environment (von Bertalanffy, 1950). An open system consists of a collection of elements or components that are linked with each other and with a common environment. Such a system can be contrasted with a closed system that has no inputs or outputs and, therefore does not change internally.

Ontological Claims

Reality, according to the Systems perspective, consists of a large number of systems that behave in a regular manner. Pickel (2007) gave an overview of the ontological claims that can be associated with Systems thinking (Table 3.1). Early concepts of systems thinking define an open system in terms of the output that the system produces. It assumes that systems have clear objectives that can easily be identified by their outputs. This perspective has more or less been abandoned when scientists applied Systems

42

Ontological claims of systems thinking

Systems are the basic entities of the natural and social world.

Systems are real entities.

There are material, mixed, and non-material systems.

Each concrete system is directly or indirectly related to all other systems which form their proximate or distal environment.

While some systems are nested and ordered hierarchically, others are non-nested and overlap.

Systems have a different spatial and temporal reach.

A system consists of components and their relations with each other. Particular important are, in addition, a system’s environment as well as the key processes that make it work.

In addition to linear or proportionate causal effects, there are non-linear or disproportionate causal effects.

Table 3.1: A modern systems ontology (Pickel, 2007).

Social Reality Explained: Systems andFunctions

thinking to understand the social world. System behavior cannot be captured by a straightforward reference as the output when humans are part of that system. Such a “hard” reference ignores the variations of human action and the underdeterminacy of social structures. Human beings have, within certain rational limits, the freedom or autonomy of choice. They seem to enjoy a kind of “bounded autonomy” (Child, 1997).

Modern Systems thinking acknowledges that social systems do not have any hard reference such as the output of the system. Still, a basic aspect of a system is the regular patterns of behavior that distinguish the system from its environment. The sustaining mechanism identified by modern Systems thinking is homeostasis, defined as the maintenance of equilibrium by a tendency to compensate for changes in the system or in its environment. This equilibrium needs not to be static; it can be dynamic and indeed may respond to outside disruptions (Harrington, 1991).

Methodological Statements

Systems thinking does not only make ontological claims, it also contains a number of methodological statements. Methodology, broadly conceived as an explanatory framework, is the necessary link between a social ontology and practical theory (Archer,

43

Methodological statements

The conception of system is basic to and relates all sciences and disciplines. Individual humans are both systems and components of systems.

Systems exist independent of the models, conceptualizations, or theories through which we try to understand and explain them.

Materialist and idealist reductionism in the social sciences are rejected.

Conceptualizations in terms of ‘part-whole’ or ‘base-superstructure’ are insufficient to capture the complex order of real social systems.

Time and space are crucial dimensions in accounting for systems.

While the concept of system as entity may suggest stasis, the mechanisms or dynamics of any system are central in explaining emergence, persistence, and dissolution of concrete systems.

Causal relationships cannot be inferred from linear correlations.

Table 3.2: Systems methodology (Pickel, 2007).

Chapter 3

1995). The explanatory framework of systems thinking consists of a set of basic principles that characterizes interpretations associated with a certain ontology. These principles lay down the kind of explanatory mechanisms that a theory may contain. Table 3.2 gives an overview of the methodological statements of Systems thinking (Pickel, 2007). These statements provide us with a framework to study groups as complex and dynamic systems.

Dynamic Systems

To capture the complexity of the social reality, modern System thinking stresses the dynamic nature of systems. Humans, and the systems they create, are characterized by a number of features of dynamical systems, most notably nonlinearity, multiple interactive parts, and system evolution (Burlingame & Fuhriman, 1997). A social entity from a dynamic perspective is a self-contained set of components that interact in complex, often nonlinear ways to form coherent patterns. (Levine & Fitzgerald, 1992; Vallacher & Nowak, 1994). The properties that emerge from these interactions may change over time. It means that the essence of a complex system can only be identified by capturing the evolution of a system’s behavior on some time scale (Abraham & Shaw, 1982).

Organized Complexity

Systems thinking emphasizes the stratified nature of social reality; its complexity cannot be captured in one description. This idea is typical for the social world, which became the subject matter of systems thinking (Checkland, 1981). On each systematic level of complexity, one can identify properties or qualities that cannot be explained on other levels. These emergent properties set Systems thinking apart from Reductionism. To speak of emergence implies a stratified social world including non-observable entities, where talk of its ultimate constituents makes no sense, given that the relational properties pertaining to each stratum are all real (Archer, 1995). The citation of Sayer (1992) at the beginning of this chapter about the characteristics of water and its compound components eloquently makes this clear: oxygen and hydrogen are highly inflammable, while water extinguishes fire.

The notion of organized complexity implies that some structural properties – like group norms or the rules to organize a sequence of talk – are situated at the level of the group. These properties cannot be reduced to the characteristics of the individuals who

44

Social Reality Explained: Systems andFunctions

form a group. Furthermore, these properties are real, and have autonomous causal powers, just like real properties at any other level of analysis (Sawyer, 2003). Due to these emergent properties, a group as a system cannot be derived from its parts. The group is an independent framework in which the parts are placed (Angyal, 1941). The notion of organized complexity has important implications for the analytical framework. A behavioral pattern that can be observed at one level should be described by taking into account 1) the elements of that systematic level, 2) the structural connections between these elements, and 3) the properties that emerge from the collection of elements that are coupled into a specific dynamic structure and allowed to change over time (Levine & Fitzgerald, 1992).

3.3 Collaborative Learning: Three Levels of ComplexityA central notion of Systems thinking is the idea that collaborative learning can be described at various levels or strata of complexity. These levels can be typified by their own emergent properties.

A group from a Systems perspective consists of a limited number of students who interact with each other. When these students interact as a group, a different kind of dynamics occurs with its own properties. In fact, the educational rationale behind collaborative learning rests on the assumption group learning is based on distinct learning mechanisms that cannot be reduced to individual cognition.

The idea of emergent properties and several levels of complexity have important implications for the analysis of collaborative learning. The dynamics of interacting students must be studied at various systematic levels, where each level has its own specific patterns and properties. We make an analytical distinction between three levels: the individual, the group, and the learning environment of the classroom (Table 3.3).

The Individual Learner

If one wants to understand aspects of group interaction process, one must take the group members' properties into account (McGrath, 1984). Group members are an important antecedent variable that needs to be foregrounded when studying group communication (Keyton & Frey, 2002).

Characteristics of the individual can only be observed from external observable cues (Ajzen, 1988). In our case, the most important cues are the communicative

45

Chapter 3

actions of the individual group members. These behaviors reflect group member's thoughts, feelings, and behaviors about the topic of discussion. This distinction between thoughts, feelings and behaviors is used to classify the different communicative acts (Hilgard, 1980; Ajzen, 1988):

1. Cognitive responses are expressions of beliefs that link an object of discussion with certain characteristics or attributes. It reflects the reasoning processes of the individual learner whereby the learner makes an appeal to the knowledge that he or she has with regard to the topic of discussion. In chapter 6, we refer to these responses as task-related communication.

2. Affective responses have to do with evaluations of, and feelings towards a topic of discussion. The object of discussion may vary: it may include responses that refer to the task, other group members or the group as a whole. In chapter 6, we call these affective responses social-emotional communication.

3. Conative responses are behavioral inclinations, intentions, commitments, and actions that can be observed during a discussion. The analyses of the computer-mediated communication patterns that are discussed in chapter 7, 8 and 9 take these conative responses into account.

An example of a property that can be situated on the individual level is dominant behavior of a group member. Dominant behavior can be considered as an individual

46

Social strata Observable behavior Examples of properties

Classroom Classroom interaction Classroom culture

Group Patterns of intragroup communication.

Group normsRules and conventions to organize the sequence of individual talk

Individual The communicative actions of individuals.

Cognitive responsesAffective responsesBehavioral inclinations

Table 3.3: Three levels of complexity to study collaborative learning in the classroom.

Social Reality Explained: Systems andFunctions

trait that manifests itself by behavioral inclinations such as coercively taking turns or interrupting other speakers. However, interpersonal dominance manifests itself differently in various groups. The way a dominant person exerts his or her influence depends on structures that are positioned at the level of the interacting group. For example, the group's authority structure influences the occurrence of dominant behavior of individual group members (Bales, 2002).

The Group of Interacting Learners

The group level focuses on processes and structures that frame the interactions between the group members. In this thesis, we focus on the communication patterns that join together the group members as a social system (Mabry, 1999). It differs considerably from the individual level of analysis, in the sense that the group is approached as a collective unit with an emphasis on the patterns of intragroup interaction (Poole, Keyton and Frey, 1999).

On this level, we make a distinction between: 1) the communication problems that the groups experience during their collaborative learning activities, 2) the communication patterns that causes these problems, and 3) the underlying mechanisms or structures that explain these patterns (Figure 3.1).

The analysis that is discussed in this thesis starts with the identification of problems that prevent the groups from creating the proper conditions for learning. The

47

Figure 3.1: Group level: processes and structures.

Chapter 3

analysis concentrates on those problems that can be traced back to the communication processes. These problems are described in terms of observable patterns of interaction. These patterns refer to the regular sequence of communicative acts. They are further described in terms of the underlying structures that organize the individual actions into a coherent and meaningful whole. These structures govern and guide the coherence of recurring patterns of group behavior (Cushman & Kovacic, 1994). By focusing on these structures, we make the underlying root causes visible. Examples of underlying structures include role differentiation and authority. In this thesis, we concentrate on the structures of the medium that organize individual exchanges into a sequence of coherent and meaningful talk. We identify two media for communication, i.e. a human medium for verbal face-to-face communication and a digital medium to support computer-mediated face-to-face communication.

The Classroom

The level of the classroom situates the group in the larger learning environment. This level focuses on the interactions that go on in the classroom. Important elements of the classroom environment are the teacher and different kinds of media that bring the knowledge domain to the learner. The teacher, for example, is an important actor on this level who shapes the learning processes. Teacher communication encompasses both establishing and maintaining order, designing effective instruction, dealing with students as a group, responding to the needs of individual students, and effectively handling the discipline and adjustment of individual students (Emmer & Stough, 2001).

The communication processes that can be situated on the level of the classroom are not an object of detailed analysis. In this thesis, we mainly focus on the communicative actions of the individual and the communication patterns of the group. The research examines the relationships among people, tools, and tasks, activated by individual behavior and group structures that change and evolve as the group interacts over time (Arrow, McGrath & Berdahl, 2000). The level of the classroom is used as background that enables us to position the research within the larger classroom environment. This level is further described in chapter 5.

48

Social Reality Explained: Systems andFunctions

3.4 Understanding Groups: A Functional PerspectiveSystems thinking makes it possible to understand collaborative learning from a functional perspective. Functional descriptions call for a concrete and detailed account of the mechanisms, which operate to perform a designated function (Merton, 1967). This perspective accords with the design-based research method that has been discussed in chapter 2. A major objective of design-based research is to generate descriptions in terms of organization and function (Simon, 1996).

Performance Differences

The functional perspective provides us with a general theoretical framework to study groups. It acknowledges that groups who learn together may differ in their performance: the processes and outcomes of groups may vary, even if they all work within the same learning environment. The functional perspective focuses on the origins of these performance differences. It is a preferred research approach of researchers who seek to understand group performance effectiveness (Wittenbaum, Hollingshead, Paulus, Hirokawa, Ancona, Peterson, Jehn & Yoon, 2004). It offers researchers a chance to test hypotheses about the success or failure of groups (Cummings & Ancona, 2005).

The functional perspective states that communication is the instrument by which group members, with varying degrees of success, achieve their learning goals (Gouran & Hirokawa, 1996). The perspective guides much of the research into collaborative learning (see e.g. Barron, 2003; Kneser & Ploetzner, 2001; Sfard & Kieran, 2001; Hogan, Nastasi & Pressley, 2000; Keefer, Zeitz & Resnick, 2000). All these studies observed that the performance of groups varies and they trace these performance differences back to variations in the communication patterns. They aim to identify communication patterns that promote or hamper group learning. For example, a number of studies identified elaborative patterns of interaction as the most productive pattern for group learning (Hogan, Nastasi & Pressley, 2000; van Boxtel, van der Linden & Kanselaar, 2000). Typical for an elaborative pattern is that students build upon each other’s contributions by means of evaluating, adding or refining the explanations shared by the group.

49

Chapter 3

Core Assumptions

The functional perspective aims to understand group performance in terms of inputs and interaction processes. It assumes a sequential, causal string: input factors influence interaction processes, which in turn influence group performance outcomes (Wittenbaum et al., 2004). The relationship between input and output is not a direct one; it is mediated by processes that occur during group interaction. The functional perspective can be defined by four core assumptions (Hollingshead et al., 2005):

1. Groups are goal oriented.

2. Group behavior and performance varies and can be evaluated.

3. Interaction processes have utility and can be regulated.

4. Internal and external factors influence group performance via interaction.

Groups are Goal Oriented

Groups have one or more goals that they want to achieve. These goals may refer to the task, the accessibility of resources or the social-emotional well being of the group members (Stangor, 2004). Much of the research from the functional perspective has focused on the effective accomplishment of task-oriented goals (Hollingshead et al., 2005). This also holds true for the research that is presented in this thesis. In our case, it means that we focus on the goals that has to do with learning achievement. The groups that are the object of study have to obtain a learning goal.

Group behavior and performance varies and can be evaluated.

The functional perspective assumes that the performance of groups varies and that it can be evaluated by some standard (Hollinghead et al., 2005). In our case, we evaluate the performance of groups against the criteria of a constructive dialogue during which students collaboratively explore their thinking and develop a shared understanding. These criteria served as a reference for the research and design activities that are discussed in this thesis.

Interaction processes have utility and they can be regulated.

The functional perspective states that the groups should perform certain learning behaviors in order to be successful. Not all groups will display these learning behaviors to the same extend. Steiner (1972) spoke in this context of “productivity losses” due to

50

Social Reality Explained: Systems andFunctions

faulty individual and collective behaviors like frustration, competing motivations or inadequate understanding. To increase the likelihood that groups will meet their goals, functionalists often design interventions to stimulate the occurrence of certain interaction that make the desired patterns of action during group interaction more likely (Hollingshead et al., 2005). In our case, the intervention will be in the form of a collaborative tool that mediates and regulate part of the face-to-face communication so that a group is better able to carry out a constructive dialogue.

Internal and external factors influence group performance via interaction.

Internal and external factors influence group performance. These factors can be positioned at the level of the individual, the group or the larger learning environment. Internal factors like the composition of the group or external factors like the instructions they receive affect the communication that goes on in the group. Two external circumstances that regulates the communication will the focus of the research: 1) the medium for communication, and 2) the learning instructions that the students receive. The aim of the research is twofold. First, two collaborative tools will be developed that offers the group an alternative medium for communication. Secondly, an instructional strategy will be formulated that makes these tools suitable for a specific context of use.

The functional perspective views group outcomes as a linear function of inputs and processes, it cannot explain cyclical, nonlinear group dynamics, or reverse causality (Hollingshead et al., 2005). The perspective might be a too simple framework to explain the complex and dynamic character of groups as a whole. Still, the functional perspective enables us to zoom in to specific communication patterns and relate these patterns to distinct mechanisms that affect the learning achievement of student groups. However, we need to keep in mind that the reality is much more dynamic.

3.5 SummaryTable 3.4 summarizes the essential elements of the methodological framework that guides the research activities that are discussed in this thesis. The human activity system that is the object of study is the group that consists of a limited number of students who communicate face-to-face. These students participate in a problem-solving discussion with the aim to develop an increased understanding of the topic of discussion.

51

Chapter 3

We adopt a Systems perspective that differentiates between three levels of complexity: 1) the individual learner, 2) the interacting group, and 3) the larger learning environment of the classroom. Each level of complexity can be described by its own emergent properties.

Systems thinking considers a group of students as a purposeful system. This opens up explanations in terms of coordinationprocesses and function. System thinking is associated with a functional perspective that focuses on performance differences and relates these differences to specific patterns of group communication. In our case, we aim to understand the functions of communication relevant for the learning task and the various constraints that might prevent groups from effective communication (Waldeck et al., 2002).

The functional perspective is in accordance with a Design-based research approach that intervenes within an existing learning situation to create desired patterns of communication and learning outcomes. These interventions should not only solve the problems that a group faces but they must also lead to an increased understanding about communication, mediation and learning of groups who interact face-to-face.

52

Characteristics Description

Ontological claims

A systems perspective towards social reality.Groups can be described at different levels of complexity.Each level of complexity is characterized by its own emergent properties.

Explanatory framework

A functional perspective towards group performance.Communication processes will be explained in terms of organization and function.

Functional theories focus on performance differences. These performance differences can be traced back to communication patterns that emerge during a problem-solving discussion. Some patterns are positively associated with learning while other patterns hamper learning.

Mode of inquiry Design-based research.

Table 3.4: Research strategy to study collaborative learning in the classroom.

4 Group Discussions: Learning as Co-

Construction

Without other people with whom we share responses to a mutual environment, there is no answer to the question what is in the

world to which we are responding

Donald Davidson1

Collaborative learning makes specific epistemological claims about the nature of knowledge and learning. In this thesis, collaborative learning is viewed from a social-constructivist perspective that stresses the “interdependence of social and individual processes in the co-construction of knowledge” (Palincsar, 1998). Social constructivism acknowledges that the construction of knowledge emerges on the social level without denying the constitutive role of the individual. It enhances the role of inter-action with others rather than actions themselves (Dillenbourg, Baker, Blaye & O'Malley, 1996).

Social-constructivist epistemology contains a clear view of how collaborative learning should proceed in practice. Several studies within this tradition examined what kind of communication contributes to the co-construction of knowledge. These studies make clear that the quality of interaction has implications for learning (Barron, 2003). In this chapter, we further elaborate on the term “quality” and identify some communication patterns that can be associated with the co-construction of knowledge. We present a set of criteria for a constructive dialogue. These criteria serve as communication demands that indicate how well a group carries out a problem-solving discussion aimed at achieving an increased understanding.

1 Davidson, D. (1999). The emergence of thought. Erkenntnis, 51, 7-17.

53

Chapter 4

4.1 Constructivist Views towards LearningConstructivist views of knowledge and learning mark a shift in focus from the object of experience or the known towards the subject of experience or the knower (Campbell, 2002). These views state that knowledge is constructed internally by the individual or socially during interactions with others. Constructivist theories come in many shapes like cognitive constructivism, radical constructivism, social constructivism, and sociocultural theory. Descriptions of these theories are open to divergent interpretations and the use of concepts varies. It is not easy to make a clear distinction between the various theoretical directions that fall back on the constructivist epistemology. The readers of constructivist literature are usually left to figure out for themselves which epistemological direction is being pursued (Phillips, 1995). One could argue that this is inherent to the position of these theories because they state that knowledge is constructed through “subjective or even intersubjective experience” (Campbell, 2002). Still, it is important to make clear which position will be adopted in this thesis because it has strong implications for research.

Subject-centered versus Social Accounts of Constructivism

Constructivist theories can be positioned on a continuum with subject-centered accounts at the one end and social accounts at the other (Davis & Sumara, 2002). The distinction has to do with the different epistemological claims about the nature of knowledge. Subject-centered accounts stress that individuals construct their own understanding; they include accounts such as cognitive theories and radical constructivism. Social accounts of constructivism situate the process of learning in the social domain and consider knowledge as something that emerges out of social interaction or social activity. Learning from this perspective is viewed as inherently social; people construct “shared versions of knowledge” (Burr, 2003) when they try to make sense of their mutual experiences. Social accounts include theoretical positions like social constructivism and sociocultural theory.

Subject-centered Constructivism

Subject-centered constructivism states that the mind does not passively accept sensory impressions; rather the mind actively imposes an interpretive framework on sense data. Reality, in other words, is constructed (Driscoll, 1994). Virtually all cognitive theories, which are inherently subject-centered, entail some form of constructivism to the extent

54

Group Discussions: Learning as Co-Construction

that the cognitive structures are typically viewed as individually constructed in a process of interpreting experiences in particular contexts (Palincsar, 1998). These cognitive theories take up a realist position; they see the internal representations that are the result of individual constructions as more or less correct representations of an external world. The challenge of teaching, from that point of view, is to present the learning material so that it allows for effective and accurate constructions of internal representations. It means that the teacher introduces the subject matter in an accessible manner that closely matches the cognitive abilities of the individual student, taking into account how people acquire and organize information cognitively. The teacher presents the knowledge in parts that are easily understood, points out connections and explains what is difficult to understand. Reiteration, multiple representations and referring to existing knowledge are seen as effective strategies for knowledge acquisition.

Radical constructivism can also be positioned at the subject-centered end of the continuum. However, radical constructivism takes up an idealist position towards the nature of knowledge. It denies that knowledge could be treated as internal representations of the external world. Radical constructivism breaks with the traditional theory of knowledge that asserts that our cognitive efforts result in a more or less objective representation of the world as it exists apart from us and our experiences (von Glasersfeld, 1991). Radical constructivism does not deny an objective world but this world is not accessible for the individual mind. Radical constructivism refers to the work of Piaget who sees learning as a process of adapting one’s internal believe system triggered by individual experiences. The constructivist epistemology of Piaget’s work refers to a process in which the individual reflects on and organizes experiences to create order in and adapt to the environment (De Lisi and Golbeck, 1999). Von Glasersfeld (1996) followed Piaget and stated that:

Knowledge, for [Piaget], arises from actions and the agent’s reflection on them. The actions take place in an environment and are grounded in and directed at objects that constitute the organism’s experiental world, not things in themselves that have an independent existence (von Glasersfeld, 1996, p. 4).

Social Accounts of Constructivism

Social accounts of constructivism also consider experiences rather than an objective source outside these experiences as the reference for knowledge construction. However,

55

Chapter 4

social accounts do not set the individual apart from the social world as radical constructivism does. The subjectivity of the individual and the objectivity of the surrounding social and cultural context exist in relation to one another (Davis & Sumara, 2002). Properties of social collectives like language constitute to the construction of knowledge. Social accounts of constructivism include theoretical positions like social constructivism and sociocultural theory.

Social constructivism covers a range of learning theories that share an antireductionist stance towards the nature of knowledge. Human learning is not a solitary activity by individuals who make sense of their world. Knowledge is temporary, developmental, non-objective, internally constructed and socially and culturally mediated (Twomey Fosnot, 1996). Social constructivism stresses the social dimension of cognition and the role that language production plays in promoting learning (Palincsar, 1998).

Two terms with more or less similar connotation appear in the literature: social constructivism and social constructionism. The former term is more frequently used in educational sciences. Social constructivism refers to the idea that knowledge can be situated on the social plane. The latter term is more widely used in psychology to indicate a tradition that can be contrasted with the reductionist tradition in psychology. Social constructionism places the social prior to the individual and acknowledges the constructive power of language (Gergen, 1995, Burr, 2003). We use the term social constructivism because our research takes place in an education setting. Social constructivism assumes that sense making is a social process of people who interpret mutual experiences. The units of analysis to study learning are: the social activity, the mutual environment that shapes thinking, the interactions between people, or a combination of these units. Some analyses concentrate on micro descriptions of human interaction, while others focus on the cultural and historical origins of knowledge.

Sociocultural theory sees human learning through the manner in which the social and cultural world codetermine the way in which people approach learning in various settings, inside and outside formal institutions (Bliss and Säljö, 1999). Thinking is culturally mediated by artifacts such as signs and tools, it is founded in purposive activity, and it develops historically (Scribner, 1997; Packer & Goicoechea, 2000). The defining concepts of sociocultural theory – human action, the use of tools and mediation – can be traced back to the work of Vygotsky. All activities contain different artifacts, and these artifacts or tools embody a certain history and culture (Vygotsky,

56

Group Discussions: Learning as Co-Construction

1986). Artifacts play an essential role in shaping action; they mediate human action (Wertsch, del Rio & Alvarez, 1995). Every mediating system has distinct features that characterize the nature of communication and learning that takes place in that system. It means that human experience is shaped by the artifacts that they use (Nardi, 1996).

Taking Up a Position

We follow a social-constructivist perspective that assumes that learners construct knowledge through social interaction and that the nature of these interactions affects collaboration and learning.

Social-constructivist theories can be positioned on both sides of the realist-idealist continuum, most theories however firmly lean towards the idealist end. Idealist accounts state that learners cannot know reality in itself, only in so far as it is given in consciousness, experience, language, or practice (Collier, 1998). However, some social-constructivist theories maintain some concept of reality that exists outside the discourse (Burr, 2003). Zuriff (1998) made a distinction between empirical and metaphysical social constructivism. These two positions can be situated at the two ends of the Realist-Idealist continuum. Empirical social constructivism distinguishes the natural world from the constructed world and admits that constructions are descriptions of the natural world. Metaphysical social constructivism rejects the view that the natural world consists of an external objective reality, independent of the human mind (Zuriff, 1998).

With regard to our position we keep away from a relativist position that suggests that essential properties of an object of the world are relative to their description (Moser, 1993). We do acknowledge that human action is based on a belief system that is socially constructed. This does not mean, however, that there is no reality outside these experiences. Although the perceptions and sensations of people do not mirror reality, they do refer to the real world in some way; they are not independent of it, produced entirely through symbolic systems such as language (Burr, 2003). Experience is always an encounter with what exist before the experience and is to a degree known to us as a result of the experience (Collier, 1998). This observation makes it possible to come up with a deliberate design that aims to change an existing practice. The design interventions are based on insights of the practice but it also incorporates knowledge in a broader sense.

57

Chapter 4

4.2 Collaborative Learning in the ClassroomSo far, we have discussed social constructivism at a rather abstract level. In this paragraph, we turn to a more practical elaboration. We discuss the implications of social constructivism by focusing on the human-activity system – i.e. small groups of interacting students who learn collaboratively – that is the object of the research. Collaborative learning has been implemented as a problem-solving discussion between a limited number of students. Collaborative problem solving can be seen as the interplay between a problem-solving task and those who have to solve that task. It is defined by the interaction between task characteristics on the one hand and person and group characteristics on the other (see for example Frensch & Funke, 1995). In this paragraph, we further elaborate on the relationship between the two. We introduce the notion of “problem space” to conceptualize the relationship between the group and the learning task. The concept has its origin in the cognitive analysis of human problem solving where it is associated with an information-processing approach. It has also been applied to describe problem solving from a social-constructivist perspective (Roschelle & Teasley, 1995; Greeno, 1998).

Information Processing: A Search for the Right Answer

The information-processing perspective focuses on distinguishable aspects of human problem solving. The complexity of human problem solving is decomposed into identifiable processes that are carefully studied by manipulating various conditions within the environment. Research topics include knowledge and its representation, inferential processes and perception. Explanations are primary in the form of models of processes of constructing, storing, retrieving and modifying representations of information (Greeno, 1998).

Problem solving from an information-processing perspective is seen as a search for the appropriate response within a set of internal representations that comprise the individual's problem space. A problem space is defined as a person’s internal representation of the task environment (Newell & Simon, 1972). It captures essential aspects of the problem as it is presented to the person. The problem space includes general knowledge associated with the problem and strategies, plans and rules for solving that problem.

58

Group Discussions: Learning as Co-Construction

The Pooling of Information

The information-processing perspective conceptualizes collaborative problem solving as a joint search in the problem space for the right means to solve the problem. Groups from this perspective are viewed as information-processing units, analogous to how cognitive psychology views individuals (Brauner & Scholl, 2000). Group members share information so that they are able to build a more accurate description of the problem. They deal with a lack of understanding by constructing a shared representation of the problem. The essence of collaborative problem solving is that group members pool information to represent the problem space. The group from this perspective needs to gather all the information needed to solve a problem and it uses the collected information to develop solutions (Hasenbein, Kopp, Mandl, 2008). Group problem solving, just like individual problem solving, is a search for the right answer within the pool of collective information. According to this view, groups outperform individuals because they have potentially more information available, are more likely to recognize valid information, and are capable to process more information than an individual (Propp, 1999).

Information Processing and the Design of Group Support Systems

Electronic meeting systems (EMS) seem to reflect an information-processing perspective towards collaborative problem solving. These systems are used in professional settings to support groups who have to solve problems or make decisions. They are employed in face-to-face settings where people gather at the same time and same place (Fjermestad, 2004). A group uses an EMS to construct a shared representation of a problem by exchanging, organizing and assessing information that is available among the group members. In general, such meetings can be characterized by a distinct communication sequence of 1) the generation of ideas, 2) the organization of these ideas into meaningful categories, and 3) the evaluation of these ideas (Stefik, Foster, Bobrow, Kahn, Lanning & Suchman, 1987).

Conclusion

The added value of groups from the information-processing perspective is their ability to better process information. This notion seems too narrow for our situation. Complex learning does not only capitalize on the pooling of information. Complex learning such as solving ill-structured problems calls for the construction of new

59

Chapter 4

knowledge (King, 1999).

Ill-structured problems give rise to confusion: for example, the exact nature of the problem and possible solutions are unclear because information can be interpreted in different ways. These differences serve as the mechanism for learning. Students can only resolve ill-structured problems by extensive reasoning during which the students address their differences in thinking. Students should be encouraged to critically address each other's belief system and discuss underlying reasons, assumptions, and perspectives. They have to engage in a constructive dialogue that marks a rather subtle shift from the pooling of information and the search for the right answer towards interaction and the joint construction of meaning. It emphasizes that collective thinking is much more than the pooling of knowledge (Barron, 2003).

Social Constructivism: The Emergence of Understanding

Social constructivism considers collaborative problem solving as the intelligent social practice of a group of students who collaborate on a common task. The denomination ‘intelligent’ refers to the ability of the group to alter their thinking in response to additional information, growing experience or increased insight. This ability cannot be traced back solely to individual cognition but emerges during group interaction when the group constructs a shared representation of the problem space that exceeds their prior individual understanding. Shared understanding cannot be established at the beginning; rather it evolves when the group jointly addresses the demands set by the task.

Shared UnderstandingA shared problem space expresses how the group understands the problem. Aspects of the problem situation that are shared may include a representation of the situation, the main goal, operators for changing the situation, and strategies, plans and knowledge of general properties and relations in the domain (Greeno, 1998). A shared problem space supports problem-solving activity by integrating goals, descriptions of the current problem state, awareness of available problem solving actions, and associations that relate goals, features of the current problem state and available actions (Roschelle & Teasley, 1995).

A shared problem space emerges as individuals find and align themselves with other members who have comparable cognitive models of the situation (Massey &

60

Group Discussions: Learning as Co-Construction

Wallace, 1996). Significant features of the problem space arise during group interaction. This alignment of cognitive models becomes more apparent for ill-structured problems that are open to multiple interpretations. The group can be considered as a means to gain access to alternative viewpoints. Groups do not only have more resources available to generate new ideas. There are also more viewpoints from which to evaluate critically those ideas (Fisher, 1980). Multiple viewpoints serve the same role as multiple examples or cases in supporting the induction of abstractions (Gick and Holyoak, 1983). Through a process of extensive reasoning group members integrate their perspectives into a more abstract, and a more accurate representation of the problem space (Schwartz 1995). Roschelle (1992) speaks in this respect of convergence that is achieved through cycles of displaying, confirming and repairing shared meanings. Individual and shared cognitive representations recursively co-evolve into a shared problem space that reflects shared knowledge and conceivably new knowledge (Massey & Wallace, 1996).

Occasions for Collaborative Problem Solving

According to Weick (1995) two aspects of the problem-solving task – uncertainty and ambiguity – determine the kind of communication that is needed to solve a problem. Uncertainty refers to those situations where the problem solvers lack sufficient knowledge and information to form a valid representation of the problem. Uncertainty comes from an ignorance or imprecision of an interpretation of the problem and possible solutions. To remove ignorance, more information is required (Weick, 1995).

Ambiguity, in contrast, refers to those situations where people are confronted with multiple viewpoints. The problem with ambiguity is that people are unsure what questions to ask and whether there even exist a problem they have to solve (Weick, 1995). People do not need new information to address ambiguity but they have to enter into a process of reasoning, argumentation and negotiation. Schrage (1990) made a similar remark when he states that the essence of collaborative problem solving is not just more communication but rather a different quality of interactions. Schrage (1990) identified two kinds of groups, those who are oriented at communication and those that are oriented at collaboration. The traditional model of communication states that listening carefully and talking clearly are essential for understanding. Groups caught up in the communication paradigm believe that more communication can compensate for a lack of understanding. Collaboration, on the other hand, assumes a different kind of

61

Chapter 4

orientation; what is needed is not more communication but rather a different quality of interaction. Students should be less interested in displaying information rather than in creating a shared space that enables them to play collectively with their knowledge, concepts and beliefs (Schrage, 1990).

4.3 A Constructive DialogueThe collaborative learning practices that are the object of the research exploit ambiguity as an opportunity for learning. The ambiguity associated with the learning task leads to a constructive dialogue during which the students address their differences in thinking. A constructive dialogue is concerned with the knowledge and the concepts that underlie the problem solving (De Vries, Lund, Baker, 2002). It uses inquiry to explore one another's assumptions and thinking with the intent of learning about them more deeply (Barge, 2002).

Collaborative learning as a constructive dialogue puts some demands on the kind of communication that the group has to display. A constructive dialogue requires the active involvement of the learners’ believe system, while better understanding is acquired when learners collaboratively reflect on their mutual experiences. Learners should align their comparable cognitive models to construct a meaningful representation of the problem space. Then, the emphasis lies on communication that creates meaning while coherence is the primary vehicle through which learning occurs (Allen & Plax, 2002).

Box 4.1 gives an overview of the communication demands that are associated with a constructive dialogue. First, the communication should be oriented at effective task performance because learning achievement is associated with task-related interactions. Task-related interactions stimulate the elaboration of conceptual knowledge (van Boxtel, van der Linden & Kanselaar, 2000).

A second criterion for the success of the groups is that the members maintain acceptable levels of participation. All the group members must be able to share their knowledge with the group. Equal participation of the group members is a fundamental element of well-performing student groups (Lindblom-Ylänne, Pihlajamäki & Kotkas, 2003).

Thirdly, the group members must organize their individual talk into a coherent

62

Group Discussions: Learning as Co-Construction

whole. If a contribution is to be used, someone must, sooner or later, relate it to the other contributions (Harnack, Fest & Schindler Jones, 1977). Barron (2003) studied the interaction patterns of groups who communicated orally in a face-to-face setting. The groups consisted of three students who participated in a problem-solving discussion. She found that more successful groups compared to less successful ones differed in how they respond to correct proposals. Successful groups carried out a productive discussion that is oriented at the exploration of ideas. These groups discussed or accepted correct proposals, whereas less successful groups showed a tendency to ignore or to reject them. Students from successful groups also showed a better transfer of their learning to an individual achievement task. Differences in level of knowledge of individual group members did not account for how successful the groups were. Barron found that the performance differences had to do with the interaction patterns. Successful groups carried out a more coherent discussion during which they directly linked proposals to the prior conversation. This observation is confirmed by a study of Kneser & Ploetzner (2001) who conclude that the successful groups produced more coherent dialogues.

Finally, successful groups can be characterized by frequent knowledge elaborations that are organized in an orderly and meaningful manner (Barron, 2003; Kneser & Ploetzner, 2001). Hogan, Nastasi and Pressley (2000), for example, identified three patterns of interaction to characterize the essence and flow of a discussion: 1) consensual, 2) responsive, and 3) elaborative. During consensual sequences, only one speaker makes substantive statements. Another speaker assents or acquiesces with what

63

Box 4.1: Communication patterns of a constructive dialogue.

– Communication is oriented at the learning task.

– Communication displays an equal pattern of participation and the free exchange of ideas.

– Communication is coherent. The successive communicative exchanges of various students should be organized in an orderly and meaningful way.

– Students go beyond the given and address their group member's assumptions, judgment and reasoning strategies. Such a group discussion can be characterized by a process of elaboration on the knowledge that is shared by the group members. Elaborative communication patterns can be characterized by a critical assessment of the knowledge that is shared by the group members.

Chapter 4

has been said. Within responsive sequences, at least two speakers contribute substantively to the discussion. These sequences usually have a question-response pattern that contains a few turns in length. Elaborative sequences resemble coherent sequences in the sense that several speakers contribute substantive statements to the discussion. The speaker in elaborative sequences makes multiple contributions that build on or clarify another’s prior statement. Hogan, Nastasi & Pressley (2000) observed that performance differences could be traced back to the kind of interactions that emerged during the discussion. In general, the elaborative sequence was most frequently associated with a productive dialogue.

A medium-neutral description

In the next chapter, where we discuss the problem analyses and the tool design whereby we take the four criteria of Box 4.1 as the point of reference. These criteria – task orientation, equal participation, coherence and knowledge elaboration – have been described in medium-neutral terms. The criteria relate the group communication with learning achievement. They will be used to analyze the existing verbal, face-to-face communication but the criteria also serve as the reference for the design of the collaborative tools. The aim of the collaborative tools is to create patterns of communication that closely match with the criteria for a constructive dialogue.

64

5 The Learning Environment of the

Classroom

The transition from Aristotelian to Galilean concepts demands that we no longer seek the “cause” of events in the nature of a

single isolated object, but in the relationship between an object and its surroundings.

Kurt Lewin1

In chapter 2, we already mentioned that we carried out the research in the classroom. In fact, the research environment within which we studied collaborative learning was also an authentic setting for learning. The students that participated in the research were not experimental subjects in the classical sense; in the first place, they were students who came to the classroom to learn. It means that the interventions that are discussed in this thesis are no isolated events in a carefully controlled setting; they can only be understood by taking into account the different elements that comprise the learning environment of the classroom. In contrast to an experimental approach, we did not neutralize the influence of these elements. Therefore, before we discuss the collaborative tools, we give an overview on the learning environment that incorporates these tools. The learning environment will be discussed from the perspective of two learning methods: direct instruction and collaborative learning.

5.1 Two Learning MethodsWe argue that the way the learning environment will look like in practice depends on the learning method that is adopted by the teacher. To make this clear we refer throughout this chapter to two learning methods: direct instruction and collaborative

1 Kurt Lewin (1936), Principles of topological psychology.

65

Chapter 5

learning.

We choose to describe direct instruction because it is a common approach for teaching (Killen, 2007). This method will be contrasted with collaborative learning that recently appeared in the classroom as a “new learning” method. Collaborative learning is frequently used as an alternative learning method that undoes some of the disadvantages associated with direct instruction. In practice, we observed that two methods frequently mingled. The students that participated in the study discussed in chapter 9, for example, worked together in small groups to discuss a topic after a series of lectures. The small-group activities encouraged students to apply the knowledge acquired during the lectures in an authentic learning situation that resembles real-life practice.

Direct Instruction

Direct instruction has been shown to be an efficient way to teach factual knowledge or procedures that are difficult for students to discover on their own (Palinscar, 1998; Klahr & Nigam, 2004). Direct instruction is a teacher-centered approach: the teacher directs the learning activities that go on in the classroom. The teacher fully explains the concepts and procedures that students are required to learn and defines the learning strategies for acquiring the knowledge (Kirschner, Sweller & Clark, 2006). The emphasis is on teaching in small steps, providing students with practical assignments after each step, guiding students during initial practice, and providing all students with a high level of successful practice (Rosenshine, 1987).

Direct instruction can be typified by a teacher-centered communication pattern: communication follows a pervasive pattern of teacher initiation, student response, and teacher evaluation or feedback (Mehan, 1979). For example, the teacher asks a question, a student answers and the teacher evaluates that answer. These questions are used to determine prior knowledge or to monitor student’s progress. The teacher uses the student’s response to fine-tune the teaching practice so that it closely matches with the abilities of the student.

The spatial arrangement of the classroom supports the asymmetrical communication pattern illustrative for direct instruction; usually the teacher stands or sits in front of the class while the students are seated with their faces oriented towards the teacher, which makes communication with the teacher quite easy.

66

The Learning Environment of theClassroom

Collaborative Learning

Collaboration between learners may be part of a teacher-centered learning method like direct instruction. For example, a discussion between students helps them to better process new information. Characteristic for these discussions is the central role for the teacher. The teacher guides the discussion by bringing forward topics, asking questions, and giving additional information. Collaborative learning, in this case, rests on cognitive theories of learning. It can be contrasted with the social-constructivist perspective that is emphasized in this thesis. The latter approach can be characterized by a changing relationship between the teacher and the learners. It marks a shift in responsibility from the teacher towards the learners. Learners operate more autonomously; to a large extent they are responsible for their own learning activities. The students organize their own learning and monitor progress, often in collaboration with their peers. It is assumed that the capacity to learn increases when the students have greater autonomy (Brockbank and McGill, 2000). It gives students the opportunity to construct their own perceptions (Driscoll, 1994).

Collaborative learning is a method where students work together in groups that are small enough so that everyone can participate on a collective task that has been clearly assigned (Cohen, 1994). It stresses that students construct knowledge in interaction with fellow students with the aim to make sense of their mutual learning experiences. Students draw upon their own knowledge, beliefs and judgments, share these concepts with their group members, and explore their assumptions and consequences.

The spatial arrangement of the classroom should differ when students learn in small groups, although the classroom does not always allow for an optimal layout. Several small groups sit together in the classroom, more or less separately from the other groups, with their faces directed towards each other. This guarantees free interaction between the group members without much disturbance from outside. The teacher remains more at a distance and monitors the learning activities from the various groups on the fly. The teacher walks around, goes from one group to another, paying attention to a group when this is needed. Exactly this situation is the focus of the research that is discussed in this thesis.

67

Chapter 5

5.2 The Classroom Learning EnvironmentThe learning environment of the classroom mainly provides interpretations of the natural world for students to interact with (Winn & Windschitl, 2001). These interpretations are brought to the learner by the teacher, their fellow learners or by different kinds of media like books or computers. These three elements of the learning environment represent the knowledge domain to the learner (Kanselaar, de Jong, Andriessen & Goodyear, 2000). Learning can be considered as an upshot of the learner’s interaction with these three elements (Figure 5.1).

Resources

The three elements - teacher, media and fellow learners - draw upon different kinds of resources to direct the interactions with the learner. These resources are used to represent the knowledge domain to the learner. They are for the most part deliberately arranged to create the optimal conditions for learning. The three elements and the different resources they call into play lead to divergent interaction patterns and distinct opportunities for learning. The interactions of the learner with the teacher, for example, differ in character from the interactions that occur among the learners. The relationship between the teacher and the learners can be characterized as one between

68

Figure 5.1: The learning environment of the classroom.

The Learning Environment of theClassroom

an expert and a novice. The teacher who is an expert in the knowledge domain under study, employs specific instructional strategies and teaching skills to present the subject matter to the learners. Learners who explore a topic together usually do not possess that kind of knowledge and skills. They interact as equal interlocutors with more or less similar conceptual models. Collaborative learning capitalizes on these interactions. When learners interact with each other, they engage in a “progressive discourse” (Wells & Arauz, 2006) during which they explore each other's understanding of things. Such a discourse must be of a high cognitive level, that is, it needs to include the mutual exchange of ideas, explanation, justifications, speculations, inferences, hypotheses, and conclusions (King, Staffieri & Adelgais, 1998).

Reciprocal Relationship between the Learner and the Three Elements

The three elements of the environment do not affect learning in a direct way. Students do not merely reproduce what is communicated but they actively integrate new information within their cognitive framework. It results in new knowledge and increased understanding based on what they already know and believe (Bransford, Brown & Cocking, 2000). It is through these mutual constitutive interactions that learning gets its shape: the teacher, media and fellow learners share their interpretations of the knowledge domain but the learner also brings in his or her own prior knowledge, feelings, motives to make sense of what is communicated.

Reciprocal Relationship between the Three Elements

The three elements that comprise the learning environment, and the interactions they give rise to, do not operate independently from each other. For example, they way that the teacher communicates with the class affects the way learners communicate with each other, while the opposite is also true. For example, the teacher may apply a specific instructional strategy to stimulate specific patterns of communication between the students. When the teacher asks questions that have multiple possible answers and stimulates learners to build upon each other’s contribution, it is more likely that a productive discussion between learners will occur (Wells & Arauz, 2006).

69

Chapter 5

5.3 The TeacherLearning is not solely an individual activity; it is also the result of the learner’s interactions with other people. The person who comes first to mind is the teacher. Classroom learning is usually centered around the teacher who serves as an important point of reference. The teacher performs various tasks: the teacher has to instruct students effectively, deal with them as a group and respond to individual needs, establish and maintain order in the class, and handle the discipline and adjustment of individual students (Emmer & Stough, 2001). The way the teacher deals with these tasks influences the motivation and achievement of the students. Research indicates that the appropriate teacher-student relationships can be characterized by a high rate of teacher influence and cooperative behavior towards the students (Wubbels & Brekelmans, 2005).

Teacher-student interaction does not proceed in a one-way direction. Students bring in their own thoughts, feelings and motives towards the learning activities that go on in the classroom. The teacher has to address the cognitive, as well as the affective and behavioral dimensions of students’ performance. Both learning methods – direct instruction and collaborative learning – recognize the active role of the learner within the teacher-learner relationship. However, their practical implications differ.

With Direct instruction, the teacher regulates the learning process although the teacher has to take into account prior knowledge of the students. If their initial understanding is not engaged, students may fail to grasp the new concepts and information that are taught (Bransford, Brown, Cocking, 2000). The teacher has to present the subject matter in a way that best fits with the abilities of the learners. Information presented by the teacher should be relevant, meaningful and, if necessary, useful to the learners (Ebbens & Ettekoven, 2005).

With collaborative learning, the students have a more active role. To a large extend they shape their own learning. The students explore a subject matter collaboratively often without close guidance from the teacher. Fellow students serve as an important reference to mirror student's belief system. The role of the teacher shifts from an expert who controls the learning processes, towards a facilitator who creates the proper conditions under which the group may learn from and with each other.

70

The Learning Environment of theClassroom

The teacher draws upon various resources to direct the teacher-student interactions. The teacher shapes learning by means of:

1. the learning task,

2. the instructional strategy,

3. the teaching skills, and

4. the communication tools that support the interactions between the teacher and the learners.

The Learning Task

A learning task exists of an assignment allocated to an individual or a group of learners to create a specific opportunity for learning. It is a way to engage students in the teaching practice. When students carry out a learning task, they are actively involved instead of listening to a teacher passively. A learning task may serve several educational goals. It can be used to motivate students, to strengthen knowledge and skills, for transfer of learning, or to engage students actively in the construction of knowledge.

Problem Solving Task

In this thesis, we concentrate on problem solving as the means by which the students develop a deeper understanding of a subject matter. Problems are commonly defined as a gap between a disruptive situation and a desired end state that can be bridged by the activities of one or more human actors. Two additional conditions are essential for problem solving to occur. First, the human actors must perceive the problem-solving task to be important enough to inspire current and prospective solution activities (Smith, 1988). This condition seems trivial but in an educational setting it cannot be taken for granted. It has a strong motivational connotation; students must perceive the problem-solving task as meaningful so that they are willing to participate in the problem-solving activities.

Active participation can be further stimulated by establishing interdependence between group members. Cohen (1994) identified three structures of interdependence that foster participation:

1. positive goal interdependence,

2. positive resource interdependence, and

3. reward interdependence.

71

Chapter 5

Collaboration will be stimulated when students depend on one another to achieve the group goal, when they have to use one another’s resources to attain that goal, and when they receive a group reward that is based on the performance of each individual member (Cohen, 1994). The three structures of interdependence do not affect participation directly. Their effect depends on factors such as task characteristics and motivation. Together they determine what kinds of interactions do occur. Reward interdependence, for example, does not appear to be necessary for achievement when students are motivated to complete a challenging and interesting group task that requires everyone’s contribution for a good outcome (Cohen, 1994).

A second condition for students to engage in problem-solving activities has to do with the gap between the disruptive situation and the desired end state. It must be difficult but not impossible for students to bridge the gap between the disruptive situation and the desired situation. Not all situations that require purposeful actions of a human actor can be labeled as problematic (Smith, 1988). Only when a solution is not ready at hand or the steps to bridge the gap are not clear, the learners must allocate energy for problem solving.

Well-structured and Ill-structured Problems

Problems can be categorized along various dimensions like the kind of knowledge required to solve them or their complexity (see Robertson, 2001). A frequently used distinction takes the complexity of problems as the starting point: problems can be situated on a continuum that goes from well-defined to ill-defined problems.

Well-defined or well-structured problems provide the students with sufficient information that enables them to solve the problem. The problem description contains information about the problem as it stands now (initial state), what the situation should be when the problem is solved (the goal state), and what actions should be taken to solve it. It does not mean that the students have all the necessary information, as they are not told what objects to perform the action on (Robertson, 2001). Robertson (2001) gives the example of algebraic problems. There are four basic arithmetic operations: multiplication, division, addition and subtraction. Students usually know these operations; it is more difficult for students to know how to apply these operations.

Wyndhamn and Saljö (1999) studied collaborative problem solving for a well-defined problem where students had to count the days between a certain calendar

72

The Learning Environment of theClassroom

period. They provided the students with a clear description of the initial state and goal state. The conditions that they created varied between the kind of information that makes students aware of the strategy to solve the problem. The groups, who received the proper information through the assistance of an adult or an artifact like a calendar, were able to solve the problem. If the groups lack the proper information, they were not able to solve the problem correctly. When the students discussed the problem, they became aware of their lack of knowledge and they were only able to solve the problem when an outsider provided them with the adequate information. The task that the students had to solve can be characterized as uncertain: students lack the proper information and they have to search for additional information.

With ill-defined or ill-structured problems, the students do not know in advance the conditions that apply to the problem-solving process and the actions that lead to a solution of the problem. Both the given state, the desired end state, and the barriers are transparent (Frensch & Funke, 1995). These novel problems require considerable knowledge or experience to solve them. They involve complicated situations or statements that are unfamiliar and difficult to understand for the students. Insights with regard to the exact nature of the problem and possible solutions emerge during the problem-solving process and constitute each other, i.e. specifying the problem goes hand in hand with solving it.

A number of studies into collaborative learning focused on novel problems. These studies stem from different knowledge domains such as biology (Coleman, 1998), physics (Roschelle & Teasley, 1995; Baker, 2002; Hogan, Nastasi & Pressley, 2000), mechanics (Kneser & Ploetzner, 2001), the study of a text (Keefer, Zeitz & Resnick, 2000), or mathematics (Sfard & Kieran, 2001).

Hogan, Nastasi and Pressley (2000), for example, studied a learning activity where students had to understand the characteristics and behaviors of solids, liquids, and gases. These students collaborated in groups to build a mental model of the nature of matter and used their model to explain the characteristics of solids, liquids, and gases. Before they entered into a discussion, they gained experience with the three states of matter by means of experiments and demonstrations.

Complex, novel problems may give rise to different interpretations and misunderstandings that can only be resolved by extensive scientific reasoning. According to Schwartz (1995), the different perspectives held by the learners create

73

Chapter 5

opportunities for higher-order thinking. When learners address their differences in thinking, they develop a deeper understanding of the problem. The process of participating in arguments or even listening to others arguing and justifying their opinions or solutions may be enough to enhance learning (Slavin, Hurley & Chamberlain, 2003).

The research that is discussed in this thesis focuses on ill-structured problems. These problems are complicated, multidimensional and relatively novel for the learners. They give rise to multiple interpretations and must be resolved by extensive inquiry and reasoning.

Authentic Tasks

Learning from a social-constructivist point of view requires the active involvement of the learner’s believe system, while better understanding is acquired through meaningful experiences. One way to engage students with meaningful experiences is to present them with a learning task that incorporates authentic activities based on real-life situations. Romiszowski (1984) defined different kinds of tasks that use reality as the basis for learning (Figure 5.2). Although each category uses real-life data, there is a shift in pedagogical goals and, subsequently, a shift in the representation of the real life situations.

The studies that are discussed in this thesis used “case study” and “role play” as authentic learning task. These studies examine the collaborative activities of students

74

Figure 5.2: Different kind of tasks that are based on ‘real life’ situations.

The Learning Environment of theClassroom

who discussed a topic whereby they draw upon a case description or they approach the topic from different roles.

A case describes a problem, situation or event from an actual context (Lkoundi & van Woerden, 1997). It presents information about a situation or a process-in-action, for analysis and discussion by a student or by a group (Romiszowski, Mulder & Pieters, 1990). Although a case is based on authentic data, it is not an exact description of a real problem, situation or event. A case differs from the latter because it takes into account the learning processes and goals. For example, the complexity of the situation or the task may be “simplified” for reasons of coherence and intelligibility.

In a role-play, students take a role and act them out in a play. A role play represents a relevant social situation in such a way that the focus is on the perspectives and behaviors of the various roles and the relationships between the roles. Role play provides students with a framework to analyze a situation and to develop ways of coping with that situation (Joyce and Weil, 1980). When students take a particular role, they use a repertoire of behaviors that are expected of the role in that situation (Holsbrink-Engels, 1998).

Instructional Strategies

Another resource that the teacher draws upon is an instructional strategy. An instructional strategy specifies how the students learn effectively. It consists of a set of guidelines or a plan that directs student's learning towards the achievement of the learning goals. An instructional strategy takes the form of a lesson plan or a set of product specifications for mediated materials (Gagné, Briggs & Wager, 1992). They reflect a number of decisions about the arrangement of the learning environment with the aim to create the proper conditions for learning.

In the case of direct instruction, the teacher clearly instructs students what they have to do. The teacher explains to the students: 1) the assignment, 2) how to carry it out, 3) what kind of aids they have at one’s disposal, 4) how much time they may use, 5) what will happen with the outcomes, and 6) what they should do when they finish the assignment (Ebbens & Ettekoven, 2005).

For collaborative learning, the picture is much more diverse: students may be given more freedom to arrange their learning themselves. Instructions help students to construct their own meaningful and conceptually functional representations (Jonassen,

75

Chapter 5

1991). For example, the teacher only presents the assignment, while it is up to the students to determine how to carry it out and what kind of aids they want to use. The students are responsible for defining their learning outcomes and for choosing the road needed to achieve these outcomes (Reigeluth & Moore, 1999).

Instructions as the Link between the Technology and the Learners

In this thesis, we focus on the instructions that make the collaborative tools applicable for a certain context of use. While the tools stay more or less the same over different contexts, its use will vary. Groups can use a collaborative tool in many ways that can be more or less appropriate to the group’s purpose and practices. Just introducing the tools in the classroom is not sufficient. In our case, the introduction of the collaborative tools is accompanied with an instructional strategy that makes the support suitable for a specific learning situation. The use of instructions results in a support environment that is “optimally adapted to the microstructure of local conditions and constraints” (Lea, 1994). The instructional strategy consists of a set of procedures and guidelines that enable the teacher to adapt the support environment to a specific learning practice. It helps the teacher to stimulate communication and behaviors that, in our case, are associated with a constructive dialogue.

Brigss, de Vreede, Nunamaker and Tobey (2001) introduced the concept of ThinkLets as a means for less experienced users to apply a groupware tool within their practices. They define ThinkLets as the “smallest unit of intellectual capital required to create repeatable, predictable patterns of thinking among people to work towards a goal” (Briggs et al., 2001). A ThinkLet has three elements:

1. A tool that refers to the specific support that the group uses.

2. A configuration of the tool that specifies how the tool is presented to the users.

3. A script that describes the sequence of events and instructions given to the group.

The notion of ThinkLets was used to development an instructional strategy that guides the introduction of the collaborative tools in the classroom. For every study, we develop a lesson plan in close cooperation with the teacher. That lesson plan describes: 1) the learning goals, 2) the different kind of activities that the students have to carry out, 3) the expected outcomes, 4) the learning materials they receive, 5) the tools they use, and 6) the configuration of these tools.

76

The Learning Environment of theClassroom

Teaching Skills

Teachers are skilled professionals who have been trained to teach students effectively: they have learned the principles of good teaching and how to apply them in the classroom (Slavin, 2000). Good teaching requires specific knowledge, attitude and skills. Teachers must have a thorough understanding of the knowledge domain and they must be able to communicate that knowledge in an attractive manner. They must understand their students and known how they learn and behave. Teachers have to monitor class activities, reflect on their teaching practice and adapt their teaching if necessary. Furthermore, teaching requires the appropriate communication skills like explaining, questioning, summarizing, use of voice, listening, eliciting and giving feedback (Kyriacou, 2001; Slavin 2000; Cohen, Manion & Morrison, 2005).

The research in this thesis does not focus that much on the teacher's behavior; rather it concentrates on the communication between the students. This choice has to do with the fact that collaborative learning considers the interactions between students as the primary vehicle for learning. It does not mean, however, that the teacher does not play a significant role. During the preparation and the evaluation of the studies, we made use of teacher's knowledge and experiences. The teachers that participated in the research helped us to develop the proper instructions to integrate the collaborative tools within the existing learning environment. In the final study – that is discussed in chapter 9 – teacher's interaction with the groups was analyzed with the aim to build a better picture of the performance of the groups.

Communication Tools

Usually, the teacher delivers knowledge and instructions verbally. Communication between the teacher and the class can be supported by various tools; the teacher may use certain media in addition to the verbal, face-to-face communication. The most striking example is the blackboard or chalkboard that has been part of the classroom from the beginning of the 19th century (Coulson, 2006). More recently, electronic devices like interactive whiteboards or handheld devices have been introduced in the classroom to supplement the verbal communication between the teacher and the students. For example, the teacher may use a handheld device to collect responses from students to questions, display them to the class and use that feedback to trigger and

77

Chapter 5

sequence a class discussion (Boyle & Nicol, 2003). Alternatively, learners may ask questions to the teacher by using personal wireless devices (e.g. Ratto, Shapiro, Minh Truong & Griswold, 2003).

5.4 MediaThe learner can study a subject matter by consulting the media available in the classroom such as a textbook or the internet. These physical or digital media represent the learning material in distinct ways. Each medium provides a particular representation of a knowledge domain that influences what can be learned and how to learn it. The attributes of a medium affect the way the learner interacts with the medium and the kind of learning that emerges from these interactions. It influences the way that information is selected and transformed in the acquisition of general cognitive skills (Clark, 1983).

Kozma (1991) came with a classification that takes into account aspects of the medium, of the learner who uses the medium and of the learning activity that the medium tends to support. He classified media by three attributes:

1. the technology,

2. the symbol systems, and

3. the processing capabilities.

Technology

Technology refers to the physical, mechanical, electronic or digital properties of the medium that determine the range of possible representations and the activities that can be performed with these representations. A textbook, the most common medium in the classroom, consists of a set of sheets that are bound together. Such a book contains text and pictures.

Learners do not interact so much with the technology. The technology enables and constrains the other two attributes of the medium. Meaning is derived from the other two attributes of the medium: the symbol systems and processing capabilities. These two attributes have more direct implications for the interactions between the learner and the medium and the cognitions that emerge from these interactions (Kozma, 1994).

78

The Learning Environment of theClassroom

Symbol Systems

Symbol systems are “structures of appearance” (Goodman, 1976) that determine how the knowledge domain is expressed by the medium. A symbol system consists of a set of elements like words, textual objects, algebraic notations or pictures that can be arranged by rules or conventions and point to a certain knowledge domain (Goodman, 1976). Different symbol systems are correlated with distinct meanings and bring about differences in information processing (Salomon, 1979). Certain symbol systems may be better at presenting certain tasks. Graphical representations such as Euler diagrams (Figure 5.3), for example, are most suitable for teaching abstract reasoning because their elements specify information in a distinct number of classes. Such representations permit a limited range of abstraction and are therefore easy to process (Stenning & Oberlander, 1995).

Processing Capabilities

Information is not only presented in memory; it is also processed (Kozma, 1991). The processing capabilities of a medium refer to the cognitive operations that are triggered by the medium and the actions that the learner can perform with that medium. The processing capabilities can be described along several dimensions like perspective, precision and complexity (de Jong, Ainsworth, Dobson, van der Hulst, Levonen, Reimann, Sime, van Someren, Spada & Swaak, 1998).

Perspective refers to decisions about “how and what to see” with regard to the knowledge domain. A software program, for example, can be looked at from different perspectives as a set of instructions that specify its operations or as a set of functions

79

Figure 5.3: Examples of Euler diagrams.

Chapter 5

that describe the purpose of the different elements of the program. Each of these perspectives provides a valid description taking into account the kind of knowledge that should be communicated.

Precision indicates how accurate or exact the knowledge domain is presented to the learner. There are situations where learners need precise descriptions of the knowledge domain that are correct in every detail. For example, programming techniques can best be explained in detail by representing the source codes of program. For users of a software application it is not necessary to go into so much detail. For them a program can be explained by the interfaces that specify the external behavior of the program.

Complexity has to do with the amount of information and how that information is organized. Complexity relates to the information content but what is considered as complex depends on those who perceive that information.

Conclusion

The characteristics of the medium determine the kinds of learning activities that will occur. The structural appearance and processing capabilities of a medium make certain actions available. This also holds true for the collaborative tools that are discussed in this thesis. The collaborative tools represent the information that is communicated in distinct ways.

In the next chapter – where we discuss the design of these tools – we refer to these two attributes of the medium that were discussed in this paragraph. These attributes emphasize that a medium represents knowledge in distinct manner. In our case, they underline certain aspects of the problem space while it “masks” other aspects. It made us aware of the different representational capabilities. That is why two collaborative tools were developed that represent a group discussion in two different ways.

5.5 Other LearnersThe teacher is not the only relevant person in the classroom; interactions with fellow learners also influence the course and outcomes of the learning process. Learning methods like collaborative learning explicitly capitalize on these learner-learner interactions. There is an important difference between teacher-learner and learner-learner interaction. Teachers rely on a variety of resources like the instructional strategy and teaching skills to guide their interactions. Learners who collaborate usually have

80

The Learning Environment of theClassroom

less means at their disposal to organize their communication effectively. For example, students may lack the appropriate discourse strategies that enable them to discuss scientific ideas effectively (Linn & Burbules, 1993). The learning environment of the classroom has to anticipate on this lack of resources. Actually, this observation provides the rationale for the design of the collaborative tools: the collaborative tools aim to facilitate a learner-centered communication that enables a group to learn without direct support of the teacher.

Computer Support for Collaborative Learning

Computers in the classroom can serve different kinds of goals. They can be used to deliver information about the knowledge domain just as the teacher does. Computer-assisted instructions and intelligent tutoring systems are examples of this approach. These systems base their instructions on a cognitive model of the competence that the student is being asked to learn (Anderson, Corbett, Koedinger & Pelletier, 1995).

The use of computer tools from a social-constructive perspective serves a different purpose. Jonassen (2000) considers computer tools as learning tools that students learn with, not from. Computer tools should be used as engagers and facilitators of thinking and knowledge construction (Jonassen, Peck & Wilson, 1999).

Jonassen (1999) makes a distinction between cognitive tools and collaborative tools. Cognitive tools interact with the learner to support their thinking and reasoning. They may help the learner to better represent the problem or task (visualization tools), represent what they know (knowledge modeling tools), offload some of the cognitive activities (performance support) or they may help learners to gather important information (Jonassen, 1999). Collaborative tools, in contrast, support the collaborative activities of a group of learners who carry out a common task. They mediate the interactions between learners with the aim to support their joint learning activities. These tools may have similar properties as the cognitive tools, only its effects are situated on the level of the group rather than on the individual level. They mediate and regulate the communication with the aim to stimulate communication that is positively associated with learning. Precisely, these tools are the focus of the tool design that is discussed in the next chapter.

81

Chapter 5

5.6 SummaryThe existing learning environment within which the new collaborative tools are integrated has a number of characteristics. These characteristics partly determine what kind of collaborative tools will be developed. Two characteristics are of crucial importance for the development of the collaborative tools:

1. the learning task, and

2. the instructional strategy.

The Task

The collaborative tools support small groups of students who are co-located and already communicate face-to-face. These groups work together on a problem-solving task that aims to establish a deeper understanding of the subject matter associated with that task. The problem-solving task can be further specified by its complexity. The problems that the students discuss are ill-structured, they refer to complex situations that are relatively unknown for the students and difficult to comprehend. To solve these novel problems, students have to enter into a constructive dialogue during which they explain their thoughts, build upon other members' understanding of things and discuss alternative interpretations.

Instructional Strategy

The collaborative tools stimulate communication that is associated with a constructive dialogue. The tools cannot be set apart from the instructional strategy that accompanies their introduction in the classroom. The instructions make the tools suitable for a specific learning situation. Such a strategy consists of procedures and guidelines “to create repeatable patterns of acting and thinking” (Briggs et al., 2001). It means that the design will not be limited to the technology but also takes into account how the technology will be made applicable for a specific context of use.

82

6 Changing Communication: Designing

Tools for Collaborative Learning

The group problem-solving process, which I was observing, was constantly interrupted, back-tracked and side tracked, with signs

of growing emotional tensions in and between persons […] A good deal of behavior seemed to be deficit driven, neither

particularly emotional, nor directly instrumental, but in the nature of patch-up work that seemed to be necessary to maintain the

communication process.

Robert Bales1

In this chapter, we discuss the design of two collaborative tools that aim to change the communication patterns of a group of students who participate in a problem-solving discussion. The challenge for the tool design is to create the right conditions for a constructive dialogue during which the students develop a deeper understanding of the subject matter. The design process that is discussed in this chapter starts with an analysis of the existing collaborative learning situation and ends with a description of two collaborative tools that aim to improve that situation. First, we identify the communication patterns that have a negative influence on the performance of the groups whereby performance is associated with the four criteria for a constructive dialogue (see chapter 4). We describe two ineffective communication patterns – interpersonal dominance and product blocking – that prevent group members from freely share and discuss their ideas. The ineffective patterns will be traced back to potentially manipulable aspects of the learning environment that are particular potent in promoting learning effectiveness (Hackman, 1987). In our case, we relate the

1 Robert Freed Bales (2002). Social Interaction Systems: Theory and measurement, page 161-162.

83

Chapter 6

patterns to the underlying structures of the medium for communication. Such a medium has certain structural features that enable the group to arrange their verbal exchanges into a meaningful whole.

The aim of the design is to develop a collaborative tool that counteracts the ineffective communication patterns associated with a verbal discussion. It is hypothesized that improvements do occur because a collaborative tool mediates and regulates the communication differently.

6.1 Ineffective Communication PatternsIn the 1950s the social psychologist Robert Bales (Bales, 2002) studied students who worked together in small groups to solve problems. He noticed that these groups often did not fulfill the rational, goal-directed strategies that are associated with individual problem solving. The group processes that he observed differed from the theoretical accounts that describe problem solving primarily as an analytical activity. Groups sometimes seemed disorientated and did not follow the logic of a rational discussion. Their sequence of talk was incoherent and was not oriented towards the achievement of an acceptable solution. It drew Bales attention away from individual mental processes towards the interaction processes. In line with this shift in focus, our analysis of small-group discussions aims to identify the ineffective communication patterns that hamper group learning. The analysis focuses on the communication patterns that inhibit the free exchange of knowledge. We identify two types of ineffective communication patterns that affect the performance of groups. These patterns are: 1) interpersonal dominance, and 2) product blocking.

Interpersonal dominance

The teachers that participated in our studies mentioned unequal participation as a major drawback of small-group discussions. They reported that a discussion is sometimes controlled by one or more rhetorically “skilled” students who constantly manage to take the turn or to overrule their fellow group members. It means that some students talk a lot while other students hardly get the opportunity to share their ideas with the group. This does not only work out badly for the individual students but it also hampers the learning potentials of the group. The group does not fully capitalize on the knowledge available.

84

Changing Communication: DesigningTools for Collaborative Learning

Dominance and control are regularities in the behavior of persons who score high on the “dominance-submissiveness trait” (Cattell, 1965). They guide much of the seemingly irrational behavior when people discuss their difference of opinion (Stein & Albro, 2001). Bales (2002) considers dominance as one of the basic problems of groups as social-interaction systems. Interpersonal dominance implies the ability and the tendency to use “raw power”, i.e. a tendency to control how others behave apart from any formal “legitimacy” (Bales, 2002). Interpersonal dominance manifests itself by behaviors like repeatedly taking the turn, interruptions and simultaneous speech. These individual behaviors reflect an individual´s disposition to monopolize a discussion.

The disruptive behaviors that are associated with interpersonal dominance can be situated on the individual level as behavior regularities. However, their appearance during a group discussion depends on how the group coordinates their interactions. The group can draw on certain structures and mechanisms that enable them to organize their individual actions into collective behavior.

The disruptive behaviors during a verbal discussion are intentional violations of the turn-taking principle. They typically occur because someone wants to take over the speaking turn (Bull, 2002). A dominant group member interferes the ongoing talk so that fellow group members are not able to finish their turn or to take the turn. Repeatedly taking the turn rests on the basic technique that “first starter gets the turn” (Sacks, Schegloff & Jefferson, 1974). Simultaneous speech refers just to two or more people talking at the same time, whereas interruptions have the intent or the effect of disrupting another person’s speech (Bull, 2002). A discussion becomes scattered when students do not manage these disruptive behaviors properly (Schegloff, 2000). Interpersonal dominance leads to an ‘asymmetrical pattern of communication’ (Zimmerman & West, 1975) whereby some students have fewer opportunities to share their knowledge with the group. Asymmetrical or unequal participation is in conflict with one of the criteria for a constructive dialogue that states that all group members should freely participate. It prevents groups from fully access the knowledge that is available so that they are not able to capitalize on the insights that fellow members have generated (Barron, 2003).

Product Blocking

Brainstorming research provides us with valuable insights in how the sequential organization of the communication affects the cognitive processes on the level of the

85

Chapter 6

group. Various studies compared brainstorming in interacting groups with brainstorming in nominal groups. Nominal groups consist of the same number of individuals as the interacting groups but their members brainstorm alone, i.e. there are no interactions between the group members. Comparing real or interacting groups with nominal groups allowed researchers to analyze the effects of the interaction processes on group performance. Early research indicated that nominal groups outperformed real groups (Diehl & Stroebe, 1987). This “productivity loss” (Steiner, 1972) in real groups is large, and increases with group size; the effect is found for quantitative as well as qualitative measures of productivity (Nijstad, 2000). Three interpretations have been offered to account for the lower productively of real groups (Diehl & Stroebe, 1987): 1) product blocking, 2) evaluation apprehension, and 3) free riding.

Product blocking occurs when individuals need to wait to verbalize an idea because of some procedural limitation (Isakesen, 1998). Production blocking has to do with the turn-taking rule that only one group member speaks at the same time. Group members, who are prohibited to verbalize their ideas on the moment they occur, forget or suppress them so that the process of idea generation stagnates or is disrupted because of blocking (Nijstad, 2000). Evaluation apprehension refers to a fear for negative evaluation from fellow group members. It prevents group members from presenting their more original ideas. Free riding states that any factor that reduces the ability of the group to monitor individual productivity is also likely to reduce a subject’s motivation to participate during a collaborative activity.

Most of the productivity losses in real brainstorming groups were due to blocking (Diehl & Stroebe, 1987; Nijstad, 2000). Nijstad (2000) concluded that idea generation could be made more effective when product blocking is eliminated or kept to a minimum. The use of a collaborative tool that is based on parallel access – as is the case with an electronic brainstorm tool – could counteract the effects of blocking. In electronic brainstorm groups there is no product blocking because group members can put forward a contribution simultaneously, whereas group members still have access to each others´ ideas. Under these conditions, moderate positive effects of idea sharing can be found (Nijstad, 2000).

86

Changing Communication: DesigningTools for Collaborative Learning

6.2 The Envisioned Direction of ChangeThe functional perspective – that has been introduced in chapter 3 – puts the communication processes at the heart of effective group performance. The functional perspective states that the communication processes have utility and may vary. Some communication patterns can be associated with effective performance, while other patterns can be ineffective and hamper learning. The functional perspective asserts that the communication processes can be controlled and regulated by internal group factors or external circumstances so that the desired group behaviors are more likely to occur (Webb & Palincsar, 1996; Hollingshead et al., 2005). This view provides the ground for the support that aims to inhibit the negative effects of the ineffective communication patterns. In this thesis, we follow the functional perspective and discuss the design of two collaborative tools that intervene in the existing group discussion. These tools offer an alternative medium for communication, a medium where group members are not able to display the disruptive behaviors of interpersonal dominance and product blocking.

Figure 6.1 presents the process of tool design that is discussed in the following paragraphs. The design process starts with the two communication problems –

87

Figure 6.1: The process of tool design.

Chapter 6

interpersonal dominance and product blocking – that have been described in the previous paragraph. These problems are further specified in terms of observable communication patterns and their underlying structures. The communication patterns describe the recurrent sequence of related communicative actions, while the underlying structures and mechanisms describe the way that the group organizes these actions into a pattern of communication. Webb and Palincsar (1998) identified a great variety of structures that influence the group processes like the reward structure, group composition and role specialization. In this thesis, we concentrate on the structures that are characteristic for a specific medium. The essence of the tools is to offer the groups an alternative medium for communication – besides human speech – that changes the communication and improves learning.

From Action towards Inter-action

The human activity system that is the focus of the tool design is the group who meets face-to-face to discuss a problem with the aim to increase their understanding of the knowledge domain associated with that problem. These discussions are a specific type of “focused interaction” (Goffman, 1963) that occurs when students collaborate to build a shared representation of the problem space. The interaction can be considered as multiparty talk that consists of loosely connected episodes of talk. Participants either develop or use specific structures to regulate their talk into a coherent whole (Schwartzman, 1989). These structures constitute the significance of the communicative actions of those who participate in a discussion (Cushman & Whiting, 1972). They can be inherent to a medium as the most sensible way to organize the communication or they emerge when the group starts to communicate. An example of the former is turn taking to organize the sequence of verbal exchanges, while an example of the latter may be a strong task focus that leads to considerable task-related communication.

Mediation and the Organization of Communication

The tool design that is discussed in this chapter exploits the structures that are inherent to a medium as an opportunity to change group communication. They do so by changing the informational, temporal, and interactional processes by which groups do their work (Hollingshead & McGrath, 1995).

First, we concentrate on the existing communication and describe the

88

Changing Communication: DesigningTools for Collaborative Learning

coordination mechanisms by which group members organize their verbal exchanges. This is because we assume that the two communication problems – interpersonal dominance and product blocking – are characteristic for human speech as medium for communication.

A group who communicates orally has to organize their interactions according to the “mechanism of turn-taking” (Sacks, Schegloff & Jefferson, 1974). Turn taking rests on the principle that only one group member talks at the same time. In the previous paragraph, we traced the two communication problems – interpersonal dominance and product blocking – back to the mechanism of turn taking. Frequent violations of the turn taking principle make a verbal discussion scattered and incoherent. Furthermore, turn taking may hamper the sharing of ideas because only one group member can put his or her thoughts into words at the same time. It is expected that the communication problems will not occur when the group organizes its communicative exchanges according to a different kind of floor-control mechanism. The collaborative tools should offer such an alternative mechanism in the form of parallel access.

With parallel access, all the group members can access a shared workspace simultaneously. They do not have to wait for their turn but immediately share their ideas with the group. Furthermore, other group members cannot interfere because a group member puts forward a contribution in a private window by typing. It is hypothesized that parallel access counteracts the communication problems that are exemplary for verbal exchanges. It ensures an equal participation of the group members during the discussion.

Guidelines for a Constructive Dialogue

The transition from turn taking to parallel access is a basic property of the collaborative tools. However, this guideline is not sufficient; it describes the tools in rudimentary terms. Parallel access as a design guideline triggered further research that aimed to come up with a collaborative tool that is specifically adapted to the specific context of use, in our case, a group who carries out a constructive dialogue. In chapter 4, we defined four criteria that are associated with a constructive dialogue. These criteria served as reference for the tool design, they are:

1. An orientation towards the achievement of the learning task.

2. Equal participation and the free exchange of ideas.

89

Chapter 6

3. A coherence discussion where successive communicative actions are connected in a meaningful way.

4. Communication pattern that are directed at knowledge elaborations.

6.3 Design GuidelinesThe remaining paragraphs of this chapter discuss the design of two collaborative tools. We present eight guidelines that capture the properties of the collaborative tools. The design guidelines contain clear expectations how the collaborative tools mediate and regulate the communication and create better opportunities for learning. The design guidelines that are discussed in this paragraph are:

1. The collaborative tools match with the characteristics of effective task performance.

2. Associate a medium with a specific communicative function.

3. Ensure joint attention by using only one collaborative tool at the time.

4. Parallel access as floor-control mechanism.

5. Functional spaces to coordinate the computer-mediated exchanges.

6. Relate functional spaces to relevant macro structures that are associated with task performance.

7. A global model of coherence instead of a local model coherence to organize the communicative exchanges.

8. A notation system that stimulates knowledge elaborations.

Effective Task Performance

Small-group discussions require a lot of effort from the students. Students have to address several aspects of their performance as a group: they have to maintain durable social relations and acceptable levels of participation, manage the relations with outsiders such as the teacher, solve a complex problem, and they have to establish the proper conditions for learning. In general, one can make a distinction between two dimensions of group performance: groups have to carry out a task and they must act as a group (Hare, 1960; Fisher, 1980; McGrath, 1984; van Rees, 1999). Groups have to put effort in the performance of the task and they have to invest energy in maintaining themselves as a group. The group members must simultaneously attend to and develop a shared problem space and a relational space consisting of interactional challenges and

90

Changing Communication: DesigningTools for Collaborative Learning

opportunities (Barron, 2003).

The first function – carrying out the group task – requires a lot of cognitive efforts, especially in the case of complex, ill-structured problems. These types of problems refer to situations where it is not clear at the beginning what the problem exactly entails and which actions may lead to the solution of the problem. Problem solving can be characterized as a nonlinear transformation process: the process has no clearly defined beginning and the output could be achieved along multiple paths simultaneously (Taylor and Felten, 1993). These problems are open to different interpretations while solving them requires the application of multiple viewpoints and mutual knowledge (Schwartz, 1995). Groups who have to solve ill-structured problems are faced with the challenge to address the different viewpoint and to develop a shared understanding.

A second complicated factor with regard to the performance of groups has to do with maintaining a sense of “groupness”, i.e. with the development and maintenance of the group as a system (McGrath, 1991). When students collaborate, they have to maintain durable relationships and acceptable levels of collaboration. This entails the development of a climate that socializes the uniqueness of the individual into the social system of the group (Fisher, 1980).

The two functions described above are fundamental for every social system. They are associated with two distinct types of communicative functions, i.e. task-related and social-emotional communication (Bales, 1950). Task-related communication is used to explicate the task processes in groups. It enables group members to share and use knowledge and information that are directly related to task performance (Propp, 1999). Social-emotional communication affects student’s perception of the other group members and the relationships they form (Propp, 1999). The minimal number of categories of social-emotional interactions includes control and affection (Hare, 1960).

The question remains which communicative function – task-related or social-emotional – should be supported by the collaborative tools? It is assumed that groups benefit from effective task performance because their task-related communication is generally related with learning. Task-related communication leads to cognitive activities often referred to as knowledge elaboration, which, in turn, facilitates the acquisition of that knowledge (Draskovic, Holdrinet, Bulte, Bolhuis and Leeuwe, 2004).

91

Chapter 6

General design guideline 1 (Box 6.1) states that the collaborative tool should provide the students with the proper environment that enables them to carry out their task effectively. The digital medium has to facilitate the communication that is associated with effective task performance. This will be the incentive for students to use the collaborative tools. The tools provide the students with a medium that better enables them to discuss their learning task.

Design guideline 1 lies at the basis of a number of subsequent design guidelines that concern the detailed design of the collaborative tool and the associated instructions.

Two Media for Face-to-face communication

Group communication in the envisioned collaborative learning situation will be distributed between two modes of communication: a verbal part and a digital, computer-mediated part. Groups can communicate verbally or by means of the computer. General design guideline 2 (Box 6.2) states that each medium has to display

92

Box 6.1: Design guideline 1.

Computer support for effective task performanceUsers are encouraged to perform their task-related communication within the shared workspace of the collaborative tool. Task-related interactions are associated with knowledge elaboration and learning. It is expected that students will be using the collaborative tools when they are configured in such a way that it closely matches the characteristics of effective task performance.

Box 6.2: Design guideline 2.

Associate a medium with a specific communicative functionThe collaborative tool has to display communication that has to do with the task. It will be more difficult for the group to focus on the task when both communicative functions – task-related and social-emotional – occur in the shared digital workspace of the collaborative tool.

Changing Communication: DesigningTools for Collaborative Learning

one communicative function so that the communication in that medium will be meaningful. The distinction between task-related and social-emotional communication will be used as the criterion to split the communication between the two media. It is assumed that a clear distinction will be beneficial for the coordination of the activities. When a specific communicative function will be displayed in one medium – verbal or computer-mediated - attention of the group will be more focused.

Face-to-face versus Online

Design guidelines 1 and 2 set the collaborative tool apart from online learning environments. Online learning environments support collaboration between students who are dispersed in time or space. In general, these environments aim to mediate a broad range of interactions. An enrichment of the information flow improves learning: for example, students are presented with multiple tools simultaneously to address the various communicative functions (see e.g. van Amelsvoort, 2006), or the students use awareness tools that provide the group with detailed information about their members or their performance (see e.g. Janssen, 2008). Online learning environments seem to reflect an attitude of “more support is better’” i.e. a rich information flow between the learners is a good guarantee for collaboration and learning. The computer support that is discussed in this thesis is more in line with the kind of support that has been developed in the field of Computer Supported Collaborative Work (CSCW). Various collaborative tools within that field have been developed to support professionals who meet face-to-face. These tools mainly focus on task performance; they provide support for specific task-related activities such as brainstorming or decision-making.

A Common Task Focus

Collaboration is a process by which individuals negotiate and share meanings relevant to the problem solving task at hand (Roschelle and Teasley, 1995). As mentioned before, it is a kind of “focused interaction” that requires a common orientation. Such an attentional engagement is a prerequisite for coordinated interactions (Barron, 2003). Attentional engagement is crucial for the success of communication during collaborative learning (Sfard & Kieran, 2001). If the learners do not focus on the same activity, it will be less likely that they built upon the knowledge that is shared by others. General design guideline 3 states that joined attention can best be ensured when all the learners work together in one collaborative tool (Box 6.3).

93

Chapter 6

Group communication already splits between an oral and a computer-mediated part. The computer-mediated part occurs in one environment so that coordination of the communicative exchanges will not become too complex. We assume that it is undesirable to divide group activities between multiple collaborative tools otherwise students may not be aware of what other students share with the group.

Implementation

Design guideline 3 implies that the task-related interactions of all the users take place within the same tool. This guideline is not so much a technical requirement. It will be implemented as an instruction that makes the final software system – that contains a number of collaborative tools – suitable to support a small-group discussion.

Parallel Access as Floor Control Mechanism

The two ineffective communication patterns – interpersonal dominance and product blocking – are associated with turn taking as floor-control mechanism. This mechanism is typical for the organizational of verbal sequences of talk.

Interpersonal dominance manifests itself as frequent violations of turn taking and in difficulties in gaining the floor (Barron, 2003). Product blocking occurs when group members forget or suppress their ideas because they are prohibited to verbalize their ideas when another group member is speaking. In these cases, the turn-taking mechanism prevents group members from “getting the floor”.

Floor control refers to a pattern of control and authorities relative to participants’ interaction in the shared workspace, i.e. the “floor” (Heeren, 1996). A change in the floor control mechanism affects the behaviors that the group members will display. It is assumed that parallel access as floor-control mechanism counteracts unequal

94

Box 6.3: Design guideline 3.

Ensuring joint attentionAll computer-mediated communication should be concentrated within one collaborative tool. Attention of the group will be more focused when all the digital communication will be concentrated in the shared workspace of one tool.

Changing Communication: DesigningTools for Collaborative Learning

participation that is caused by interpersonal dominance (Box 6.4). Parallel access has a direct impact on productivity because a wider bandwidth allows for parallel entries, coupled with instantaneous dissemination of ideas to group members (Gallupe, Bastianutti & Cooper, 1991).

Implementation

In the case of “free for all” or “parallel access”, the coordination of interactions is a shared responsibility of all the users. All the students can put forward their contributions at the same time without any disturbance. They use a text-based, digital medium to share their ideas with the group. Users type their ideas in a private window that is not accessible by other users. Only after they submit their text it becomes visible in the shared workspace.

It is hypothesized that parallel access stimulates the free exchange of ideas because participants can put forward their ideas without being interrupted. They can directly share their ideas without having to wait for their turn.

Task- or Problem Decomposition and Functional Spaces

Design guideline 4 suggests the use of parallel access as floor control mechanism. With parallel access, all participants contribute to the discussion without any direct constraints from their peers. They can simultaneously access the shared digital workspace to put forward their ideas in writing.

Parallel access brings along its own dynamics with regard to the coordination and organization of the communicative exchanges. The study that is discussed in chapter 7 indicates that parallel access encompasses the danger of disorganization; it is just as if

95

Box 6.4: Design guideline 4.

Parallel access for the free expression of ideasParallel access enables users to share their ideas directly and fully with the group. A user who wants to contribute does not have to wait until another group member finished speaking like in the case of a verbal discussion. A user who puts forward an idea cannot be interrupted by fellow group members. It is expected that this may lead to equal participation.

Chapter 6

several people talk at the same time. Communication problems that are associated with parallel access are: 1) extensiveness of contributions in the shared workspace, and 2) a lack of a common group focus with regard to the topic of discussion. The two issues are related in the sense that they strengthen each other. There is the danger that the shared workspace becomes crowded. The number of contributions increases quickly, which makes it difficult for group members to monitor what is going on. Group members are not able to keep track of the various topics that are put forward by their fellow group members. Without timely joint attention, basic processes as sharing perspectives, increased monitoring and providing explanations will be compromised (Barron, 2003).

A number of studies investigated how persons coordinate their actions when they access a shared workspace simultaneously (Tang, 1991; Dwyer & Suthers, 2006). Tang (1991) emphasized the spatial properties of the medium. He observed that the spatial orientation of the shared workspace could serve as a resource that allows persons to establish a meaningful context for their contribution.

Dwyer and Suthers (2006) studied how groups of two people construct a representation collaboratively. Participants in their study typed messages in a two-dimensional shared workspace that was positioned horizontal between them. They provided all groups with large sheets of paper, some groups also received yellow notes, to brainstorm ideas, discuss them and come to some kind of final agreement. The participants could access the workspace simultaneously to read and write independently from each other. Dwyer and Suthers (2006) observed that the participants in the study created functional spaces in the shared workspace like a personal area, a shared work area or an area to list conclusions. Some of these areas emerged in every group; they seemed to fulfill a common goal that could be associated with the nature of the task.

The studies of Tang (1991) and of Dwyer and Suthers (2006) indicated that the participants used the spatial properties of the shared workspace to coordinate their actions into a coherent whole. It is assumed that a shared workspace that offers a structure that divides the two-dimensional workspace counteracts the coordination problems associated with parallel access. Such a structure divides the shared workspace into functional spaces that have a specific meaning. They are meant to organize the contributions in a coherent manner. This brings us to next design guideline that is presented in Box 6.5.

96

Changing Communication: DesigningTools for Collaborative Learning

Implementation

Design guideline 5 proposes a solution for the coordination problem that is associated with parallel access as floor-control mechanisms. The shared workspace will be structured into functional spaces by the use of categories in the case of the Threaded-discussion tool or areas in the case of the Graphical tool.

The Threaded-discussion tool – the first envisioned tool – is a tool that enables users to communicate with each other by means of textual messages. The tool provides the users with a stratified structure where each layer reflects different levels of elaboration. Categories resemble the first level of elaboration. A category is linked to a specific shared workspace. When a user selects a category, the associated workspace becomes visible. The discussion that occurs in that workspace refers to that specific category.

With the Graphical tool – the second envisioned tool – users make spatial representations of concepts and their interrelationships. An area is a demarcated space within the two-dimensional workspace of the graphical tool. It contains contributions that are related.

Categories and areas can be considered as representation aids that help users to organize their digital discussion according to some meaningful structure. They help users to apply coherence and meaning to the ongoing discussion. This brings us to a next question: What kind of macro structures can be used to define functional spaces?

Functional Spaces and a Relevant Macro Structure

Students benefit from representation aids that make them aware of relevant aspects of the problem space. It is assumed that a macro structure that displays an essential aspect of the problem helps learners to construct a shared representation of the problem space.

97

Box 6.5: Design guideline 5.

Functional spaces for more coherenceContributions of the group members that are displayed in the shared workspace of the collaborative tools become more coherent if they are organized according to predefined structure that divides the shared workspace into functional spaces that have distinct meanings.

Chapter 6

One of the macro structures is “task or problem decomposition” that breaks down a problem or a task to its component pieces. Task decomposition reduces the cognitive load resulting from the demands of group interaction (Coskun, Paulus, Brown and Sherwood, 2000). It is assumed that task or problem decomposition proceeds along structures – i.e. categories or functional areas – that are relevant for solving the problem (Box 6.6).

Implementation

Design guideline 6 is part of the instructional strategy. Decomposing a problem into meaningful categories depends on the kind of learning task that students have to carry out. Decomposition helps students to construct a shared problem space that contains all the relevant aspects of the problem. The categories – that visible for all group members – display essential characteristics of the problem space.

Coherence

Coherence refers to the fact that a sequence of utterances is usually connected to each other in orderly and meaningful ways (Craig & Tracy, 1983). Patterns of coherence relate sentences or propositions as wholes (van Dijk, 1985). Coherence displays itself in different manners. The patterns that generate coherence may include rules about turn taking and the kinds of utterances that may follow each other or structural formats for introducing and developing topics or telling stories (Craig & Tracey, 1983). In general, coherence can be described in terms of local and global models.

98

Box 6.6: Design guideline 6.

Associate functional spaces with a relevant macro structureDecomposing the problem into meaningful subproblems provides the learners with task-relevant macro structure that helps them to better analyze and discuss the problem at hand. Categories or meaningful areas that subdivide the shared workspace are used to prime learners cognitively. It structures the problem into smaller pieces that are better comprehensible.

Changing Communication: DesigningTools for Collaborative Learning

Local Coherence

Local models of coherence derive meaning from a sequence of contributions that directly succeed each other. Figure 6.2 represents the principle of local coherence: a statement receives its meaning in relation to a statement that directly precedes in time while different topics of discussion are temporally ordered into a coherent whole. This pattern is typical for verbal communication. The coherence of a verbal discussion is based on the assumption that each act is an appropriate response to a previous act (Littlejohn, 2002).

Coherence during a verbal discussion is realized by principles of turn taking and adjacency pairs or “nextness”. Time is a determining feature of local coherence; the adjacency of utterances in time is used to interpret meaningful sequences in a discussion. Turn taking and adjacency pair are units for sequence construction that transform individual contributions into a coherent and meaningful whole. Turn taking serves as a means to coordinate the sequence of turns whereby one person needs to talk after the other. Adjacency pair serves as a principle to extract meaning from a sequence of communicate acts during a discussion that is based on turn taking. The basic form of adjacency pair is characterized by certain features. It is composed of two turns by

99

Figure 6.2: Local coherence.

Chapter 6

different speakers that are adjacently placed, that is, one after the other. The two turns are relatively ordered into a first and second pair set. The first pair set contains utterance types that initiate some exchange while the second pair type is responsive to the action of a prior turn. Adjacency pairs compose pair types such as question-answer (Schegloff, 2007).

Global Coherence

Global models of coherence assume that a discourse is a dynamic process of back-and-forth practical reasoning. These models judge the coherence of a discourse on the basis of the overall reasoning (Littlejohn, 2002). A discourse may be coherent with respect to a particular theme, topic or event. Global models of coherence are less visible in verbal discussions where adjacency pair is a leading principle for sequence construction (Schegloff, 2007). This in contrast to, for example, written texts like a scientific article where a generic structure of meaning seems to be more vividly displayed (Goldberg, 1983). Such an overarching structure of meaning guides the readers and writers during their comprehension of the text. They model the cognitive mechanisms that readers and writers apply when they process the text (Knott & Sanders, 1998).

Implementation

Text-based, computer-mediated communication that is based on turn taking as in the case of chat tools shows a high degree of disrupted adjacency, overlapping exchanges and topic delay. These tools are claimed to be interactionally incoherent (Herring, 1999). It is hypothesized that a global model for coherence gives students more “cognitive freedom” in the sense that they can follow their own line of thinking more

100

Box 6.7: Design guideline 7.

Global coherenceA mediated discussion should be based on a global model of coherence. The use of “connections” enable students to respond to a previous contribution that does not directly precede in time. Users can relate contributions based on a macro structure of meaning instead of a purely temporal structure. It is expected that the students experience more “cognitive freedom” when they discuss a problem based on a global model of coherence.

Changing Communication: DesigningTools for Collaborative Learning

easily during the group discussion (Box 6.7). Their thinking is not primary guided by the current topic of discussion, as is the case during a verbal discussion. Instead, students follow general structure of reasoning that enables them to address different topics continuously.

Figure 6.3 displays the principle of global coherence as it is implemented in the two collaborative tools. Global coherence is stimulated by a macro structure like categories in the case of the Threaded-discussion tool that divides a discussion into a number of topics. During the whole discussion, participants can contribute to these topics. They are not hampered by a temporal sequence of communicative exchanges that is characteristic for local coherence.

The collaborative tools provide the students with an alternative way to organize their discussion in a coherent manner (Box 6.7). Students organize their discussion based on a macro structure of meaning that better reflects their individual thinking. This results in more cognitive freedom within the context of the group discussion. The group discussion moves away from a temporal sequence of related messages towards a form of collaborative reasoning based on a general structure of meaning.

101

Figure 6.3: Global coherence.

Chapter 6

The permanence of contributions in the shared workspace is a necessary condition for global coherence. The contributions in the workspace remain visible so that users can relate their message to a relevant contribution that does not directly precedes in time. To summarize, global coherence is achieved by:

– representational aids like categories and meaningful areas that reflects a macro structure of meaning that guides the discussion,

– permanence of messages in the shared workspace, and

– the possibility to link a new contribution to a contribution in the shared workspace to display relatedness.

Encouraging Elaboration: The Use of a Notation System

Collaborative learning requires active participation of all learners and extensive task-related interactions. The task-related communication has to be oriented at the elaboration of the knowledge that is shared by the group. It does not only include the exchange of ideas but the students also build upon the shared knowledge by offering explanation or justifications, making inferences or drawing conclusions. These kinds of interactions make a group discussion into a constructive dialogue.

It is assumed that a notation system encourages learners to display communicative actions that can be associated with a constructive dialogue (Box 6.8). A notation system labels a contribution: a user chooses between a limited set of meanings or functions to label a message. This can be done before a messages is placed in the shared workplace or afterwards. Baker and Lund (1997), for example, provided students with a structured interface with a restricted set of communicative acts to guide their communication. Their analysis indicated that a structured interface promoted interactions that enabled learners to collaborate effectively.

102

Box 6.8: Design guideline 8.

Notation system to stimulate elaborationIt is expected that a notation system guides the communication between students. A notation system displays a limited range of meanings or functions – that are called labels – that are presented to the user as choices. These labels serve as a semantic guidance that keeps a discussion focused.

Changing Communication: DesigningTools for Collaborative Learning

Implementation

A notation system consists of a set of labels that represents certain meanings or functions. These labels mark a communicative act. This will stimulate users to put forward contributions that are associated with that label, i.e. the labels that are displayed by the notation system guide the discussion. For example, the label “Question” encourages users to ask questions.

The labels that will be used depend on the specific pedagogical objectives, the learning task and the learning activities that the teacher wants to stimulate. For example, a notation system may encourage the occurrence argumentation by provide the learners with a set of labels that represent a formal argumentation scheme.

We make a distinction between two kinds of notation systems: 1) structured communication where students label a contribution beforehand, and 2) unhampered communication where students label a contribution afterwards with the support of post-hoc notation system.

With the first option, users choose a label before they start to type a text message. This option explicitly restricts the communication because the users can only choose between a limited set of labels. With the second option, users label a contribution after it has been placed in the shared workspace. This option should stimulate reflection on what has been put forward digitally.

6.4 Functional DesignAt the start of the LEAD project, we envisioned two collaborative tools: a Graphical tool and a Threaded-discussion tool. These tools were only rudimentary described.

The Graphical tool allows users to make abstract representations of the problem. These representations describe concepts and their interrelationships. Examples of such representations are argumentative diagrams that show the structure of arguments and conclusion or causal maps that show influence and causality (Huff, 1990). The Threaded-discussion tool is a collaborative tool that supports meaningful conversations among students (Jonassen, 2000). With this tool, users share, discuss and organize their ideas into a shared representation that addresses various aspects of the problem.

In this paragraph, we describe how the design guidelines that were presented in the

103

Chapter 6

previous paragraph were translated into a number of services. These services describe a set of tangible actions that the user can perform with the collaborative tool. They lay down the external behavior of a tool from the user's point of view (Davis, 1990). The services are functional descriptions that served as input for the software development process. This process resulted in a software system – which has been named CoFFEE – that was developed by the technical partner2 of LEAD's project team.

The functional design has been described in a number of documents that the software designers used as reference. It did not only include a description of the services but also several images or mock-ups of the user-interface. We also used data from existing studies to develop scenarios that describe how students use the envisioned tools (see Appendix A).

The Graphical Tool

The Graphical tool has – in contrast to the Threaded-discussion tool – several predecessors. Graphical tools to support learning include single user tools (e.g Inspiration3) and multiple users tools (e.g. Digalo4 and DREW5). These tools have been taken into account whereby we analyzed how certain user-interface issues were resolved by these tools.

The Graphical tool is based on a concept-mapping interface. A concept-mapping interface represents concepts and the relationships between concepts in a two-dimensional space. It displays nodes (concepts or ideas) and links (a relationship between two concepts). A node is visualized as a shape (text box) that contains textual information, while a link is visualized as a line between two shapes (Figure 6.4).

The Graphical tool has a number of structural features that help users to represent ideas into a coherent and meaningful way. These structures can be considered as representation aids that are relevant for effective task performance. It is expected that the representational aids stimulate the occurrence of communication patterns that are

2 ISISLab, the research lab of the Dipartimento di Informatica e Applicazioni “R.M. Capocelli” of the University of Salerno

3 http://www.inspiration.com/

4 http://www.argunaut.org/

5 http://drew.emse.fr/

104

Changing Communication: DesigningTools for Collaborative Learning

beneficial for problem solving and learning. For example, the labels of the notation system can be used to distinguish between different types of contributions (see Figure6.4). In that case, users select a label to mark a contribution. The representational aids that are characteristics for the Graphical tool are:

– The representation of ideas as textual objects in a two-dimensional graphical workspace that can be accessed simultaneously by the users.

– Functional spaces to allocate a specific meaning to areas in the shared workspace.

– A notation system that elicits specific communicative actions in the shared workspace of the tool.

– The possibility to connect related concepts so that the sequence of contributions can be organized in a logical order that reflects the reasoning of the group.

Appendix B provides an example of how we further elaborated on these representation aids. It describes the initial functional design of the Graphical tool.

105

Figure 6.4: A first mock-up of the graphical tool.

Chapter 6

The Threaded-discussion Tool

The Threaded-discussion tool supports a discussion where students have to elaborate on various topics related to the problem. Students share, discuss and organize their ideas in a systematic manner.

The Threaded-discussion tool has a stratified structure that represents different levels of abstraction. The different layers serve as representation aids that help users to organize their discussion. Categories, for example, arrange the discussion based on most essential points.

Figure 6.5 gives a first impression of the user interface of the Threaded-discussion tool. Users type their contribution in the “add contribution window” and by pressing the “submit” button the contribution is placed in the shared workspace. Users respond to a contribution in the workspace by selecting that contribution and then press the “comment’ button”.

The Threaded-discussion tool has a number of structures that help users to organize their discussion into a coherent and meaningful whole. These structures resemble the structures that are implemented in the Graphical tool. The tool represents and relates contributions on the basis of three criteria:

106

Figure 6.5: A first mock-up of the threaded-discussion tool.

Changing Communication: DesigningTools for Collaborative Learning

1. Categories to display various topics of the discussion.

2. A notation system that elicits specific communicative actions in the shared workspace of the tool.

3. Connections that enables a user to link a new contribution with a contribution in the shared workspace to indicate relatedness.

A separate window displays the categories. This Category window is always visible unless there is only one category. The users select a certain category in the Category window, which opens a shared workspace – associated with that category – within which the users place a contribution. Categories are defined beforehand by the teacher and they are the same for all the students that compose the group.

Users connect a new statement to a contribution in the shared workplace. A connection is made by placing the new contribution underneath the ‘parent’ contribution (left structure of Figure 6.6). A comment could also become a parent contribution (right structure of Figure 6.6).

The Threaded-discussion tool has a notation system similar to that of the Graphical tool. The notation system consists of a set of labels that represent certain communicative actions.

Why Two Collaborative Tools?

The two collaborative tools support particular group activities during a discussion. The Graphical tool allows students to make abstract representations that specify their ideas into a distinct number of classes like argument in favor, argument against and conclusions. It restricts the communication of the group to a limited number of

107

Figure 6.6: Connections as a means to organize related contributions.

Chapter 6

communicative acts. The spatial limitations of the two-dimensional workspace further confine the number of contributions that can be placed in the workspace. Sometimes, the task makes it impossible to construct an abstract representation that contains a limited number of messages. The Graphical tool seems to be less useful to represent the richness of a discussion that addresses various topics. Then the students need more freedom of expression. The Threaded-discussion tool provides the students with that freedom. The tool has a shared workspace that more closely resembles the way students share ideas during a verbal discussion. In addition, the Threaded-discussion tool has some representational aids that help students to organize their digital discussion into a coherent and meaningful whole.

Relationships between the two collaborative tools

The two collaborative tools show a large degree of correspondence. The Graphical tool has structural features similar to the Threaded-discussion tool. They are implemented differently because the two tools support different kinds of collaborative activities. The shared workspace of the Threaded-discussion tool is based on a stratified structure of: 1) categories, 2) threads that consist of main ideas and underlying elaborations, and 3) labels that classify a single contribution. Correspondingly, the shared workspace of the Graphic tool has three representational aids: 1) meaningful areas, 2) links, and 3) labels.

To summarize, both collaborative tools offer the users the ability:

– to divide the shared workspace in functional spaces, i.e. categories for the threaded-discussion tool and meaningful areas for the graphical tool,

– to link ideas that are related,

– to arrange a discussion according to a macro structure, i.e. categories and threads for the Threaded-discussion tool and meaningful areas and links for the Graphical tool, and

– to label contributions according to a notation system.

6.5 CoFFEE: The Software SystemThe two collaborative tools that have been discussed in this chapter are part of a networked-learning environment named CoFFEE that aims to support group

108

Changing Communication: DesigningTools for Collaborative Learning

discussions in the classroom. The CoFFEE system makes a distinction between four kinds of tools: 1) collaborative tools, 2) shared tools, 3) personal tools, and 4) communication tools (see Appendix C). In this paragraph, we present the two collaborative tools of CoFFEE.

The Final Graphical Tool

The final Graphical tool (Figure 6.7) has several structural features that help the students to represent and organize their ideas. The representational aids that are characteristics for the basic graphical tool are:

– the possibility to divide shared workspace into meaningful areas that have a specific meaning,

109

Figure 6.7: User interface of the Graphical tool.

Chapter 6

– the possibility to link related concepts so that the external representation can be

organized in a logical order that reflects learners’ reasoning, and

– a notation system to elicit specific communicative actions.

Meaningful areas

Design guidelines 5 and 6 state that students who discuss a complex problem may benefit from a decomposition of that problem into component pieces. Dividing the shared workspace into meaningful areas provide the students with a task-relevant macro structure that helps them to address the complexity of the problem. The teacher divides the shared workspace beforehand into a number of columns or quadrants. Each column or quadrant can be named. There is a navigation window where users select the designated area so that they can move quickly to that specific part of the shared workspace.

Link

A concept-mapping interface enables users to link related contributions. Users relate a contribution by drawing a line between two contributions. They can also classify the link by selecting a label. The teacher has to define the labels beforehand. Examples of labels are causal relationship for causal maps, support or opposition for argumentative maps or association in the case of a concept map.

Notation System

A notation system consists of a set of labels or contribution cards that represents certain communicative acts. There are two strategies to attach a label to a contribution; these strategies are the same for both tool. Users attach a label before they put forward a contribution or they label a contribution after it has been placed in the shared workspace.

The Final Threaded-discussion Tool

The final Threaded-discussion tool has a shared workspace where users put forward written contributions (Figure 6.8). The tool provides the users with a number of representational aids to organize their discussion. The representational aids include:

1. categories,

110

Changing Communication: DesigningTools for Collaborative Learning

2. the function to connect a new contribution to an existing one, and

3. a notation system that enables users to label a contribution.

Categories

Design guidelines 5 and 6 form the rationale for the use of categories. Categories decompose the topic of discussion into a number of essential points that make the students aware of the different aspects of the problem or task. A user who selects a certain category in the Category window sees the shared workspace associated with that category. A contribution that is placed in that workspace refers to that category, and hence to a specific topic.

The Category window is always visible for the users. Categories must be defined beforehand by the teacher. This is done at the beginning of the session after the teacher has started Threaded-discussion tool.

111

Figure 6.8: User-interface of the Threaded-discussion tool.

Chapter 6

Connections

Students connect a new contribution to any contribution that is already placed in the shared workspace. The existing contribution does not have to precede directly in time. This rule sets the discussion that occurs in the Threaded-discussion tool apart from its verbal counterpart. Users organize their discussion according to a global model of coherence that enables them to carry out a discussion based on back-and-forth practical reasoning. The resulting discussion can be visualized as a tree where each branch or thread represents a sequence of related contributions.

Connections in the Threaded-discussion tool differ from the linking mechanism in the Graphical tool. In the Threaded-discussion tool, users must connect a new contribution to an existing one while in the Graphical tool the use of links is optional.

Notation System

A notation system consists of a set of labels that attach a certain meaning to a communicative act. These labels have to be defined in advance by the teacher. The notation system can be applied in two different ways:

1. as a semantic guidance when the students put forward a contribution, or

2. as a means for reflection when a contribution is labeled after is has been placed in the shared workspace.

In the case of semantic guidance, a label is selected before the user types in a text. The text is submitted in the shared workspace with the label attached to it. Figure 6.9 displays a notation system that consists of three labels: problem, cause or sign. The users have to decide in advance what kind of label they want to attach to a new contribution. Next, they type in a text, submit the text to the shared workspace where it is displayed with the label attached to it.

With the second option (Figure 6.10), the users classify a contribution after it has been placed in the shared workspace. A user selects a contribution and right clicks with the mouse button to attach a label to a contribution.

112

Changing Communication: DesigningTools for Collaborative Learning

6.6 SummaryIn this chapter, we discussed the design of two collaborative tools to support a constructive dialogue between a group of students who communicate face-to-face. We faced the challenge to define the proper conditions for such a dialogue that combines verbal with computer-mediated communication. To do this, we identified two ineffective communication patterns – interpersonal dominance and product blocking –

113

Figure 6.10: Label a contribution afterwards.

Figure 6.9: Label a contribution in advance.

Chapter 6

that can be associated with a verbal discussion. These communication patterns hamper the free exchange of ideas. From there we moved towards the design of two collaborative tools that counteract these problems. The heart of the tool design consists of a set of eight design guidelines that contain some clear expectations or hypotheses of how the collaborative tools change the group communication for the better.

Table 6.1 gives an overview of the eight design guidelines that comprise the conceptual part of the tool design. In chapter 4, we stated that the design activities have a twofold focus: we have to develop guidelines for the design of the collaborative

114

Designguidelin

e

Description Object of the design

Studies into the design guidelines

1 The collaborative tools closely matches the characteristics of effective task performance.

Tools and instructions

Chapter 8 & Chapter 9

2 Associate a medium with a specific communicative function.

Tools and Instructions

Chapter 9

3 Ensure joint attention by concentrating the computer-mediated part of the communication in one collaborative tool.

Instruction Chapter 7, Chapter 8 & Chapter 9

4 Use parallel access as floor-control mechanism for the free expression of ideas.

Tools Chapter 7 & Chapter 9

5 Divide the shared workspace into functional spaces to stimulate coherence of the computer-mediated part of the communication.

Tools Chapter 7 & Chapter 9

6 Associate functional spaces with a relevant macro structure that supports the construction of a shared problem space.

Tools and instructions

Chapter 7 & Chapter 9

7 Establish global coherence by using connections that enables users to respond to contributions that do not directly precede in time.

Tools Chapter 8 & Chapter 9

8 Use a notation system to stimulate elaboration. Tools Chapter 8 & Chapter 9

Table 6.1: Design guidelines.

Changing Communication: DesigningTools for Collaborative Learning

tools but we also have to formulate instructions that help students to use the tools in an appropriate manner. The third column of Table 6.1 “Object of the Design” displays the scope of the design guidelines: a guideline applies to the tool design, the instructional design or to both.

In the following three chapters, we evaluate the hypotheses that lay at the basis of the design. Table 6.1 reports which guidelines will be discussed in the three chapters. The findings that are presented in these chapters reflect a progress in understanding with regard to the use of the collaborative tools. This understanding cannot be set apart from the design decisions that have been discussed in this chapter. In fact, some design guidelines were only formulated after other guidelines had been evaluated in practice. This repeated sequence of design, evaluation and adaptation of the design reflects the iterative character of the research.

6.7 ResearchIn this chapter, we presented the eight design guidelines as a whole, which gives the impression that they were developed simultaneously. This was however not the case; it took several successive cycles of research which ended in eight guidelines. Initially, four design guidelines were formulated that set the direction of change. These guidelines reflect two basic properties of the envisioned collaborative tools:

1. The collaborative tools mark a transition from “turn taking” towards “parallel access” as floor-control mechanism to stimulate the free expression of ideas.

2. The collaborative tools emphasize a focus on effective task performance as a means to stimulate group learning.

We expected that the two basic properties reduce or neutralize a number of the communicative barriers associated with a verbal face-to-face discussion. The first property, parallel access, counteracts the ineffective communication patterns that have to do with interpersonal dominance and product blocking. The second property, a focus on task performance, provides the group with a collaborative tool to support their task performance, and by doing so, creates the proper conditions for learning.

115

Chapter 6

Validation

The guidelines that relate to the two basic properties of the tools were separately addressed in different studies. These studies – that are discussed in chapter 7 and 8 – validated one of the basic properties of the tool. Validation is the process of demonstrating that the design meets the user's true requirements (Hall, Hall & Zeleznikow, 2003). Validation refers to research that examines if a collaborative tool has been developed right. For example, such kind of research analyses if parallel access really enables the students to freely share their ideas. Initial research into the design guideline “parallel access” revealed some hurdles with regard to the communication that was mediated by the tool. The groups faced a new problem that mainly had to do with the coordination of the communicative exchanges in the digital workspace. This issue will be addressed in the next chapter.

Educational Impact

The eight design guidelines served as the basis for the development of two collaborative tools that are part of the CoFFEE software system. These tools – the Threaded- discussion tool and the Graphical tool – were evaluated in a study that is presented in chapter 9. That study had a different perspective than the studies in chapter 7 and 8. It did not evaluate a single design guideline. Rather, the study focused on the educational impact of a collaborative tool as a whole. Educational impact concerns the effect of the tool on the processes of communication, collaboration and learning. This effect depends on the instructional strategy that makes the computer support suitable for a specific context of use. The study that is discussed in chapter 9 also pays attention to the development of an instructional strategy that leads to an optimal use of the collaborative tools.

116

7 The Organization of Computer-

mediated Communication

In chapters 7 and 8, we discuss two studies that have been carried at the beginning of the research project. We deliberately planned these studies at the start as part of an exploratory phase. The exploratory studies prevented us from making any premature decisions with regard to the requirements specification. First, we studied the basic properties of the envisioned collaborative tools. It helped us to get a grip on a relatively new topic of networked learning in face-to-face situations.

Chapter 7 discusses one of the initial guidelines – parallel access as floor-control mechanism – that aims to stimulate equal participation and the free exchange of ideas. We present a study that investigates the effects of parallel access on the coordination of the communicative actions in the shared digital workspace. The outcomes of that study provided the ground for the development of two additional design guidelines – i.e. design guidelines 5 and 6 – that aim to make the sequence of computer-mediated acts more coherent.

7.1 Problem Analysis: Parallel Access as Floor-control Mechanism

As mentioned in chapter 6, it is expected that parallel access as floor control mechanism counteracts the ineffective communication patterns “interruption” and “product blocking”. It is expected that parallel access leads to a more even participation because group members can express their ideas without any direct interference of other group members. Parallel access allows users to share their ideas directly with the group without any delay or interruption. Users can express their ideas more freely. In sum, this should result in a more balanced pattern of participation.

117

Chapter 7

Parallel access opens up new possibilities for communication but it also creates new problems. Parallel access comes at some cost; it leads to new communication challenges for the group. These challenges have mainly to do with the coordination of the individual communicative actions.

Each floor control mechanism – turn taking and parallel access – has its own dynamics. They bring about their own specific opportunities and problems with regard to organization of interactions. Students in the parallel-access condition have more time available to perform actions, i.e. they do not have to wait for their turn but can put forward their contributions without any delay. This results in a sense of freedom to act, whereas working in the turn taking condition gives a feeling of being restricted. Parallel access as a permissive floor control mechanism leads to a greater number and a greater diversity of actions being performed in the shared workspace (Heeren, 1996). The group must manage this expansion and diversity of actions properly. They have to organize the constant flow of individual messages into a meaningful representation of their digital discussion.

Initial studies with parallel access as floor-control mechanism revealed that the number of contributions in the shared digital workspace expanded quickly. With parallel access, students hardy experienced any limitations with regard to the expression of their ideas. Students added many contributions in a relatively short period without any constraints. Consequently, the shared workspace of the collaborative tool became crowded with contributions and the students spent a lot of time organizing their contributions into a meaningful whole. The permanence of contributions even worsens the problems that students experience because all the contributions remained visible in the shared workspace.

The groups were faced with the challenge to organize their profusion of ideas into a meaningful whole. They must relate the individual contributions into a comprehensible representation that displays the line of reasoning of the group. The initial studies showed that the groups differed in respect to manner in which they achieved that task: some groups organized their interactions better than others did. These observations made it clear that the groups may benefit from some additional support that would guide their interactions in the shared digital workspace. In this chapter, we discuss the development of such kind of support. Therefore, we had to revisit research question three and four of chapter 1 that states:

118

The Organization of Computer-mediatedCommunication

Q3: How do the structural features of the medium relate to the ineffective communication patterns?

Q4: Which structures makes the desired patterns of communication more likely?

Research questions 3 and 4 were further specified by the following two questions:

1. What kind of principles do groups apply to organize their computer-mediated exchanges into a coherent and meaningful whole?

2. Could the insights, which we gained from answering the previous question, be used to improve the support? Moreover, how should such an improvement look like?

7.2 Study I: The Organization of Computer-mediated Communication

We carried out a study at a secondary school with fourth level students who varied in age between 15 and 17 years. The study involved one class with 7 groups of three students. The aim of the exploratory study was to investigate the computer-mediated communication patterns associated with parallel access. The students in our study took part in a geography course. The topic that the students had to address was the social image of a certain geographical region, in their case the view that people have of Salou, a popular holiday destination in Spain.

The learning method that the teachers applied was collaborative learning. Students discussed the topic in small groups of three students. The pedagogical objective of these discussions was to address the students' (pre)conceptions with regard to Salou as a popular holiday destination for young people. The students had to formulate arguments in favor or against the claim that this region was mainly a holiday destination for youngsters. Such an argumentative discussion is an important pedagogical strategy for knowledge construction (Mason, 1998).When a learner argues he or she constructs and put forth an argument that another learner will interpret and criticize, after which the first learner will respond, perhaps by revising the argument (Hitchcock, 2002). The notion of ‘winning’ an argument is completely irrelevant – if both counterproductive – to the primary goal of clearly articulating ideas (Schrage, 1991).

119

Chapter 7

Learning Environment

As mentioned at the introduction, the studies of chapter 6 and 7 have been carried out at the beginning of the research project before any software prototype was available. It means that we used an existing collaborative tool – the Digalo tool – to study the dynamics of parallel access. The Digalo tool was developed by the DUNES project, initially to support online collaborative learning but the tool has also been applied in face-to-face learning situations (Schwarz & De Groot, 2007; Overdijk & van Diggelen, 2008). The Digalo offers two modes of floor control: turn-taking and parallel access. In our study, we used parallel access as a means to access the shared workspace.

The Digalo Tool

The Digalo is a Graphical tool that offers its users a two-dimensional shared workspace that is based on a concept-mapping interface. Such an interface allows users to construct a spatial representation of concepts and their interrelationships. Users put forward their ideas in the shared workspace as textual objects and they visualize a relationship between two concepts by drawing a line between the two (see Figure 7.1).

The Digalo supports collaborative argumentation. The tool has a pre-defined notation system that labels the textual contributions. A user selects a specific label from the menu at the top of the window and attaches that label to the textual contribution, for example an argument in favor or argument against a proposition. The shared workspace represents the contribution as a card that has a distinct shape and color. Shape and color represents the specific label attached to the contribution.

A contribution only becomes visible for the other users after the user types in a text and presses on the “submit” button. The text that is displayed in the card is called the “title”. Users can also add an additional text associated in a comment window. The comment text is not directly visible in the workspace but it appears on the screen when a user selects the card.

Another feature of the Digalo is the possibility to link contributions to indicate a relationship between the two. Users can link two contributions that, in their view, are related by drawing a line between the two contributions.

120

The Organization of Computer-mediatedCommunication

The Learning Task

The study took place during two lessons of 45 minutes each. During the first lesson, the students received their task instructions, studied information that has to do with the task and they became familiar with the basic functions of the Digalo. In the second lesson, the students performed a role-play discussion. Role-play is one of the tasks that incorporate authentic activities based on a real-life situation (see chapter 5). The three roles that were defined beforehand are: 1) a youngster, 2) a parent with young children, and 3) an elderly person. As a preparation, the students received role-specific information about the popular holiday destination Salou.

The discussion in the Digalo focused on the central claim: “Salou means discos, beaches, parties and flirtation”. The argumentative diagrams that the students constructed with support of the Digalo visualized the students’ argumentation. During their discussion in the Digalo, students could choose between three types of contributions: 1) arguments pro, 2) arguments contra, and 3) information source.

The Analysis of Student's Actions within the Shared Workspace

The analysis of the collaborative activity focused on students’ ongoing actions within the shared digital workspace of the Digalo. The Digalo tool makes certain actions available for the students. We identified 10 different kinds of actions that the students

121

Figure 7.1: The user-interface of the Digalo tool.

Chapter 7

could perform with the support of the Digalo (Table 7.1). These 10 different actions enable the students to put forward contributions and to organize them into an argumentative diagram that represents their line of argumentation.

A sequential analysis of the students’ actions in the Digalo showed a typical pattern. Students put forward their ideas with the actions “Add shape” and “Add title”. When the students placed 4 or 5 shapes in the shared workspace, they began to organize the diagrams with the action 'Moved shape' that moves a contributions to a different location in the shared workspace. Students also linked contributions to indicate relatedness. In general, all groups added links between cards when the shared workspace contained approximately 9 shapes. As mentioned before, synchronous action in a shared workspace quickly resulted in a crowded diagram. When three group members submit contributions more or less simultaneously, it is hard to keep track of all entries, let alone to get an overview of the discussion. This is why most groups organized their diagrams when complexity increased.

122

Action Description

Add shape Add a shape in the drawing area.

Add title Add a title in the shape.

Add comment Add a comment with a title (the comment is not directly visible. After one double click on a shape a comment window will appear).

Add link Draw a line between two shapes that are associated together.

Move shape Move a shape within the drawing area.

Change title Change the title of the shape.

Change shape Change the shape, i.e. change the type of contribution.

Delete shape Delete a shape.

Delete link Delete a link.

Resize shape Change the size of the shape. For example, the user can enlarge the shape when a title is a few sentences long to make the whole title visible.

Table 7.1: Actions that can be performed in the shared environment.

The Organization of Computer-mediatedCommunication

Three Organization principles

Students used a lot of time organizing their contributions into a comprehensible diagram. Some groups managed to coordinate their actions into a clear direction; they applied a clear principle that guided the actions in the shared workspace. The diagrams of these groups looked organized. They came up with more rigidly structured diagrams. Other groups failed to do this, their diagrams remained scattered. These groups constructed less structured, extensive diagrams that may look chaotic for an outsider. An analysis of the structure of diagrams revealed that the students applied three principles to organize the contributions that were placed in the shared workspace (van Diggelen, Overdijk & Andriessen, 2004). These principles are (Figure 7.2):

– type of contribution,

– a link between two contributions,

– the spatial position of the contributions.

Type of Contribution

First, users can organize a diagram with the support of a notation system that attaches a label to a contribution. These labels are visualized in the shared workspace by their shape and color. They give additional meaning to a contribution. For example, the notation system that we applied made a distinction in shape and color between arguments in favor or against. These different types of arguments are clearly recognized in the shared workspace. For example, one can see at first glance if an argumentative diagram contains many arguments against (a red color) or in favor (a green color).

123

Figure 7.2: Three principles for organizing a Digalo diagram.

Chapter 7

A Link between Two Contributions

Secondly, the use of a link helps students to indicate coherence and meaning. A user adds a link between two cards to visualize relatedness. Links make it possible to organize the sequence of messages differently; it exceeds the temporal order associated with verbal exchanges where coherence is based on turn taking and adjacency pairs (see Chapter 5, design guideline 7). The principle of linking cards enables users to go beyond the limits of a temporal sequence of ordering. Users organize their contributions based on related meanings so that they create a logical order that reflects their line of reasoning.

The Spatial Position or Grouping of a Contribution

Users do not always use a link to indicate a relationship between two contributions. Sometimes they position contributions near each other to indicate relatedness. In that case, users fully use the spatial characteristics of the two-dimensional workspace. As a result, they create meaningful areas that contain cards that belong to each other. Cards that are in close proximity suggest similarity, association, inferences or causality. This principle for organizing the contributions based on closeness is called 'spatial grouping'.

Organization Principles and the Appearance of the Diagrams

During the first studies, students worked within the two-dimensional workspace of the Digalo without hardly any restrictions. They could access the shared workspace simultaneously without any temporal or spatial constraints. Furthermore, students did not receive any specific instruction with regard to the organization of the diagram. This led to diversity in appearances of the argumentative diagrams: the patterns and number of contributions differed to a large extend.

Some groups constructed more rigidly structured diagrams with a clear pattern. However, the leading pattern differed between groups. We observed, for example, that one group constructed a diagram that represents one line of reasoning (diagram 1, Figure 7.3). Contributions were places beneath and next to each other in a logical order. A second group constructed a diagram that emphasized opposing standpoints (diagram 2, Figure 7.3). Arguments in favor and arguments against were separately grouped together in the left side and the right side of the workspace. The maps that displayed a more rigid structure can be characterized by the fact that all the

124

The Organization of Computer-mediatedCommunication

125

Figure 7.3: Schematic examples of Digalo diagrams.

Chapter 7

organization principles strengthen the representation of one leading pattern. For example, in diagram 2 the arguments in favor and against were: 1) linked, 2) grouped together spatially, and 3) recognizable by their shape and color.

The more unstructured diagrams lacked one clear leading pattern. From diagram 3 (Figure 7.3) one can see that the students used two of the organizing principles: 1) different types of contributions, and 2) links between the contributions. What the diagram lacked was the use of the third organization principle: the spatial grouping of contributions. It resulted in rather complex diagrams. This became even more apparent when the number of contributions increased.

Conclusion

To summarize, observations suggest that the diagrams became crowded and scattered if the groups did not use the third organizing principle of spatial grouping. By far the most common method for indicating relatedness was to position contributions that are related close to each other. This observation is consistent with the studies about functional spaces that were discussed in chapter 6. Dwyer and Suthers (2006), for example, observed that spatial proximity was used to group contributions into meaningful threads. If the groups did not appropriate the spatial dimensions of the shared workspace, it became less like that they could organize their interactions into a coherent whole. This insight served as inspiration for the development of design guidelines 5 and 6 (Table 7.2). Table 7.2 describes the two design guidelines in the form of the pattern language (Alexander, 1979) that has been introduced in chapter 2. The description makes a distinction between problems, solutions and the context within which the problems and solutions emerge.

7.3 Study II: A Refinement of the DesignDesign guideline 5 and 6 aim to solve the coordination problem that users experience when they can access the workspace simultaneously. Design guideline 5 assumes that students are better able to coordinate their actions when the shared workspace is divided into functional spaces. Design guideline 6 states that these functional spaces should have a distinct meaning. The guidelines assume that it will be less likely that students act arbitrary when the workspace is divided into functional spaces relevant for their task performance. The tool must offer the students a meaningful structure that is

126

The Organization of Computer-mediatedCommunication

related to the task that the students have to carry out. Such a structure or representational aid guides the joint actions of the group members.

The Implementation of the New Design Guidelines

Design guideline 5 states that functional spaces provide the users with a predefined structure that facilitates the coordination of their actions in the shared workspace. Design guideline 5 was implemented in the graphical tool as a grid that divided the shared workspace into four areas (Figure 7.4). As indicated by design guideline 6 these areas should have a meaning in relation to the learning task. Actually, this design guideline makes design guideline 5 suitable for a specific context of use.

As mentioned in the previous paragraph, the students had to discuss the social image of Salou from three different perspectives, i.e. of 1) a youngster, 2) a parent with young children, and 3) an elderly person. They should formulate arguments in favor or against the claim that “Salou means discos, beaches, parties and flirtation” from the perspectives of those three roles. The three student roles were taken as a reference; they were associated with a meaningful area that we called the “individual role areas”. All students could access these areas; however, the students received the instruction to start

127

Design pattern for guidelines 5 and 6

Problem A diagram becomes crowed in a relatively short time when users put forward contributions simultaneously. As a consequence, users spend a lot of time organizing their contributions in the shared workspace into a coherent whole.

Solution It is hypothesized that a structure that divides the workspace into functional spaces helps users to organize the diagram into a coherent whole. It makes the users aware of the spatial properties of the shared workspace as a means to the structure their shared representation. These functional spaces have a meaning relevant for the problem or task at hand. They display basic aspects of the problem space.

Context The implementation of the two design guidelines depends on the learning task. These functional spaces have a meaning relevant for the problem or task at hand. They display some basic properties of the problem space. This may vary from situation to situation.

Table 7.2: Design guidelines 5 and 6.

Chapter 7

working in the area associated with their role. The fourth quadrant was appointed for a general discussion.

The use of a grid provided the students with clear directions how to organize their joint actions in the shared workspace. It was expected that the students should spent less time organizing their communicative exchanges into a coherent whole because they were provided with a structure relevant for their learning task. We used a semi-experimental design to study the effects of design guideline 5 and 6. Two conditions were created: one condition where a grid was displayed as a representation aid and one condition where the shared workspace did not contain a grid.

The Analyzing of Students' Actions with and without a Grid

The groups that participated in the study were divided over two conditions (Table 7.3). The groups in both conditions received the same learning task and preparation. All the groups used the Digalo to share and organize their arguments in favor or against the claim. The groups differed with regard to the appearance of the shared workspace. In condition one (H4a) the students used the shared workspace without a grid as representation aid. The students in this condition discussed the topic in a ‘default’ shared workspace with no additional structure. In condition two (H4b) design guideline 5 and 6 were applied: the shared digital workspace was divided into four functional areas with a distinct meaning attached to them. The shared workspace had a grid structure that divided the shared workspace into 4 meaningful areas (Figure 7.4).

128

Figure 7.4: The shared workspace divided into four functional spaces.

The Organization of Computer-mediatedCommunication

Each condition involved one class: a class with 21 students (H4a) and a class with 22 students (H4b). Students were randomly appointed to groups of three persons. It is hypothesized that the coordination of the joint actions would differ for the two conditions. The division of the shared workspace into function spaces would help the students to organize their contributions into a meaningful way. The grid structure makes the students aware of the spatial dimensions of the workspace. It is expected that students use this knowledge to coordinate their actions and locate contributions in the appropriate area.

Analysis

Table 7.4 and Table 7.5 give the frequencies of the different actions (see Table 7.1) that the students can perform in the shared workspace of the Digalo. The actions “Move

129

Condition 1 (H4a) Condition 2 (H4b)

Lesson 1 Students became familiar with the basic functions of the tool

Students became familiar with the basic functions of the tool

Lesson 2 One ‘default’ shared graphical workspace with no additional structures

A shared graphical workspace divided into four meaningful areas

Table 7.3: Two conditions to study the effects of design guideline 5.

Actions in the shared digital workspace H4a (condition 1)

Add

shap

e

Add

title

Add

link

Com

men

t

Mov

e sh

ape

Cha

nge

title

Cha

nge

shap

e

Del

ete

link

Del

ete

shap

e

Resiz

e sh

ape

Org

anize

Group 1 15 15 23 0 42 2 0 0 0 2 65

Group 2 19 13 5 0 50 2 0 0 2 10 55

Group 3 12 11 7 1 55 2 0 0 1 11 62

Group 4 21 21 2 0 40 2 0 0 0 13 42

Group 5 10 9 9 7 60 1 0 0 0 1 69

Group 6 17 16 15 0 48 0 1 1 1 2 63

Mean 16 14 10 1 49 2 1 0 1 7 59

St. dev. 4.2 4.2 7.7 2.8 7.6 0.8 1.2 0.4 0.8 5.4 9.6

Table 7.4: Frequencies of Digalo-mediated actions (Condition 1).

Chapter 7

shape” and “Add link” were added up and labeled “Organize”. Organize refers to the actions of students aimed at arranging the diagram into a coherent and meaningful whole.

A large part of the students’ actions dealt with organization of the contributions (59% of the actions for H4a and 37% of the actions for H4b). Parallel access in the shared digital workspace quickly filled up the workspace with a large number of contributions. That is why most groups started to organize the diagram when the number of cards increased; and they kept on doing this. On average, the groups in condition 1 who did not worked with functional spaces (M=59, SE=3.9) spent significantly more time organizing the diagram then students who worked in condition 2 with a workspace divided into meaningful areas (M=37, SE=3.6, t(9)=4, p<.05). On the average, 59 actions in conditions 1 had to do with the organization of the

130

Actions in the shared digital workspace H4b (condition 2)

Add

shap

e

Add

title

Add

link

Com

men

t

Mov

e sh

ape

Cha

nge

title

Cha

nge

shap

e

Del

ete

link

Del

ete

shap

e

Resiz

e sh

ape

Org

anize

Group 1 31 28 26 1 11 0 0 0 2 2 37

Group 2 30 22 12 0 18 6 0 0 4 8 30

Group 3 41 21 7 0 21 0 2 1 2 4 28

Group 4 29 20 8 0 35 0 0 0 7 0 43

Group 5 20 14 5 8 42 2 0 1 5 2 47

Mean 30 21 21 2 25 20 0 0 4 3 37

St. dev. 7.5 5 5 3.5 12.7 2.6 0.9 0.5 2.1 3 8.2

Table 7.5: Frequencies of Digalo mediated actions (Condition 2).

Figure 7.5: Digalo-mediated actions (Condition 1).

Add shapeAdd title

Add linkComment

Move shapeChange title

Change shapeDelete link

Delete shapeResize shape

0102030405060

Mea

n

The Organization of Computer-mediatedCommunication

contributions into a meaningful representation. The groups in condition 2 only used 37 actions for the organization of the diagram. Moreover, the students in condition 1 (M=14, SE=4.2) put forward significantly less contributions than the students in condition 2 (M=21, SE=5.0, t(9)=4, p<.05). Figure 7.5 and Figure 7.6 graphically represents the average frequencies of the different Digalo-mediated actions for both conditions. It clearly shows how the dominant course of actions moves from organizing the diagram in condition 1 towards putting forward contributions in condition 2.

7.4 DiscussionIn this chapter, we studied three design guidelines – design guidelines 4, 5 and 6 – that were developed during two successive cycles of problem identification, abstraction, theory formulation, implementation and evaluation (Figure 7.7).

First, we identified interpersonal dominance as a group behavior that leads to unequal participation during a group discussion. Dominance as an ineffective communication pattern was associated with turn taking as floor-control mechanism. We hypothesized that parallel access as an alternative model for access leads to a more balanced participation because group members express their ideas without any hindrances. We lay down this hypothesis as design guidelines 4, which rounded off the first research cycle. The implementation of guideline 4 led to a second research cycle that focused on the coordination problem that students faced when their communication is based on parallel access.

Students found it difficult to coordinate their joint actions and to come up with a diagram that represents their contributions into a meaningful whole. An analysis of the diagrams indicated that the spatial grouping of contributions made the digital

131

Figure 7.6: Digalo-mediated actions (Condition 2).

Add shapeAdd title

Add linkComment

Move shapeChange title

Change shapeDelete link

Delete shapeResize shape

0102030405060

Mea

n

Chapter 7

interactions more coherent. This led to the development of two additional design guidelines: guideline 5 and 6. These guidelines state that meaningfull areas help users to organize their contributions into a coherent and meaningful whole. We evaluated this hypothesis in a study where the groups were divided over two conditions. In one condition, the use of meaningful areas was implemented as a grid, while in the other condition the groups did not use such a representational aid. Results indicated that the use of the grid significantly reduced the coordination problem that groups experienced in the shared digital workspace. This reduction had a positive effect on the discussion: students put forward substantial more contributions.

132

Figure 7.7: Two successive cycles of design.

8 Computer-Mediated Communication

and Task Performance

The envisioned collaborative tools aim to support groups who carry out a problem-solving discussion in the classroom. The tools mediate part of the group communication. It is expected that the tools create the proper conditions for groups to learn. This is not self-evident because the groups already communicate verbally. They can talk with each other but should also use a collaborative tool as a medium for communication. The two opportunities for face-to-face communication – verbal and computer mediated – raise the question which communicative functions should be computer mediated. Design guideline 2 addresses this issue. The distinction between task-related and social-emotional communication is used as a criterion “to split” the communication between the two media. Design guideline 2 states that the computer-mediated communication must contribute to effective task performance. This guideline provides us with a general direction. In chapter 6, we further elaborated on the requirement that the computer-mediated communication should mirror effective task performance. The collaborative tool should support the groups so that they will carry out their learning task effectively. Group learning is most likely to occur when the students share knowledge about the task and when they elaborate on the knowledge that is shared (Hogan, Nastasi & Pressley, 2000). We developed two additional guidelines – design guidelines 7 and 8 – that are based on the two criteria for a constructive dialogue. These two criteria emphasize coherence and elaboration. Table8.1 describes the two design guidelines that should stimulate groups to carry out a coherent discussion that is oriented at knowledge elaborations.

In this chapter, we present a study that further elaborates on design guideline 7 and 8 that aim to establish a proper fit between the characteristics of the collaborative tool and the communicative demands set by the learning task. The study addresses

133

Chapter 8

research questions 4 that can be stated as follows:

Q4: How do the structures of the medium organize the computer-mediated part of the communication?

8.1 Computer Support for Task-related CommunicationThe task-related communication is considered as the gist of group learning, difficulties with regard to the task-related communication will have a direct effect on the learning achievements. It does not mean that the other functions of communication are less important. Task-related communication should be balanced by communication that is directed towards the relationship between group members. Social-emotional communication is crucial for the group’s well being. The quality of these expressive

134

Designguideline

Description

7 A mediated discussion should be based on global principle of coherence. The use of “connections” enable students to respond to a previous contribution that does not directly precede in time. Users can relate contributions based on a general structure of meaning instead of a purely temporal structure. It is expected that the students experience more “cognitive freedom” with they discuss a topic based on a global structure of coherence.

8 It is expected that a notation system that structures the communicative acts that students can display in the shared workspace stimulates the occurrence of certain behaviors or cognitions that are beneficial for group learning. A structured- communication notation system displays a limited range of acts that serve semantic guidelines during a problem-solving discussion.

With a post-hoc notation system students label a contribution after it has been placed in the shared workspace. It triggers student to reflect on the content of the discussion.

Table 8.1: The guidelines that are associated with task performance.

Computer-Mediated Communication andTask Performance

interactions in groups may be very important in influencing the attitudinal and affective consequences of being in a group (Guzzo & Shea, 1992).

The collaborative tools support the task-related interactions because they are directly related to learning (Draskovic et al., 2004). The tools should stimulate groups to carry out their task-related communication in the shared digital workplace. Design guidelines 7 and 8 (Table 8.1) point out how the task-related communication should look like. These guidelines refer to two criteria for a constructive dialogue: 1) the sequence of task-related communication should be coherent, and 2) the task-related communication should display patterns of knowledge elaboration.

Coherence

Communication is more than a collection of random sentences (Knott & Sanders, 1998). Groups have to organize their individual talk into a coherent whole. A speaker must relate his or her contribution to what has been said previously so that a meaningful discussion emerges (Pavitt & Kline Johnson, 1999). Design guideline 7 addresses this issue of coherence. The guideline states that the use of connections enables students to respond to contributions that do not directly precede in time. It “overrules” the principle of adjacency pair or “nextness” that is the leading principle to relate successive verbal contributions during a verbal discussion. Design guideline 7 offers an alternative structure to organize communicative exchanges into a coherent whole. Students can directly react on all the contributions that have been put forward in the shared workspace. This provides students with more freedom of expression; they are no longer confined to a temporal order where topics follow each other.

Design guideline 7 cannot be set apart from other structural features of the collaborative tool. The different characteristics of the tool make it possible to organize a discussion based on a global structure of coherence. These characteristics are:

– the permanence of contributions in the shared workspace,

– a shared workspace divided into functional spaces,

– a division of the shared workspace based on an information structures that is associated with relevant aspects of the problem or the task, and

– connections as a means to relate contributions.

135

Chapter 8

A global structure of coherence seems to be a more proper method for organizing the communication that uses parallel access as floor control mechanism. It offers the group a structure that helps them to organize their parallel talk into a coherent whole.

Elaboration

Design guideline 8 states that a notation system stimulates the elaboration of knowledge shared by the group. A notation system offers student semantic guidance; it displays a limited number of suggestions for the kind of statements that the students can put forward in the shared workspace. The study that is discussed in this chapter analyzes the effects of a notation system that contains two options: question and comment. It is supposed that “asking a question” and “making a comment” would stimulate a constructive dialogue between the students. Asking a question and giving a comment have the function to elicit a specific response; they encourage students to elaborate on a contribution. A question is more explicit in triggering such a response. A comment expresses an opinion or attitude that, on it turn, stimulates further elaboration. As a response, students might give examples to explain an idea, provide evidence for a statement or give reasons as grounds for a conclusion. The function of asking a question is to elicit a verbal response from those to whom the question is addressed (Keatsley, 1976). Questions may serve several functions that depend on the context of the interaction (Hargie & Dickson, 2004). For example, a question might encourage students to elaborate further on a statement. Research indicates that students who were instructed to ask high-level thought-provoking questions elicited more knowledge-construction responses. These questions lead to explanations, inferences, speculations, and other such elaborated responses, which have a direct positive effect on individual achievement. Questions encourage students to elaborate on existing knowledge, which, on its turn, facilitates the acquisition of that knowledge (King, 1999; King, Staffieri & Adelgais, 1998; King, 1994).

8.2 The Learning Environment

The research has been carried out at the secondary school with 5th grade students who attended a Dutch language course. These students followed a series of six lessons that had as central theme “Respectful discourse in the classroom”. The learning goal of the lessons was to develop the argumentative skills of the students. The students had to

136

Computer-Mediated Communication andTask Performance

develop well-founded solutions that would encourage a respectful discourse. Small-group discussion was considered as an appropriate and useful method for practicing argumentation. The 21 students worked in small groups of two or three persons. They had to formulate a solution for the social issue, share their solution with their group members and discuss the assumptions and implications of their solutions.

We studied the effects of design guidelines 7 and 8 in a study where the students used the Digalo tool to discuss a problem. Insights obtained from this study should provide us with valuable insights about the effects of these design guidelines on the task-related communication.

The Learning Task

The central theme of the lessons was “Respectful discourse in the classroom”. During the first lesson, the teacher introduced the topic and explained to the students which products they had to deliver for evaluation. The students had to deliver two products:

– A policy note for the school’s board of directors. The policy note should contain an advice about how to promote a respectful discourse in the classroom. The policy note should be a group product.

– An argumentative text, written individually.

As preparation, the students watched two short videos about the topic. One video about a school program that aimed to solve conflicts between pupils through mediation, and one video about a discussion between students, teachers, experts and politicians about respect.

The students carried out a number of collaborative activities before they start writing the policy note. During the second and third lesson, the students worked in small groups on two assignments that should advance their argumentative skills. The second lesson consisted of an argumentative discussion that the students had to carry out with the support of the Digalo tool. This activity was similar to the one discussed in chapter 7. Each group had to formulate a proposition like: “The teacher is responsible for respectful discourse in the classroom”. During the argumentative discussion, the group members formulated arguments in favor or against the proposition. They did not have to take a position; it was important that their arguments covered various perspectives of student, teacher and parent. During the third lesson, the students had to

137

Chapter 8

develop solutions that would promote a respectful discourse in the classroom. This activity served as a preparation for lesson four, where the students were expected to write a policy note for the school's board of directors. That policy note should contain a set of solutions. During the remaining lessons, the groups had to write the policy note. In this policy note, the students have to offer an advice to the board of directors.

The study that is discussed in this chapter is about lesson 3 where students used the Digalo tool to formulate and discuss solutions. Student had to propose a solution for the social problem and discuss their proposals with the group members.

The Digalo

Students formulated and discussed their proposals with the support of the Digalo tool. The Digalo tool offers the students a number of meaningful areas that support the coordination and organization of the communicative exchanges in the shared workspace. The shared workspace was divided into three areas that have a distinct meaning; each area is associated with a specific solution (Figure 8.1). Each student was allocated to such an area to put forward their solution. The students were instructed to take some time to formulate their proposal in their own area before they start to elaborate on the various proposals suggested by fellow group members.

The Digalo had a notation system to stimulate knowledge elaboration. The students were presented with three options to label their contribution: 1) solution, 2) question and 3) comment (Figure 8.2). Students had to choose one of these labels before they put forward a contribution in the shared workspace. They could use the label

138

Figure 8.1: Meaningful areas.

Computer-Mediated Communication andTask Performance

“Solution” to submit a solution, and the other two labels to elaborate on the proposed solutions.

Instructions

At the beginning of the discussion, the students were encouraged to work in their own area and to take some time to formulate a solution. These solutions might serve as the starting point for a discussion during which the group assessed their solutions critically. By dividing the shared workspace into separate quadrants, where each quadrant contains one solution from a specific student, we provided the students with a meaningful structure or representation aid that directs the communication in the shared workspace.

8.3 Analysis I: The Content of the Computer-mediated Interactions

There are at least two ways to look at the communication. One can look at the content and analyze what has been put forward by the students or one can focus at the process and analyze the sequence of communicative acts of the various group members. In this study we did both. First, we analyzed the content to understand what kind of reasoning went on in the shared environment of the Digalo tool. Secondly, we analyzed how the individual actions in the digital workspace were organized into a pattern of related communicative acts.

139

Figure 8.2: The notation system of the Digalo.

Chapter 8

Coding Scheme

Various authors developed coding systems to analyze computer-mediated communication (e.g. Henri, 1992; Newman, Johnson, Webb & Cochrane, 1997; Pena-Shaff and Nicholls, 2004). Henri (1992) was one of the first researchers who developed a framework for analyzing computer-mediated interactions between students. She identified five dimensions of the learning process:

– Participation: The participation dimension gives an overview of student participation indicated by the number of messages or statements made by the students.

– Interaction: The interaction dimension makes a distinction between explicit and implicit interactions. Explicit interactions are statements that refer specifically to one or more messages while implicit interactions do not specifically mention the connection.

– Social: The social dimension contains all statements that are not related to the formal content of the subject matter.

– Cognitive: The cognitive dimension is used to evaluate reasoning which uses critical thought. Critical thought is measured by criteria that distinguish surface processing from in-depth processing.

– Metacognitive: The metacognitive dimension is divided into metacognitive knowledge and skills Metacognitive knowledge refers to declarative knowledge concerning the person, the task, and the strategies. Metacognitive skills have to do with procedural knowledge relating to evaluation, planning, regulation and self-awareness.

Newman et al. (1997) combines Henri’s (1992) framework with Garrison’s (1992) model of critical thinking that considers critical thinking as a sequential problem solving process with five stages: problem identification, problem definition, problem exploration, problem applicability and problem integration. Pena-Shaff and Nicholls (2004) developed an extensive category system to analyze knowledge construction that assesses sophisticated cognitive skills such as clarification, conflict, assertion, judgment and reflection that appear to be most directly related to the process of knowledge construction.

140

Computer-Mediated Communication andTask Performance

141

Table 8.2: Coding scheme for Digalo-mediated communicative acts.

Chapter 8

It was difficult to use an existing coding framework for the analysis of the actions and interactions within the Digalo environment. For example, the frameworks discussed above were developed for asynchronous computer-mediated communication. Furthermore, they are associated with cognitive acquisition-oriented perspective that assesses knowledge construction as an individual activity, fed by a social context (Veldhuis-Diermanse, 2002). No existing coding framework matched with the data, so we decided to develop our own classification scheme (Table 8.2).

To analyze the content of the discussion we focus on the sequence of related pairs where an utterance is followed by a response. An obvious action-response example is a “question-answer” relationship. However, an utterance can also elicit a response in a less straightforward manner. For example, a comment that expresses an opinion can trigger a response in the form of a further elaboration of the initial statement. It means that we only coded the student’s actions that are a part of an interaction sequence because these actions are indicative for the co-construction. The diagrams that represent a sequence of related actions were used to (re)organize the data for coding (see Figure 8.5).

We identified six categories of actions that can be associated with the in-depth elaboration of knowledge (see King, 1994; Hargie and Dickson, 2004; Pena-Shaff and Nicholls, 2004). These categories are analysis, inference, judgment, evaluation, application, comparison and contras, and conflict. These categories where associated with the two labels of the notation system – question and comment – that should trigger elaboration. This results in 12 categories. Each sentence that students put forward in the Digalo – and that is a part of sequence or related responses– was coded by two coders (interrater reliability 0.9). We choose for the sentence as unit for analysis because students sometimes change theme within a contribution. For example, they ask a question but also comment upon the question. Sentences that were divided by the words “but” or “for” were considered as two separate sentences because these words also reflected a change in focus.

Results

Table 8.3 gives the amount of sentence for each category of the coding scheme. It is noteworthy to notice that all students’ utterances were task-related. There were no statements that could be typified as social-emotional. Figure 8.3 presents the percentages of sentences identified in each category. If we look at “questioning” as a communicative act for in-depth elaboration that we may conclude that students mainly

142

Computer-Mediated Communication andTask Performance

asked “specifying questions” (18% of all statements) that encouraged respondents to examine an idea in more detail. The comments that students gave are more diverse: students have more detailed account of their ideas (18 % Cspe), they provided reasons, evidence or arguments on which they based their ideas or thoughts (22% Cinf ) or they expressed an opinion or made value judgments (22% Cjud).

One can conclude that the students used the Digalo to remove uncertainty within the student groups that is caused by ignorance or imprecision of a shared interpretation of the situation. The communication was directed at acquiring new information that enables them to form a more precise interpretation of the proposal.

143

Qspe Qinf Qjud Qapp Qcom Qcfl Cspe Cinf Cjud Capp Ccom Ccfl

Group 1 5 1 0 1 1 0 5 18 9 1 1 0

Group 2 7 3 1 1 0 0 3 8 11 1 2 0

Group 3 5 1 0 4 0 0 6 4 3 5 0 0

Group 4 9 0 0 0 0 0 10 9 12 1 0 0

Group 5 4 0 0 0 0 0 4 1 3 0 0 0

Group 6 3 0 0 1 0 0 4 16 8 7 0 2

Group 7 2 0 0 1 0 0 2 1 2 2 0 1

Total 35 5 1 8 1 0 34 57 48 17 3 3

Table 8.3: Number of statements indicating knowledge construction.

Figure 8.3: Frequencies for different types of task-related messages.

Qspe Qinf Qjud Qapp Qcom Qcfl Cspe Cinf Cjud Capp Ccom Ccfl0

10

20

30

40

50

60

Contribution type

No.

of c

ontri

butio

ns

Chapter 8

Discussion

The Digalo tool in combination with verbal interaction led to a specific kind of problem-solving discussion. Only the task-related interactions were mediated by the tool. The utterances expressed in the Digalo concerned the topical content of the discussion. This in contrast to the findings with regard to online collaborative learning, where a considerable amount of the messages is of a social-emotional or meta-cognitive nature (e.g. Pena-Shaff and Nicholls, 2004; Hara, Bonk and Angeli, 2000). This difference in findings may be due to the fact that in our research the students were co-located and could also communicate verbally. In our study, the students sat next to each other. We observed that the students also communicated verbally; for example, they asked for help, discussed their planning or decided when to end the Digalo discussion. It seems that the communication mediated by the tool only concerned messages with respect to the content. An analysis of the recorded face-to-face discussions revealed that this mode of communication was used infrequently. Students could be silent for 2 to 3 minutes. When they did communicate verbally, their utterances referred to:

– social-emotional aspects of the collaboration, for example asking for help, tension release by telling a joke, giving positive feedback, keeping group members focused on the task;

– planning of the activities, for example discussing the assignment;

– regulative aspects of the collaboration, for example discussing rules for computer-mediated interactions.

8.4 Analysis II: The Sequence of Task-related InteractionsIn this paragraph, we examine the patterns of related communicative acts that emerge in the shared digital workspace. We describe how the groups organize their communicative exchanges digitally. The analysis takes into account the structural features of the Digalo that give rise to certain rules for regulating the communication. These structural features are: 1) parallel access as floor-control mechanism, 2) permanence of contributions, 3) a shared workspace divided into functional spaces, 4) a predefined information structures (meaningful areas) to organize the computer-mediated communication, and 5) the ability to link contributions to indicate a relationship between the two. These characteristics help users to organize their

144

Computer-Mediated Communication andTask Performance

individual messages into a coherent and meaningful discussion.

Communication patterns

A first look at the diagrams that the groups created revealed that all students started to work in their own area where they put forward a solution. Furthermore, their fellow group members elaborated on these solutions by asking questions and giving comments. Most groups created a coherent diagram based on the meaningful areas – the Grid – that were presented to them. It triggered the students to use spatial grouping as a leading principle for organizing their contributions.

Time to Think

We divided the shared workspace into a number of meaningful areas. Each area in the two-dimensional workspace represented a solution of one of the group members. At the beginning of the session, students were instructed to put forward a solution in their own area. The teacher stressed that the students should take ample time to formulate a solution. They were encouraged to work in their own area before they would participate in the group discussion. Table 8.4 represents the time between the moment that the students started to work in the shared workspace and the moment they participated in the discussion. On the average students worked individually during the first 10 minutes (M=10:10 min./sec., SE=3:38 min./sec.). During that period, students

145

Student A Student B Student CTime (min:sec) Time (min:sec) Time (min:sec)

of f

irst

cont

ribut

ion

of f

irst r

espo

nse

to fo

rmul

ate

a so

lutio

n

no.

of s

ente

nces

of f

irst

cont

ribut

ion

of f

irst r

espo

nse

to fo

rmul

ate

a so

lutio

n

no.

of s

ente

nces

of f

irst

cont

ribut

ion

of f

irst r

espo

nse

to fo

rmul

ate

a so

lutio

n

no.

of s

ente

nces

Group 1 00:48 17:16 16:28 11 04:30 12:58 8:28 11 01:39 15:08 13:29 7

Group 2 05:43 12:08 06:25 5 04:38 16:14 11:36 9 09:14 12:08 02:54 4

Group 3 04:33 14:15 09:42 1 02:31 13:36 11:05 1 09:21 13:36 04:15 1

Group 4 02:25 11:59 09:34 6 03:38 14:37 10:59 9 03:11 15:02 11:51 5

Group 5 02:44 25:30 22:46 12 02:25 29:26 27:01 12

Group 6 04:06 12:50 08:44 5 04:55 13:56 09:01 4 04:07 12:50 08:43 2

Group 7 00:22 15:46 15:24 8 01:29 15:46 14:17 5

Table 8.4: Time to formulate a proposal at the beginning of the session.

Chapter 8

did not interact with each other in the shared workspace. Table 8.4 shows that the groups differed from each other for the time they needed to work individually on a solution. Group members within a group also showed some variation. For example, in the case of group 1 and 2 there is a difference of 8 minutes between the moment that the first and the last student participated in the group discussion. This would indicate that the students felt hardly any pressure to participate in the discussion when other students did so.

The number of sentences that students used to write down a proposal indicates how detailed the description is. Within the groups, there is little variation between the number of sentences that the students produced. While some group members took more time to formulate a solution, the detailedness of the solutions did not differ much.

The Pattern of Related Communicative Acts

The analysis focuses on sequence of related communicative acts that might be a sign for knowledge elaboration. A sequence consists of minimal three related actions. Weick (1979) defines such a sequence as a double interact. An action by actor A evokes a specific response in actor B (so far this is an interact), which is then responded to by actor A (this complete sequence is a double interact). We broaden this definition to a group of at least three students. The minimal amount of actions to make up a sequence consists of at least 3 related contributions within the shared workspace. Furthermore, the successive contributions are made by two different group members (see Figure 8.4). To identify the interaction sequences we used two of the three organizing principles – link and spatial grouping – that students used to organize their contributions (see chapter 7).

The seven groups produced 37 double interacts. Four times two successive contributions were made by the same student. Group 6 is responsible for three of these “successive contributions made by the same student”. Table 8.5 displays the length of

146

Figure 8.4: Actions that make up a meaningful sequence.

Computer-Mediated Communication andTask Performance

the sequences as the number of communicative acts – i.e. contributions or messages – made within the sequence. Most sequences are relatively short; they contain three to five contributions. Three sequences contain 10 related contributions.

They seven groups produced 6 single interacts, i.e. interaction sequences that only contain two related contributions. These single interacts were excluded from further analysis (12,5 % of all contributions). It is worth to notice that the fast majority messages in the shared workspace were part of a sequence of related messages of several students. It means that the digital workspace with its functional spaces is really a medium for communication.

We took the proposed solution as a criterion to formulate meaningful areas in the shared workspace. Each quadrant in the shared workspace contained a solution. Table8.6 shows the number of contributions for each functional area (quadrant). All the members of the groups 1, 2 and 6 formulated proposals that were discussed by the group. Only two of the three members of the groups 3 and 4 formulated a solution. These two solutions were discussed by all the group members. Groups 4 and 7 only had two group members. These groups also entered into a discussion but put forward a smaller amount of contributions.

147

No. of related contributions within a sequence

No. of sequences

3 14

4 9

5 7

6 2

7 0

8 2

9 0

10 3

Table 8.5: The “length” of a sequence.

Chapter 8

Interaction Sequences

Figure 8.5 shows a graphical representation of the sequence of related contributions. These figures display the contributions on a time axis; the contributions that are related are represented on the same horizontal level. The first contribution is the proposed solution. A line between two contributions indicates a link between the two contributions. The absence a line means that the students used the principle of spatial grouping to signify relatedness.

Figure 8.5 shows that most interactions sequences occurred parallel in time. These interaction sequences can be considered as focused discussions within in a general discussion. These focused discussions were named discussion lines. Group 2, for example, constructed six discussion lines in parallel (2.1, 2.2., 2.3, 2.4, 2.5, 2.6).

The digital workspace represent the history of their discussion, the contributions remained visible throughout the session. It was easy for group members to return to something that has been “said” earlier. The groups created a number of focused discussions – within the general discussion – that remained active.

The number of discussion lines for the seven groups varied between two and six. Groups of two students produced less discussion lines than groups of three students. The greater part of the sequences – 96 % of all sequences – had a typical pattern: a group member put forward a solution; another group member reacted on that proposal after which the group member who had put forward the solution gave a reaction. It seems that the “owner” of the solution keeps track of the questions or comments raised by other group members. Subsequently, these questions and comments triggered a reaction from the owner of the solution. These findings combined with the results of the first analyzes suggest that the groups display short sequences of knowledge elaborations.

148

Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7

N=3 N=3 N=3 N=3 N=2 N=3 N=3

Quadrant 1 6 8 0 0 3 19 6

Quadrant 2 5 6 10 13 4 9

Quadrant 3 7 8 9 5 6 6

Total 18 22 19 18 7 34 12

Table 8.6: Number of contributions that are part of a sequence.

Computer-Mediated Communication andTask Performance

149

Figure 8.5: Sequence of related contributions.

Chapter 8

Spatial behavior

In the previous paragraph, we already noticed that several discussion lines existed in parallel and that these discussion lines remained active for a long period. The discussion lines are situated within the different meaningful areas of the shared workspace. It means that students moved through the different areas of the workspace to contribute to the different focused discussions. This spatial behavior is displayed in Figure 8.6.

An analysis of students’ pattern of participation revealed that they constantly switched between discussions lines. Students “jumped” from one quadrant to another; added a contribution to the discussion within that quadrant and moved on to the next discussion. Students were less constrained to one course of action. They easily changed the subject of the discussion by moving to another discussion line. This in contrast to oral, face-to-face discussions, where only one topic is prominent for all group members. It seems that a general structure – i.e. meaningful areas that represent relevant aspects of the problem space – helps students to represent their discussion according to a global model of coherence. The parallel discussion lines that represent different topics do not make a discussion scattered. Although students frequently switched between topics; they still responded to what has been put forward earlier.

150

Figure 8.5: Sequence of related contributions.

Computer-Mediated Communication andTask Performance

151

Figure 8.6: Spatial behavior in the shared workspace.

Chapter 8

Discussion

In this chapter, we presented a study that examined the communication patterns in the shared digital workspace. The goal of the study was to find out if students would use the collaborative tool to communication about the task. Furthermore, we were interested in how the computer-mediated part of the communication would look like whereby we focused on the relationship between characteristics of the tool and the communication patterns they give rise to.

The notation system of the Graphical tool made the students aware of the possibility to ask questions or to give comments. These two options stimulated further elaboration of the proposed solutions. Students put forward their solutions without any interference of the other students. Their solutions remained visible during the whole discussion. This kept the discussion open; the group explored the various directions, i.e. the different solutions that the members put forward at the beginning of the discussion. The computer support has a similarity to structured group interventions like the Devil's advocate (Janis, 1982) or the Stepladder technique (Rogelberg, Barnes-Farrell & Lowe, 1992), in the sense that these techniques aim to prevent groups from “a premature adoption of a preferred solution” (Nemeth, Brown & Rogers, 2001).

The sequence of communicative acts within the digital workspace looked quite different from the way verbal talk is organized. Groups no longer organized their discussion based on the principle that only one member speaks at the same time. Instead, group members put forward their ideas without any delay and they can directly respond to all the statements made by their fellow group members. It gave the students a sense of “cognitive 'freedom”: they followed their own line of thinking and they directly shared their ideas with the group. In the next chapter, we further elaborate on this issue by taking into account the relationship between verbal and computer-mediated part of the face-to-face communication.

152

9 A Diagnostic-reasoning Discussion

with the Support of CoFFEE

The studies that have been discussed in chapter 7 and 8 evaluated the initial design guidelines that lay down the basic properties of the envisioned collaborative tools. With these guidelines, we expected to realize the intended outcomes: a productive discussion between a small group of students. The initial guidelines were implemented in a Graphical tool. We introduced this tool in the educational practice to study the communication patterns that occurred in the shared workspace of the tool. Through these studies, we got a better understanding of how students organize their digital discussion. Findings from these studies were used to fine-tune the tool design so that the basic properties of the tools would be fully realized in practice. These improvements reflect the iterative character of the research. The first research cycle led to a number of findings that were used to adapt the design. These adaptations resulted in eight design guidelines that form the basis of the two collaborative tools that are part of the newly developed CoFFEE software system1.

The study that is discussed in this chapter investigates the effects of the CoFFEE system on the group communication and the learning activities. The study moves away from examining a single guideline and concentrates on the two basic properties of the collaborative tool: 1) parallel access as floor control mechanism and 2) the use of representational aids to stimulate effective task performance. These two properties are associated with some clear expectations of how the digital discussion will proceed.

First, it is hypothesized that parallel access leads to participation that is more equal. The digital part of the discussion will be more balanced in terms of the number of contributions put forward by the different group members. This may be due to the fact that parallel access makes it more difficult for an individual student to control the

1 http://www.coffee-soft.org/

153

Chapter 9

discussion. Students immediately put an idea into words without much disturbance from other group members. They can express their ideas without direct interruption or delay.

Secondly, it is expected that the representational aids of the collaborative tools stimulate certain cognitions and behaviors that are associated with a productive group discussion. Representational aids like categories and a notation system make students aware of essential aspects of the learning task, which on its turn should improve the performance of the group.

This chapter is divided in three parts. First, we discuss the development of an instructional strategy that aims to establish a good match between the learning task and the computer support. The instructional strategy makes the CoFFEE software system suitable for a specific context of use. The strategy describes how the students use the CoFFEE system effectively during the collaborative learning activity. It lays down how the group discussion should look like in practice. Secondly, we examine if the digital part of the communication is really a discussion, i.e. does the computer-mediated part of the communication fulfills the minimal requirements for interaction? The criteria that will be evaluated are “joint attention” and “the ability to express one's ideas”. These criteria emphasize that the students who interact must be able to 1) establish an attended focus and 2) to express themselves digitally. Finally, we describe how a face-to-face discussion that has a digital and verbal component looks like in practice. We present a number of analyses that compare the pattern of communication of groups who work with and without the CoFFEE software. These analyses focus on the sequential organization of the communicative acts. They examine how the sequential organization made available by the two media – verbal and computer-mediated – affects the process of knowledge construction.

9.1 Diagnostic reasoning in the ClassroomWe introduced the CoFFEE system at a secondary school for vocational education – a regional educational center (ROC) – that provides professional training for several occupational groups. In our case, we worked with students who attended a two-year training for assistant nurse. These students followed a course about basic health care in the nursing home. An essential element of that course is diagnostic reasoning.

154

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

Clinical Problem Solving

Diagnostic reasoning is a kind of clinical problem solving that includes the recognition of health problems and the development of an appropriate health care plan. It has to do with drawing inferences about the nature of health problems that can be helped or changed by nursing action (Cholowski & Chan, 1992). These inferences are derived from cues about patient’s health condition while their outcomes are used to develop a care plan that addresses the patient’s health problems.

The ability of the nurse to provide safe and competent care depends on good diagnostic-reasoning skills (Taylor, 2000). Nurses must be able to make relevant observations, recognize health problems and develop an appropriate care plan to address those problems (Tanner, Padrick, Westfall & Putzier, 1987). Developing these skills among nursing students deserves particular attention. It is a major subject of most nursing curricula. The teachers that participated in the study mentioned that students find it difficult to acquire diagnostic-reasoning skills. The course “basic care” pays a lot of attention to these skills: diagnostic reasoning is addressed during a number of lectures and it is practiced during a series of lessons.

The students that participated in the research developed their diagnostic-reasoning skills during small-group discussions. These discussions followed after a number of lectures during which the teacher explained the theory of diagnostic reasoning. Group discussions should stimulate the students to apply the newly learned knowledge in practice. It is assumed that students who discuss a patient's case in small groups will engage in a more active learning. The students should apply prior clinical knowledge and skills to identify and interpret patient's health cues. They had to construct a shared representation of the problem space that contains appropriate clinical inferences of the patient health. This kind of learning requires active participation of all group members and frequent task-related interactions. The computer support that was provided to the groups focused on these two requirements for a productive discussion.

Objectives of the Study

The CoFFEE software system was used to support an authentic learning activity, i.e. a diagnostic-reasoning discussion between a small group of students. It is assumed that the structured nature of this discussion makes that learning activity suitable for

155

Chapter 9

computer support. In this chapter, we focus on the two collaborative tools that are part of CoFFEE software. We examine how the collaborative tools affect the process of knowledge construction. The study that is presented in this chapter has two main objectives:

1. To development and evaluation an instructional strategy that makes the CoFFEE software suitable for a specific learning context.

2. To study the effects of the collaborative tools on the diagnostic-reasoning discussion.

In chapter 4, we mentioned that CoFFEE as a software system cannot be set apart from the instructional strategy that makes the system applicable for a certain learning activity within a specific education context. The instructional strategy helps students to make effective use of the software. With an appropriate strategy, it becomes more likely that the students display specific patterns of thinking and acting that enables them to achieve their learning goals. A first objective of the study is to develop the proper instructional strategy that utilizes the potentials of the CoFFEE software. This is done through an iterative design cycle that covers the first four lessons.

A second objective of the study is to examine the effects of the collaborative tools on the communication, collaboration and learning activities. To study these effects we used a quasi-experimental design with two conditions: 1) face-to-face groups who communicated by means of verbal exchanges, and 2) face-to-face groups who communicated verbally but also used CoFFEE to mediate part of their exchanges. The two conditions make it possible to identify possible changes in the communication patterns brought about by the collaborative tools.

9.2 MethodThe students that participated in the study worked together in small groups of three or four students. They received a brief case description of a patient's health condition at the start of the lesson. The case description was simplified so that the students were able to carry out their learning task within one lesson. Furthermore, an abridged case description makes the occurrence of certain learning activities more likely.

The pedagogical objective of the group discussion is to improve the diagnostic-reasoning skills. The students should analyze the patient’s health condition and develop

156

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

an appropriate care plan. They had to identify, interpret and cluster the health cues, establish a diagnosis that provides insight into the patient’s malady and write a health care plan that contains interventions to decrease or solve patient’s health problems.

The study covered five lessons that spread out over 8 weeks. Each lesson lasted approximately 1½ hours. During all these lessons, the students carried out the same learning task. At the beginning of the lesson, the students received a short case description of a patient health condition. All the lessons had the same pattern: students read the case description after which they discussed that case and formulated a health care plan.

Learning Task

Diagnostic reasoning involves the identification and management of patient’s health needs (O’Neill. Dluhy & Chin, 2005). It can be described as clinical problem solving, which includes the identification of problems and the formulation of solutions (Gordon, 1994). Essential elements of that problem-solving process are the generation and verification of diagnostic hypotheses (Kassirer, 1989). In our study, diagnostic reasoning consisted of four related problem-solving activities (Table 9.1): 1) interpreting health cues, 2) classifying cues, 3) determining the nursing goals, and 4) developing interventions.

157

Four problem-solving activities

1. During the first step, the students share the health cues they have identified from the case. They should interpret and organize these cues according to a formal classification method.

2. Students used the results from step 1 to describe the patient’s health condition. They have to make a distinction between the signs and symptoms, the health problems, and the causes of these problems.

3. During the third step, the students have to determine the nursing goals that describes the desired health condition of the patient.

4. During the fourth step, students choose interventions or actions that enable them to achieve the nursing goals.

Table 9.1: Diagnostic Reasoning as problem solving.

Chapter 9

During the first step, the students used a formal classification method to organize and interpret information about a patient's health condition. The classification method is based on the work of Gordon (1994). Gordon identifies 11 health patterns that describe a set of behaviors that negatively influence patient’s functioning (Table 9.2).

Gordon's classification helps students to assess the patient’s condition by organizing health cues into meaningful categories. It decomposes the problem into a number of important health aspects. The students in our studies only used a limited number of the health patterns to analyze a case. At the start of the lesson, they received a list of the patterns that are relevant for the case.

The students applied a second classification method during the second step. This classification method labels the health cues identified earlier according to the “PES format”. The PES format makes a distinction between a health problem, its causes and its consequences (Table 9.3). The three letters refer to the health problem (P), the causes or etiological factors (E), and the consequences or clusters of signs and symptoms (S). This PES format is used to develop a diagnosis that consists of well-founded hypotheses about the causes of the health problems.

Classifying health cues according to the PES format is crucial for the development of a health care plan. It informs students which actions will be effective. For example, students have to realize that an intervention that addresses only signs or symptoms does not really solve the patient's health problem; it only suppresses the observable aspects of the health problem.

158

Health patterns

Health perception and management Self-perception and self-concept

Activity and exercise Roles and relationships

Nutrition and metabolism Coping and stress management

Elimination Sexuality and reproduction

Sleep and rest Values and beliefs

Cognition and perception

Table 9.2: Gordon's health patterns (Gordon, 1994).

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

After the students formulated a diagnosis, they determined one or more nursing goals during the third step. These goals describe the desired condition of the patient’s health. They are formulated according to the RUMBA criteria. The abbreviation RUMBA emphasize that a nursing goal should be relevant (R), understandable (U), measurable (M), behavioral (B), and attainable (A).

Finally, in the fourth step the students formulated a set of interventions to achieve the nursing goals defined during the previous step. They have to choose a set of actions that solve or decrease the patient's health problem. These actions should relate to the nursing goal, the problem and the situation of the patient.

Computer Support

The structured nature of the learning task makes the diagnostic-reasoning discussions especially suitable for computer support. Aspects of the learning task that appear in the support environment are:

1. the sequence of problem-solving activities,

2. the two classification methods, and

3. the relationship between signs and symptoms, problems, causes, nursing goals and interventions.

Problem-solving activities as steps

The students had to follow a distinct sequence of problem-solving activities (see Table9.1). This sequence was implemented in CoFFEE as a number of successive steps whereby each step was defined by a specific activity that the students had to carry out.

159

PES format

Health problem A health-related state or process, manifest by the individual, family or community

Etiological or related factors

The probable factors causing or maintaining the client’s health problems.

Signs and symptoms Critical defining or supporting characteristics that are an indication of the condition present in nearly all clients with the problem.

Table 9.3: PES Format (Gordon, 1994).

Chapter 9

It makes students aware of the different, related activities that comprise a diagnostic discussion. Each step had its own arrangement with regard to the kind of tools that the students used and the configuration of these tools.

Figure 9.1 shows the “SessionControl” window of the CoFFEE Controller, the tool that the teacher used during the lessons to manage a CoFFEE session. The SessionControl window displays the various steps that the students had to carry out. These steps divided the problem-solving discussion into a sequence of four related activities (see Table 9.1).

It is worth to notify that the CoFFEE Controller runs on the computer of the teacher so that the teacher is able to direct the activities of the groups. The teacher determines when an activity (step) ends or when a new group activity (step) begins. The CoFFEE Controller also allows the teacher to assign students to a group or to change the composition of the group before or during a session.

Classification Methods as Categories and Post-hoc Notation

The students had to apply two formal classification methods, i.e. Gordon's health patterns and the PES format during the second step of the problem-solving sequence

160

Figure 9.1: Various steps displayed on the teacher's computer.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

(see Table 9.1). These classification methods were implemented in the Threaded-discussion tool as representational aids. These aids guide the task-related communication and facilitate the construction of a shared problem space.

Gordon's health patterns were implemented as categories in the Threaded-discussion tool (Figure 9.2). They provided the students with a relevant information structure that made it easier to construct a shared representation of the problem space. Before the student shared a health cue with the group, they had to choose the appropriate health pattern. It means that the health cues that were placed in the digital workspace were organized according to Gordon's classification method.

The students classified the health cues according to the PES format. This was implemented in the Threaded-discussion tool by means of the post-hoc notation system (Figure 9.3). Students could classify a contribution that has been placed in the shared workspace according to the PES format. To do so, they had to right click on a contribution with their mouse and choose the option “Set the contribution type to” from the drop down menu. Now the contributions could be tagged. Three options were available: P for health problem, E for the causes or etiological factors, or S for signs and symptoms.

161

Figure 9.2: Gordon's health patterns as categories.

Chapter 9

Representing the Relation between Causes and Interventions

The students had to describe the relationship between signs and symptoms, health problems, causes, nursing goals and interventions (Table 9.1). This relationship could be visualized in the two-dimensional workspace of the Graphical tool (Figure 9.4).

A notation system forces the students to choose between the five labels that describe the relationship between causes of the health problem and the interventions. The student typed a text after a selection has been made. That text was displayed as a

162

Figure 9.3: PES format as post-hoc notation.

Figure 9.4: Relationship displayed by the Graphical tool.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

box with a specific label, i.e. signs and symptoms, health problem, cause, nursing goal or intervention.

Two versions of the Software

The study that is discussed in this chapter has been carried out after the major usability issues were solved. This enabled the researchers to focus on the educational effects of CoFFEE without major disturbances that have to do with the accessibility and operability of the system itself.

During the first four series of experiments, the students worked with CoFFEE version 1.11. The last two experiments were carried with version 2.0, the Beta version of CoFFEE (Table 9.4). The Beta version had some new functionality like the possibility to transfer outcomes from a finished step to a next step. This functionality was used during the last two lessons.

Participants

The students that participated in the experiments came from two classes, the BOL class and the BBL class. Both classes combine school with an apprenticeship. For the BOL class the emphasis is on education training. These students spend most of their time at

163

Lesson Condition1(only face-to-face)

Condition2(face-to-face and CoFFEE)

Lesson 1Case: Kim

3 groups (2 BOL & 1 BBL) 5 groups (3 BOL & 2 BBL)CoFFEE version 1.1

Lesson 2Case: Benny

3 groups (2 BOL & 1 BBL) 4 groups (3 BOL & 1 BBL)CoFFEE version 1.1

Lesson 3Case: Mr. Doster

3 groups (2 BOL & 1 BBL) 5 groups (3 BOL & 2 BBL)CoFFEE version 1.1

Lesson 4Case: Mrs. Winkel

3 groups (2 BOL & 1 BBL) 4 groups (3 BOL& 1 BOL)CoFFEE version 2.0

Lesson 5Case: Mrs. Verbeek

3 groups (2 BOL & 1 BBL) 3 groups (2 BOL & 1 BBL)CoFFEE version 2.0

Table 9.4: Overview of the five lessons.

Chapter 9

school. The students in the BBL class worked part-time in a nursing home and only attended school for one day a week. The BOL class consisted of 18 students ranging in age from 16 to 18 years, while the BBL class consisted of 12 students ranging in age from 16 to 20 years.

Two Conditions

All the groups were situated in the same classroom. The members of a group sat near each other so that they could communicate face-to-face. To examine the effects of the CoFFEE software system we divided the groups between two conditions. The only difference between the two conditions was the kind of support available (see Table 9.4):

1. Only verbal.

These face-to-face groups carried out a discussion in the regular way. They communicate verbally. The Only-verbal condition served as a reference that helped us to interpret the findings of the CoFFEE supported condition.

2. CoFFEE supported.

The CoFFEE supported groups could also use a collaborative tool for discussion. These groups had two modes of communication at their disposal: verbal and computer-mediated.

The teacher divided the students into groups of three or four students and allocated these groups to one of the two conditions. Group composition slightly changed during the six lessons because of absentees. Still, the students did not change between the two conditions during the sequence of lessons.

Data Collected

Every lesson was evaluated with the teacher right after the lesson was finished. We also took into account the comments that were made by the students with regard to the planning and execution of the lessons. Each group was also videotaped. These videotapes were transcribed with the support of ELAN, a software application for the creation of annotations on video resources. The communication that was supported by CoFFEE system is recorded by the software in the form of traces. These traces were annotated with Tatiana, a software application that has been developed by the LEAD project. This application transforms CoFFEE traces into meaningful data. We also

164

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

developed a questionnaire to collect data about student's attitude about the collaborative tools that are part of the CoFFEE system. Only the students in condition 2 filled in the questionnaire during lesson one, two and four. To summarize, the following methods for data collection were used:

1. observations from the teacher and researcher,

2. feedback of students,

3. video recordings,

4. CoFFEE traces, and

5. questionnaires.

9.3 Instructional DesignOne of the objectives of the study is to develop an instructional strategy that enables students to make effective use of the CoFFEE system. Effectiveness was evaluated from a task perspective. Decisions about the planning of the activities and the application of CoFFEE had to do with effective task performance. After each lesson, the instructional strategy was evaluated with the teacher. This led to a number of adaptations. Gradually an instructional strategy emerged that made the CoFFEE system suitable for a specific educational task and context: a diagnostic discussion carried out by secondary students who followed a vocational training for assistant nurse.

The First Lesson

The instructional strategy that was implemented during the first lesson made a distinction between 5 activities (Table 9.5) that were based on the four activities that comprise a diagnostic discussion (see Table 9.1).

At the beginning of the lesson, the groups received a case description and a list of the health patterns that applied to the case. They also received a planning that describes the sequence of activities that they had to follow. First, the students read the case individually, after which they started with the group discussion. The group discussion was divided into a number of activities that are called steps in the “CoFFEE terminology”.

The CoFFEE groups (condition 2) used the Threaded-discussion tool during step 2 to share the health cues they identified from the case. The Threaded-discussion tool

165

Chapter 9

represented Gordon's health patterns as categories and used a pre-defined notation system to stimulate elaboration. The notation system had three options to label a contribution: 1) cue from case, 2) kind of cue, and 3) comment. Students should use the first label “cue from the case” to put forward a health cue from the case. They should classify a cue according to the PES format with the second label “kind of cue”. The students could also elaborate on the cues that were placed in the shared workspace with the third label “comment”.

During step 3, the CoFFEE groups used the Graphical tool to represent the relationship between the health problem, the signs and symptoms, and the causes. The Graphical tool had a notation system that had four options: 1) Cause, 2) Problem, 3) Signs and Symptoms, and 4) Comment.

The CoFFEE groups defined the nursing goals in step 4 with the support of the Threaded-discussion tool. Each student was assigned to a category where they put forward a nursing goal. The other group members reacted on the proposed nursing goal by asking a question or providing comment. It meant that the Threaded-discussion tool had a notation system with three options: 1) Nursing goal, 2) Comment, and 3) Question.

Finally, in step 5 the students developed an intervention with the support of the Graphical tool. Now the notation system had four labels: 1) Problem, 2) Cause, 3) Nursing goal, and 4) Intervention. The two-dimensional workspace was divided into

166

Step Activity Support Condition 2

1 Students read the case.

2 The groups identify and organize health cues according to Gordon's health patterns.

Threaded discussion tool(categories + notation system)

3 The groups define the patient's condition by using the PES format.

Graphical tool(notation system)

4 The groups define the nursing goals. Threaded discussion tool(categories + notation system)

5 The groups develop interventions. Graphical tool(meaningful areas + notation system)

Table 9.5: The instructional strategy for the first lesson.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

four meaningful areas. These areas were associated with an intervention.

Changes after the First Lesson

The teacher and the students reported some difficulties that had to do with the coherence between the various activities that the CoFFEE groups had to carry out. Students found it difficult to understand the relationship between the various steps. They alternately used a Threaded-discussion tool and a Graphical tool. Those tools were configured differently during the various steps. These changes were simply too hard to grasp for the students. It prevented them so see the overall picture and it made the CoFFEE-supported discussion rather complex.

The evaluation of the first lesson resulted in two major adaptations of the instructional strategy. These adaptations created more coherence between the various activities. Both changes had a similar outcome: they combined two separate steps into one new step.

A first change was to join the steps 2 and 3 into one new step (step 2 of Table 9.6). During this step, the students used the Threaded-discussion tool to share health cues. They classified the cues according to the health patterns of Gordon and they indicated if a cue was a problem, a cause or a sign of a symptom (PES format). The health

167

Step Activity Support for Condition 2

1 Students read the case.

2 The groups identify and organize the health cues and come up with a diagnosis

Threaded discussion tool with:

– categories that display Gordon's health patterns, and

– a post-hoc notation based on the PES format.

3 The groups represents the relationship between the problem, its causes, the sign and symptoms and the nursing goal to conclude what interventions are needed.

Graphical tool with:

– a notation system with five labels, and

– a shared workspace divided into four meaningful areas.

Table 9.6: The instructional strategy for the second lesson.

Chapter 9

patterns of Gordon were displayed as categories (see Figure 9.2), while the PES format was implemented by means of the post-hoc notation system (see Figure 9.3). First, students had to select the right category – i.e. health pattern – before they share a heath cue with the group. A health cue was shared as a written contribution that could be classified with the post-hoc notation system, i.e. the “set the contribution type to” functionality. A student could classify a contribution any time after it is displayed in the shared workspace.

A second change was the combination of the steps 4 and 5 into one new step (step 3 of Table 9.6) that was supported by the Graphical tool. During this step, students used the two-dimensional workspace of the graphical tool to represent the relationship between 1) the problem, 2) its causes and consequences, 3) the sign and symptoms, 4) the nursing goal, and 5) the interventions. The notation system of the Graphical tool distinguished between these five contribution types. The workspace was divided in four areas where each area represents one solution.

The two adaptations of the instructional strategy that were implemented after the first lesson were major changes. They seemed to solve the drawbacks that the students experienced during the first lesson. The other changes that were made after the second and third lesson were less inspired by the evaluation of the software. The aim of these latter changes was to test new functionalities of the software.

Changes after the Second Lesson

The new instructional method was implemented during the second lesson. One of the characteristics of the new methods was that she students started to work individually during step 3 (Table 9.6). First, they had to represent the relationship between cause and intervention in their own area of the shared workspace. Next, they could comment on the maps that their fellow group members had constructed. Before the students began with activity 3, they had to agree which problem would be addressed by the different group members. There was no explicit reference in CoFFEE that each student should choose a health problem. A new step was added in CoFFEE that followed after step 2. The purpose of that step was to pinpoint the group that they had to agree upon who would solve which health problem. This was done by means of a verbal discussion during which the group members have to agree which health problems they would address. Within the CoFFEE system, this activity is displayed as an empty step (Figure

168

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

9.5). An empty step means that no tools were offered to the students. The CoFFEE software only displays a screen that contains a brief instruction about the activity they had to carry out.

Changes after the Third Lesson

After the third lesson, the students worked with a new version of CoFFEE (version 2.0) that had some new functionalities. This version had the “copy-paste between steps” functionality: the outcome of a step – for example a threaded discussion or a diagram – can be copied into the next step. The advantage of this functionality is that students further elaborate on their group product with a collaborative tool that is configured differently. This functionality was used during the fourth lesson. A new step was added (step 5, Table 9.7) during which the students reflect on the outcomes of the previous step.

Analysis I: The Instructional Strategy

The instructional strategy that has been discussed earlier made the CoFFEE software feasible for supporting a diagnostic-reasoning discussion. The strategy had the form of a lesson plan that described the sequence of activities and the kind of support that enables students to carry out those activities effectively. Each activity is further

169

Figure 9.5: Empty step displayed on a student's PC.

Chapter 9

described in terms of 1) method of working, 2) the support, and 3) the expected outcomes.

We evaluated the instructional strategy by asking the students how satisfied they were with the support provided by the collaborative tools. Bailey and Pearson (1983) define satisfaction as the sum of the person's feelings of attitudes towards a variety of factors affecting that situation (Bailey and Pearson, 1983). In our case, we relate satisfaction with student's attitude about the computer support. This was further specified by the actions that the students could carry out with the tool. For the Threaded-discussion tool, the following actions were evaluated: 1) sharing health cues, 2) the use of categories to represent Gordon's health patterns, 3) the use of a notation system to encourage elaboration, and 4) classifying cues according to the PES format.

170

Step Activity Support Condition 2

1 Students read the case.

2 The groups identify and organize the health cues and come up with a diagnosis.

Threaded discussion tool with:

– categories that display Gordon's health patterns, and

– a post-hoc notation based on the PES format.

3 Each group members choose a problem they will work on.

Verbal discussion (empty step).

4 The groups represent the relationship between the problem, its causes, the sign and symptoms and the nursing goal to conclude what interventions are needed.

Graphical tool with:

– a notation system with five labels, and

– a shared workspace divided into four meaningful areas.

5 The group members provide feedback on the diagrams that were created during the previous step.

Graphical tool:

– result from the previous step is copied into the graphical tool.

– a notation system with two contributions: feedback and question.

Table 9.7: Instructional strategy for the fourth lesson.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

For the Graphical tool, the following three activities were evaluated: 1) sharing ideas, 2) representing the relationship between cause, health problem, signs and symptoms, nursing goal and intervention, and 3) providing feedback. It is worth to notify, that not all the actions were evaluated after every lesson because the instructional strategy changed during the course of the lessons.

Method

The students received a questionnaire after lesson 1, 2 and 4 that was filled in by the 9 students of the three BOL groups (see Table 9.4). We only used the data from the BOL groups because these groups remained the same during all three lessons. Satisfaction was measured by a number of items with a 5-points Likert scale. This scale had the following options: 1) very dissatisfied, 2) dissatisfied, 3) neutral, 4) sufficient satisfied, and 5) very satisfied.

For each measurement, we added up the responses of the students for the individual satisfaction ratings. This resulted in two overall measures of satisfaction: 1) satisfaction with the support provided by the Threaded-discussion tool, and 2) satisfaction with support provided by the Graphical tool.

Results

The students reported that they were moderately satisfied with the Threaded-discussion tool. The level of satisfaction slightly improved after the first lesson: M=3.3, SE= 0.44 for Lesson 1, M=3.7, SE=0.67 for Lesson 2, and M=3.7, M=0,68 for Lesson 4. Satisfaction with the Graphical tool showed a slightly different pattern: M=3,3, SE=0,44 for Lesson1, M=3,9, SE=0,96 for Lesson 2, and M=3,4, SE=0,44 for Lesson 4.

Figure 9.6 represents the overall satisfaction with the support provided by the collaborative tools. Satisfaction slightly increased after the first lesson. It seems that the changes in the instructional strategy that have been implemented after Lesson 1 had their effect. This effect remained visible during Lesson 4. Satisfaction with the Graphical tool improved after Lesson 1 but it decreased during Lesson 4. A closer look at the individual items revealed that the students were not that satisfied with the new activity (step 5 of Table 9.7) that was introduced during Lesson 4. Some students reported that this step was unnecessary. During the previous activity (step 4 of Table9.7) they already gave feedback verbally.

171

Chapter 9

9.4 Analysis II: The Minimal Requirements for a Group Discussion

One of the issues that was addressed in the study was the utility of the CoFFEE system. Utility refers to the usefulness of the CoFFEE system in relation to the task that the students have to carry out. Usability and utility are two distinct concepts; the former refers to the ease by which the students employ the CoFFEE system while the latter indicates if students are able to carry out their task and achieve their learning goals with the support of CoFFEE. Utility in our case refers to how well the students are able to discuss a patient’s case with the support of the CoFFEE system. If CoFFEE does not support the task performance effectively than the utility of CoFFEE will be low.

First, we examine if the CoFFEE system supports a discussion. In the next

172

Figure 9.6: Satisfaction with the support provided by the collaborative tools.

Lesson 1 Lesson 2 Lesson 41

2

3

4

5

3,3

3,9

3,4

Lesson 1 Lesson 2 Lesson 41

2

3

4

5

3,3

3,7 3,7

Very satisfied

Satisfaction with the Threaded discussion tool

Satisfaction with the Graphical tool

Neutral

Very satisfied

Neutral

Very dissatisfied

Very dissatisfied

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

paragraph, we focus on the learning processes. This order stresses that collaboration precedes learning. In this paragraph, we will answer the following question: does the CoFFEE system meets the minimal requirements of a small-group discussion?

Expression of Ideas and Joint Attention

The use of CoFFEE is by no means a guarantee that the students actually have a discussion during which they share and integrate knowledge and information. Communication with the support of CoFFEE differs from the oral, face-to-face communication. Students have to type their ideas instead of expressing them orally. They must be able to express their ideas as text, instead of spoken words. Furthermore, parallel access and the use of categories might prevent students from actually “meeting each other” in the shared digital workspace. As has been mentioned in chapter 5, joint attention is a prerequisite for a group discussion that aims to develop a shared understanding among group members (Barron, 2003; Sfard & Kieran, 2001). The question is if the students are able to establish a common focus. Are they aware of the contributions put forward by others and do they react on these contributions?

Method

The minimal requirements for a CoFFEE discussion refer to the ability to express ideas into words and the possibility to interact with other group members. These requirements are measured by means of a questionnaire. The students had to respond to four questions that address different aspects of the digital communication. These four questions are:

1. I can put forward my ideas well in the CoFFEE system?2. I have the feeling that my group members have read my contributions?3. I regularly read the contributions of my group members?4. My group members react on my contributions?

The four questions address three aspects of the communication process: 1) expression of ideas into writing, 2) awareness of contributions of other, and 3) group interaction. These aspects of the communication process ware measured three times: during the first, second and fourth series of experiments. The response of the students was measured on a five-point scale with the options: 1) never, 2) sometimes, 3) regularly, 4) often, and 5) always. Students filled in the questionnaire at the end of the lesson after

173

Chapter 9

they have finished working with CoFFEE.

Results

Results indicate that during the first lesson students found it sometimes difficult to express their ideas. This was hardly the case during the second and the fourth lesson (Figure 9.7). This may be because the instructional strategy changed fundamentally between the first and the second lesson. An alternative explanation would be the occurrence of a learning effect. After the first lesson, the students became more skilled with the CoFFEE software.

The students could simultaneously put forward a contribution in the shared workspace. This could make it difficult to keep track of what the other group members had written in the shared workspace. The use of categories may even strengthen this problem. Each category has its own shared workspace. Group members who work in different categories may not be aware of what is going on in the other categories. Furthermore, sometimes they discussed things verbally so that attention shifted from the digital workspace towards the verbal discussion. Establishing an attended focus is a prerequisite for a group discussion. The group members should know what other members have said. It is almost impossible to discuss a topic if the students cannot follow their fellow group members' reasoning processes.

174

Figure 9.7: The expression of ideas in CoFFEE.

Lesson 1 Lesson 2 Lesson 41

2

3

4

5

3,2

4,13,9

Very often

I can put forward my ideas well with the support of the CoFFEE software

Regularly

Never

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

The students reported that they were aware of what other group members put forward in the shared workspace. There is an attended focus on what takes place within the digital workspace. The group members frequently read the contributions of their group members (Figure 9.8). Again, there is a slight difference between the first lesson on the one hand and the second and fourth lesson on the other. Awareness improved after the first experiment.

Interaction also improved throughout the experiments (Figure 9.9). During the fourth lesson, students indicate that there is frequent interaction between the students within shared digital workspace. This may be due to the fact that the instructional strategy during the second and fourth lesson better fits the task that the students had to carry out. Students may also able to provide better comments because their diagnostic-reasoning skills had improved during the course of the lessons.

175

Figure 9.8: Keep oneself informed about the communication in CoFFEE.

Lesson 1 Lesson 2 Lesson 41

2

3

4

5

3,8

4,4 4,5Very often

I have the impression that my group members read my contributions

Regularly

Never

Lesson 1 Lesson 2 Lesson 41

2

3

4

5

4 4,1

4,5Very often

Regularly

Never

I read the contributions of my group members

Chapter 9

9.5 Analysis III: The Communication of KnowledgeIn chapter 6, we argued that each medium provides distinct opportunities for communication. The different opportunities have to do with the way that the communicative acts are organized into a coherent and meaningful whole. In this study, we compare two media for face-to-face communication: verbal speech and computer-mediated interaction. We examine how the groups communicate verbally and how they interact digitally. We expect that the use of a collaborative tool changes the nature of the communication, which on its turn affects the processes of collaboration and learning. We adopt a functional perspective towards group communication. Communication is analyzed to determine which functions it serves in term of the requisite needs, goals, problems and challenges that the group must satisfy or overcome to be effective and to maintain its well-being (Poole, 1999). In our case, the communicative functions are related to the requirements of a constructive dialogue that have been presented in chapter 4. The communicative demands that are addressed in the analysis are: group member's participation, group's orientation towards the learning task, the coherence of the communication and knowledge elaboration.

The five groups worked more or less autonomously. Most of the communication is communication between group members. Still, the teacher had a significant role in the performance of some groups. Therefore, we also examine the interactions of the groups with the teacher. First, we present the degree of student participation in both media. Secondly, we discuss the task-related and off-task communication for the verbal and the computer-mediated part of the discussion. Thirdly, we focus more deeply on the

176

Figure 9.9: Perceived interaction within the CoFFEE environment.

Lesson 1 Lesson 2 Lesson 41

2

3

4

5

3,7 3,74

Very often

My group members react on my contributions.

Regularly

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

different aspects of the task-related communication.

Method

Five groups of the BOL class took part in the study; three groups used the CoFFEE system (group 1, 2 and 3) while two groups (group 4 and 5) only discussed things orally. Group 1 and 3 had one male student; all the other students were female. All the groups were situated in the same room (Figure 9.10). The group members sat behind a table with their faces opposite to each other. The teacher was also present in the classroom. A researcher sat behind the teacher's laptop to control the CoFFEE session.

The lesson that we used for further analysis was lesson 2. Significant changes in the instructional strategy were implemented during that lesson. The results that were discussed in paragraph 9.3 and 9.4 indicate that these changes led to a better use of the collaborative tools.

177

Figure 9.10: Classroom layout.

Chapter 9

Analytical tools

The five groups were all video-taped while the communication that occurred in the digital workspace of the CoFFEE tools was recorded and stored as log files. A number of software applications were used to transcribe, annotate, combine and analyze these two data sources.

With the ELAN software2 we transcribed the spoken text into a data. ELAN is an annotation tool that helps researchers to create, edit and visualize transcripts from video recordings. ELAN automatically calculates the number of contributions and speaking time for each participant. The ELAN transcripts were imported into Tatiana software3 for annotation. The CoFFEE data was directly imported in Tatiana and filtered so that only the information necessary for further analysis was displayed. The CoFFEE traces and the oral transcripts were also annotated with Tatiana. Finally, we used the Sigmaplot software4 to visualize patterns of communication on a time axis.

Group Member's Participation

Participation is defined as the degree that group members contribute to the discussion. In chapter 4, we argued that even participation would be beneficial for group learning. Even participation refers to those situations where all group members are able to share their thoughts with the group. Then the group makes optimal use of the knowledge and skills of its members. The group fully exploits its learning potentials when all members participate substantially.

Even participation during a verbal discussion may be hampered due to ineffective communication patterns. A discussion can be dominated by one or more rhetorically skilled students who frequently interrupt or overrule other group members to control the discussion. This leads to an unequal participation whereby less dominant or less communicative students have fewer opportunities to share their ideas. Furthermore, the temporal, turn-taking mechanism of spoken communication may prevent the free expression of ideas. This problem of “product blocking” refers to the rule that only one group member speaks at the same time. Group members who are prohibited to

2 ELAN was developed at the Max Planck institute for Psycholinguistics, Nijmegen, The Netherlands.

3 www.lead.emse.fr

4 www.sigmaplot.com

178

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

verbalize their ideas at the moment they occur forget or suppress them (Nijstad, 2000).

It is assumed that parallel access as floor control mechanism counteracts dominant behavior and product blocking. The collaborative tools of CoFFEE are based on parallel access. They enable group members to put forward an idea directly without any delay. Users type an idea in a private text box that cannot be accessed by other users. It means that other group members are not able to interrupt that process abruptly. Dominant group members have fewer opportunities to control the discussion. It is hypothesized that the group communication based on parallel access will show a more even participation in comparison with the verbal communication that is based on turn taking. To test this statement, we analyzed the degree of participation for both the oral and the computer-mediated part of the group discussion.

Participation in the Verbal Discussion

Group member's speaking turns during the verbal discussion were added up for a total score (Table 9.8). Next, we calculated the degree of participation as the part of each member's score in the total amount of group messages (Table 9.9).

179

Participation (number of contributions)

Member 1 Member 2 Member 3 Member 4 Total TeacherGroup 1 (Cond. 2) 275 268 327 870 128

Group 2 (Cond. 2) 120 384 269 773 106

Group 3 (Cond. 2) 149 187 295 631 85

Group 4 (Cond. 1) 48 388 315 144 895 264

Group 5 (Cond. 1) 230 55 167 215 667 42

Table 9.8: Number of spoken contributions during the oral discussion.

Participation (proportion)

Member 1 Member 2 Member 3 Member 4 Standard deviation

Group 1 (Cond. 2) .31 .31 .38 .04

Group 2 (Cond. 2) .16 .50 .35 .17

Group 3 (Cond. 2) .24 .30 .47 .12

Group 4 (Cond. 1) .05 .43 .35 .16 .17

Group 5 (Cond. 1) .34 .08 .25 .32 .12

Table 9.9: Group member’s degree of participation in the oral discussion.

Chapter 9

Table 9.9gives an overview of the group member's participation in the verbal discussion for all the groups, i.e. for the CoFFEE condition (group 1, 2 and 3) and the face-to-face condition (group 4 and 5). It is worth to notify that the groups in the CoFFEE condition also communicated verbally. The last column of Table 9.9 shows how much the groups deviate from the ideal situation of equal participation. The degree of participation varied substantially. For almost all the groups, one can conclude that there is at least one student who hardy said anything. Only group 1 showed a rather equal distribution of messages between its members.

Participation in the Digital Discussion

Table 9.10 displays the number of messages that the group members put forward in the shared workspace of the Threaded-discussion tool. Table 9.11 gives the degree of participation for the computer-mediated part of the discussion.

The communication patterns in CoFFEE differed from its verbal counterpart. First, the quantity of messages was limited. This difference between verbal and computer-mediated communication will be further discussed in the next paragraph. Secondly, all the group members contributed substantially to the discussion. The groups had a fairly even participation pattern. Their willingness to participate was irrespectively of their

180

Participation in CoFFEE (no. of contributions)

Member 1 Member 2 Member 3 TotalGroup 1 6 9 10 25

Group 2 10 8 10 28

Group 3 13 12 13 38

Table 9.10: Number of contributions in CoFFEE (Threaded discussion tool).

Participation (proportion)

Member 1 Member 2 Member 3 Standard deviation

Group 1 .24 .36 .40 .08

Group 2 .36 .29 .36 .04

Group 3 .34 .32 .34 .02

Table 9.11: Group member’s degree of participation in CoFFEE.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

behavior during the verbal discussion. Group members who hardly said anything, communicated digitally. This finding indicates that parallel access stimulates even participation. Group members who did not talk much were active in the digital workspace. These silent students did not differ from their talkative peers when it comes to digital communication.

Task-related Communication

Providing support for effective task performance was one of the objectives of the tool design. The collaborative tools should provide an optimal environment for task performance. This environment makes students aware of crucial aspects of the task that affect their performance. In this paragraph, we analyzed the task-related communication, verbally as well as digitally. The first coding schema that we applied made an analytic distinction between task and off-task communication (Table 9.12).

The task-related communication refers to the learning task that the group has to perform. It does not only include the knowledge and information that the group members share about the patient's health. It is also about the coordination of the learning activities, providing explanations and feedback.

The off-task communication contains a variety of communicative acts that can be classified into two categories: 1) communication about other topics than the learning task, and 2) social-emotional communication. The former has to do with talk that falls outside the scope of the lesson. One group, for example, discussed an assignment that they had to carry out for a different course. The latter category of social-emotional

181

Task-related communication

Communication Description

Task-related Communicative acts that directly relate to the task dimension of the group processes.

Off-task Communicative acts that refer to the social-emotional dimension of the group processes and communicative acts that address other topics than those specified by the task.

Indefinite Communicative acts that cannot be clearly defined.

Inaudible Unable to hear what the participant was saying.

Table 9.12: Coding schema that distinguishes between task and off-task communication.

Chapter 9

communication has to do with maintaining durable relationships within the group. This kind of communication affects the way group members perceive each other and the relationships they form (Prop, 1999). It includes issues of control and affection. Control has to do with managing dominance, status and power that emerge when the group collaborates. Affection refers to relationships between the group members.

The Verbal Discussion

Table 9.13 gives an overview of the task-related and off-task communication that occurred during the verbal part of the discussion. All groups – including the CoFFEE supported groups – frequently talked about the task but there was also a lot of off-task communication. The pattern of task and off-task communication differed considerably between the five groups. One can position groups 5 at one end of a continuum and group 1 at the other end. Group 5 was a group that mainly focused on the task. Group 1, in contrast, showed a lot of tension between some group members which led to a lot of social-emotional communication.

A closer look revealed some divergent patterns between the groups. Group 1 showed some personal tension at the start of the lesson. Two group members made a number of unfriendly remarks to each other. This tension remained visible during the whole lesson and went at the expense of the task-performance of that group.

Group 4 and 5 are two groups that had a high quantity of task-related communication. However, the nature of that communication differed considerably between the two groups. Group 4 asked a lot of help from the teacher. One student in that group was very eager to ask for support. That student set the stage for that group which resulted in frequent interaction with the teacher. The support that the teacher provided triggered most of the task-related communication in that group. Group 5, in

182

Task versus off-task / verbal discussion(number of sentences)

Group 1 Group 2 Group 3 Group 4 Group 5

Task-related 345 489 276 664 553

Off-task 482 284 363 373 71

Indefinite 71 28 27 36 15

Inaudible 98 74 48 85 89

Table 9.13: task versus off-task communication during the oral, face-to-face discussion.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

contrast, hardly received any help from the teacher. This group had two dominant students who were very eager to share their knowledge with the group. These two members kept the group focused on the task but they did not allow much input from the other group members.

The Computer-mediated Discussion

Table 9.14 gives the task-related communication that occurred in the Threaded-discussion tool. Most communication had to do with the task. Group 1 put forward some “test” messages at the beginning of the session to make the other groups members aware of their presence in the shared workspace. Group 3 made some off-task comments at the end of the step.

The computer-mediated communication had mainly to do with the task. The group members shared the patient's health cues in the shared digital workspace. They also respond to what others had written in the shared workspace.

The quantity of messages in the digital workspace was much smaller than in the verbal part of the discussion. Jonassen and Kwon (2001) found similar results when they compared student groups who communicated face-to-face or at a distance through the Internet. This difference can be traced back to the way group members share their ideas, i.e. by speaking or by typing. Furthermore, the discussion that emerged in CoFFEE was much more focused. Only the task-related exchanges that had to do with sharing and organizing the patient's health cues appeared in the digital workspace.

The Course of Communication

Figure 9.11 visualizes the course of the communication for groups 1, 2 and 3. The horizontal time axis covers the first part of the discussion that included two activities of reading the case and sharing the health cues with the support of the Threaded-

183

Task versus off-task / oral(number of messages)

Group 1 Group 2 Group 3

Task-related 23 58 31

Off-task 8 - 5

Indefinite - - 2

Table 9.14: task versus off-task communication in CoFFEE (Threaded-discussion tool).

Chapter 9

discussion tool (see Table 9.6). Each message is displayed as a single line on the time axis. The thickness of the line is an indication of the quantity of messages that has been put forward during a certain time interval.

Figure 9.11: The course of communicative acts.

184

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

Figure 9.11 displays three sequences: 1) verbal, task-related part of communication, 2) computer-mediated part of communication, and 3) the off-task part of the communication. The verbal and computer-mediated part of the communication was synchronized so that the three sequences could be compared. From Figure 9.11 one can see that it took some time before the students began to work with CoFFEE. The computer-mediated part of communication started later than the verbal part. Group 1 started to use CoFFEE after 12 minutes, group 2 after 4 minutes, and group 3 after 7 minutes. This is because the students first had to read the case description. All the groups had some brief interactions at the start of the lesson about how the they would carry out the task.

As we mentioned before, group 1 had a lot of social-emotional communication. It prevented the group from working on the task effectively. As Figure 9.11 displays, short intervals of task-related communication alternated with short intervals of social-emotional communication. The two types of communication seemed to parallel each other. When group 1 started to work in CoFFEE, their verbal communication decreased. Group 2 and 3 had less off-task communication. Most of the off-task communication for all groups occurred at the end. When the groups had the feeling that they were done with the task they talked about other things.

185

Figure 9.11: The course of communicative acts (continued).

Chapter 9

Figure 9.11 reveals that the contributions during the computer-mediated part of the communication occurred at large time intervals. It took the group members much more time to put forward an idea in CoFFEE. It seems that students took this time for typing. They put their thoughts into words without being disturbed by others.

The verbal and computer-mediated part of the communication paralleled each other. This is mainly because the activities that went on in the digital workspace triggers verbal reactions among the group members. The group members frequently commented on what was written in the digital workspace.

Talking about the Task

From Table 9.13 we can conclude that all the five groups talked a lot about the task. We further analyzed these task-related communicative whereby we made a distinction between five different types of communicative acts. Each individual contribution was labeled according to the coding scheme of Table 9.15.

Figure 9.12 gives an overview of the different aspects of the task-related communication for the five groups. The results only refer to the verbal, face-to-face

186

Task-related communication

Communication Description

Teacher’s intervention

Communicative acts of by the teacher like explaining task-related issues, checking progress of the group or focusing the students on the task.

Coordination Communicative acts that have to do with task execution like planning activities, discussing progress or focus attention of the group on an issue.

Factual knowledge Communication about factual knowledge, for example facts that students had learned or rules of grammar.

Diagnostic reasoning Communicative acts that have to do with sharing, organizing and the evaluation of information and knowledge with regard patient’s health with the aim to develop a health care plan.

Feedback A reaction to group member's diagnostic reasoning which is used as a basis for improvement.

Table 9.15: Coding schema for the task-related messages.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

part of the communication. It is worth to notice that groups 1, 2 and 3 also used the collaborative tool to discuss the patient's case. Not all diagnostic reasoning for these groups is displayed in Figure 9.12.

Group 4 – the non-supported group – frequently interacted with the teacher. One student in that group usually took the initiative to ask the teacher for help. This student dominated the group discussion in a specific manner. The behavior of that student prevented the group from addressing issues themselves. As a result, the teacher had a strong influence on the learning processes of that group.

Group 5 had also a strong task focus. This is clearly visible in Figure 9.12. Group 5 scored high on diagnostic reasoning and received limited support from the teacher. It seems that this group performed fairly well. A closer look at the discussion revealed that two group members dominated the discussion while one group member hardly participated in the discussion (see Table 9.9).

Coordination also differed for the two conditions. Coordination for the CoFFEE groups included communication that drew the attention of the group to something that happened in the shared workspace. Focusing the attention on CoFFEE was done verbally. Student frequently reacted verbally on something that was written in the shared digital workspace.

187

Figure 9.12: Functions of task-related communication.

Teacher's intervention

Coordination

Factual knowledge

Diagnostic reasoning

Feedback

0

50

100

150

200

250

300

350Group 1Group 2Group 3Group 4Group 5

No.

of s

ente

nces

.

Chapter 9

Another difference was the kind of feedback provided by the group members. Feedback in the face-to-face condition was generally about establishing agreement between group members, e.g. “I think this is correct” or “Down syndrome has to do with health”. It hardly triggered any further elaborations. Feedback in the CoFFEE condition was generally about differences in thinking. Differences of thinking occurred in the digital workspace because that workspace allows for much more freedom in thinking. Examples of this kind of feedback are: “I think that what you say is a problem”, “Is not eating a serious problem?”, and “That’s a symptom because it is visible”. Two of the three CoFFEE groups explicitly mentioned to the teacher that discussing their differences in thinking was characteristic for their way of working.

Representation Aids

The representational aids of the collaborative tools should makes students aware of task-relevant information structures. Two information structures were implemented during the first part of the discussion when the students worked with the Threaded-discussion tool. The first structure was Gordon's health patterns that were represented as different categories (see Figure 9.2). These categories were constantly displayed in the left window of the digital workspace. Students organized the health cues by associating them with the appropriate category. The second structure was the PES format (see Figure 9.3). This representational aid was not directly visible. Students could attach a label to a contribution by selecting that contribution, right click with the mouse and select the appropriate label from a drop down menu (see Figure 9.3).

All the groups frequently referred to Gordon and PES during the verbal part of the

188

No. of verbal references (verbally) No. of references in CoFFEE

Gordon's health patterns

PES PES

Group 1 3 1 4

Group 2 3 9 26

Group 3 2 1 6

Group 4 1 6

Group 5 17 3

Table 9.16: Number of verbal references to Gordon and PES.

A Diagnostic-reasoning Discussion withthe Support of CoFFEE

discussion (Table 9.16). The last column of Table 9.16 also displayed the number of times that Group 1, 2 and 3 refers to the PES in the shared workspace (post-hoc notation).

Interaction Sequences

So far we focused on single contributions. The task-related communication was classified on the level of the individual group members. A discussion, however, consists of a sequence of individual talk that is organized in a coherent and meaningful manner. It would be interesting to see how the communication is patterned because the sequence of related contributions gives an insight in the process of knowledge elaboration. It tells us, for example, how the group members build a shared representation of the problem space. We used the “double interact” (Weick, 1979) to identify related communicative acts. A sequence consists of minimal three related acts made by at least two group members. Furthermore, a change of topic signals a new sequence. These two criteria were used to extract the sequences from the verbal, task-related communication. The first column of Table 9.17 gives the total number of task-related sequences for each group. The second column displays the average length of these sequences as the number of separate contributions. The figures that are given in the second column can also be read as the average number of turns. From these figures one can conclude that the sequence length of the CoFFEE supported groups is smaller than that of the groups that only communicated verbally. A cause for this difference may be that the communication for groups 1, 2 and 3 constantly alters between the digital “talk” and verbal talk.

189

Task-related sequences (verbal)

Total number of sequences

Length(no. of

contr./seq.)

Teacher led sequences

Peer led sequences

Group 1 23 9,8 11 12

Group 2 41 9,6 2 39

Group 3 25 7,4 8 17

Group 4 39 16,6 20 19

Group 5 40 12,6 2 17

Table 9.17: Number of contributions in CoFFEE (Threaded discussion tool).

Chapter 9

Discussion

Group communication is considered to be the functional means by which the groups accomplish their learning goal and, even more important, groups themselves are best regarded as emerging from or constituted in communication (Frey, 1999). Communication turns the individual group members into a recognizable whole and it lays the foundation of their performance as a group. A communication pattern that deals with the group as a whole is that of participation. The communication that was mediated by the collaborative tool showed a different pattern of participation. The groups who had an uneven pattern of participation during the verbal part of the discussion showed a fairly equal participation digitally.

Communication in CoFFEE was primarily task-oriented. The students used the representational aids of the Threaded-discussion tool to frame their discussion. For example, every group used the categories and the post-hoc notation system in an appropriate manner.

The task-related communication was not limited to CoFFEE. There is a complex interplay between the verbal and computer-mediated part of the communication. The permanence of the contributions in the digital workplace makes constant reflection possible. Students do not only put forward their thoughts, they also read what other group members have written and they respond to these statements. This is done digitally but also verbally. Verbal communication is used for coordination, for quick responses (e.g. factual knowledge), and to align the group's view with regard to the differences that are visible in the shared workspace.

190

10 Conclusions

One can and must communicate; it is a useful and easy way to contribute to the peace of others and one’s own, because silence, the absence of signals, is in its turn a signal, but it is ambiguous,

and ambiguity generates anxiety and suspicion.

Primo Levi1

The research that is described in this thesis goes into two major topics: one of practical nature and one of more methodological nature. The first topic had to do with the objective of the research: the development of two collaborative tools that facilitate face-to-face discussions. These tools mediate and regulate part of the communication so that it becomes easier for the group to elaborate on the knowledge that is shared by its members. A central question that guided the research was how that support should look like in order to be effective.

The second topic of methodological nature had to do with the research approach that provided us with the rationale to intervene in an existing learning situation in a scientific manner. The research that is discussed in this thesis was governed by a design-based approach. This approach can be characterized by the development of interventions that aim to improve a group discussion for the better. In our case, the approach exploits the design of a collaborative tool as a means to develop a deeper understanding of mediation and group discussions. The issue that we faced at the start was how such a design-based research approach will look like in practice.

Both issues are addressed in this concluding chapter. First, we reflect on the changes that have been observed during a discussion that has a verbal and a digital component. Next, we discuss our experiences with the design-based research approach.

1 Primo Levi (1986). The drowned and the saved.

191

Chapter 10

10.1 Mediated Face-to-Face CommunicationIn this thesis, we focused on small-group discussions in the classroom. These discussions enable students to develop a better understanding of the knowledge domain. A productive discussion in terms of learning outcomes puts some demands on the communication that the group has to display. The first research question (Q1) addressed this issue. In chapter 4, we formulated four criteria for a constructive dialogue. These criteria stem from prior research into group learning. These studies focused on performance differences and related these differences to group communication. The first criterion states that the communication must be oriented towards effective task performance because learning is associated with task performance. Secondly, the group members have to participate fairly equally so that they are able to share their knowledge with the group. Thirdly, the communication must be coherent in the sense that a contribution relates to other contributions. Finally, a discussion should show frequent knowledge elaborations whereby several speakers contribute to the discussion.

The teachers that participated in the studies reported that a discussion is sometimes controlled by one or more dominant students. In chapter 6, we argued that these observations were confirmed by previous research. The citation of Robert Bales (2002) at the beginning of chapter 6 makes this eloquently clear. Interpersonal dominance is one of the ineffective communication patterns that the second research question (Q2) aims to identify. The study in chapter 9 returns to this issue of control. Results from this study indicate that the direction of control varies from group to group. Interpersonal dominance could result in a discussion that is governed by interpersonal conflict, by unbalanced task-related communication or by a strong tendency to ask for help from the teacher. Characteristic for all these forms of an unbalanced discussion was that the sharing of knowledge was hampered.

We traced the issue of interpersonal dominance back to the way that the verbal exchanges are patterned. In chapter 6, we argued that frequent violations of the turn-taking rule – like interruptions or abrupt topic change – make a verbal discussion seemingly irrational. This analysis provided us with the ground for answering the third research question (Q3) that aims to indicate the relationship between the ineffective communication patterns and their underlying cause. The underlying cause can be

192

Conclusions

typified as an issue of coupling.

Coupling

Face-to-face discussions can be characterized by multiparty talk that is episodic in nature (Schwartzman, 1989). The episodic character implies that the utterances of the group members are only loosely coupled. A sequence of individual talk becomes a meaningful dialogue when the group manages to organize the exchanges into a comprehensible whole. A meaningful verbal discussion, for example, refers to the course of actions enacted through turns-at-talk to create coherence, order and meaning (Schegloff, 2007). Coherence, order and meaning should not be taken for granted. Group members have to follow certain principles and rules for coupling to make their individual actions into a productive discussion.

A central premise of this thesis is that each medium for communication provides specific rules for coupling. These rules do not only determine what can be expressed and how, they also affect the behavior of the group as an interacting system. In our case, we compared verbal communication with computer-mediated communication. The two modes of communication were associated with different rules for coupling: turn taking and adjacency pair for the verbal part of the communication and parallel access and global coherence for the computer-mediated part of the communication. This sets the direction for answering the fourth research question (Q4) that aims to find out how the collaborative tools change the group communication for the better. In chapter 6, we further specified this question by a set of expectations or hypotheses about how the face-to-face communication will change. These hypotheses guided the tool design.

Parallel Access as Floor-control Mechanism

The orientation taken by the research – computer-mediated communication that parallels verbal communication – provides an alternative perspective to the study of networked learning. Usually, educational researchers set face-to-face discussions apart from online discussions. The two situations are incompatible with distance as an essential characteristic. A number of studies compared face-to-face with digital communication (see e.g. Meyer, 2003; Jonassen & Kwon, 2001; Marttunen & Laurinen, 2001). This is mainly done at a general level: the two situations are described in general terms with access to information as the determining factor. Such a perspective was of no help in our case. The computer-mediated communication that is

193

Chapter 10

described in this thesis is also face-to-face. So overcoming distance is not an issue for support. This drew our attention to the characteristics of face-to-face communication.

Face-to-face communication offers many opportunities for collaboration and learning. It has been with us from the beginning and it is something that we have in common with our closest relatives. Face-to-face communication among primates, for example, seems exceedingly complex: monkeys and apes integrate each message with previous and ongoing actions, so that not only the detailed structure of the call, but also its context contributes to the eventual meaning (Jolly, 1999). Still, as we stated in chapter 6, face-to-face discussions have some drawbacks that relate to the way humans communicate verbally. We argued that these difficulties could be traced back to the coupling of individual utterances to generate order, coherence and meaning.

We identified interpersonal dominance as an ineffective communication pattern that hampers the sharing of knowledge. This verbal communication pattern was further described by the rules that unify individual speech into a meaningful discourse. The rules for sequential organization that characteristic for verbal exchanges are turn taking and adjacency pair. Turn taking refers to the rule that only one person talks at a time and the frequent changes between speakers do occurs. Adjacency pair or “nextness” is a basic unit for sequence construction whereby the meaning of an utterance relates to the preceding utterance. This relationship is central to the ways in which talk-in-action is organized and understood (Schegloff, 2007).

Unequal participation

At the start of the research, we identified unequal participation as one of the drawbacks of a verbal, face-to-face discussion. Unequal participation means that not all the group members are able to participate fully in the discussion. Some group members have fewer opportunities to share their knowledge with the group. This hampers the performance because the group does not make full use of knowledge and skills available. We traced that communication problem back to characteristics of the medium, i.e. human speech. Unequal participation was associated with interpersonal dominance that manifests itself by talking out of turn. Talking out of turn was further described in terms of the turn-taking rule that organizes the sequence of verbal exchanges. A group member is able to control a verbal discussion by frequently taking the turn through interruptions, repeatedly taking turns and simultaneous speech.

The attention for turn taking as underlying cause brought us to “product

194

Conclusions

blocking”, another ineffective communication pattern that is associated with verbal communication. Product blocking refers the principle that only one group member speaks at the same time. This prevents other group members to share their ideas with the group. The delay that these members experience hampers the free exchange of ideas. An idea that is not verbalized directly will fade into the background due to a shift in attention or because the topic of the discussion changes.

The direction of change

We expected that parallel access as floor-control mechanism would counteract the negative consequences of interpersonal dominance and product blocking. Parallel access became one of the basic properties of the collaborative tools. Parallel access makes it possible to put forward a contribution without any interruption or delay because all the group members have unhindered access to the discussion. We expected that parallel access affects the process of control that has to do interpersonal dominance that emerges when a group collaborates. The study that is discussed in chapter 9 evaluated the effects of parallel access. The digital part of the discussion that uses parallel access showed an even pattern of participation. This finding indicates that the digital medium allows students to share their ideas freely with their peers. Parallel access changed the pattern of group communication in several ways: students worked at their own pace, they focused on a topic of immediate interest, and they expressed their thoughts without being interrupted.

Equal participation in the digital workspace did not expand to the verbal part of the discussion. Some group members who participated much in the digital part of the discussion, hardly contributed to the verbal discussion. Dominance was still clearly visible during the verbal discussion. From these findings, one can conclude that the medium reduces or stimulates certain individual behaviors on the level of the group.

Parallel Access and Global Coherence

Turn taking as floor control mechanism is closely associated with adjacency pair as unit for sequence construction (Schegloff, 2007). Adjacency pair works less well for parallel access because there is no clear time frame. Text-only communication like chat shows a high degree of disrupted adjacency, overlapping exchanges and topic decay (Herring, 1999). Online chat-rooms, for example, constitute a communication environment where these basic rules and assumptions of conversation do not hold (Greenfield &

195

Chapter 10

Subrahmanyam, 2003). We observed the same problem. The study that is discussed in chapter 7 showed that the shared digital workspace became crowded with contributions that lacked a clear relationship. To solve this issue we developed some additional guidelines:

– a shared digital workspace that is divided into functional spaces,

– associate the functional spaces with a macro structures that is associated with the problem or the task, and

– the possibility to link a new contribution to a contribution that is already present in the shared workspace.

The three properties mentioned above are the elementary sequential units for the computer-mediated part of the communication. We expected that they would create coherence, order and meaning in a shared digital workplace. The properties organize a digital discussion based on a global structure of coherence. Global coherence refers to a macro structure that is based on topics or themes in a discourse (van Dijk, 1985). Such a macro structure was implemented in the collaborative tools whereby we took in mind that the digital medium had spatial properties. The two-dimensional character of the shared workspace was used to break through the temporal limitations of verbal talk. The digital communication does not only reflect a temporal ordering, it also displays spatial relationships that are based on a macro structure and graphical relationships:

– categories and connections for the Threaded discussion tool, and

– meaningful areas and for the Graphical tool.

The studies of chapters 8 and 9 show how such a macro structure affects the discussion. Several discussion lines occurred in parallel and most of them remained active during the whole discussion. These discussion lines addressed specific topics that had a direct relation with the macro structures, i.e. categories or meaningful areas. A global model of coherence – for example, various topics that are displayed by different categories – became more apparent. It seems that such a macro structure broadens up the discussion, i.e. they gave the students more freedom to follow their own lines of thinking.

In chapter 6, we explicitly distinguished local coherence from global coherence.

196

Conclusions

They are associated with the two forms of communication, i.e. verbal and computer-mediated. In practice, a verbal discussion also displays some form of global coherence. A problem-solving discussion, for example, usually follows a general sequence like problem analysis, idea generation, evaluation and selection. However, such a structure is not that self-evident as in the case of the collaborative tools where they are implemented as representational aids.

Mediated Communication as an External Representation

Licklider and Taylor published an article in 1968 that discussed the use of computers during face-to-face meetings. Computer support according to their view could help participants to align their mental models. This stresses the modeling function of the computer: the computer could be used as an “interactive, cooperative modeling facility” (Licklider & Taylor, 1968). The authors presented an example where a speaker shares more detailed information about the topic by means of the computer. Listeners who have a question can check that information without interrupting the speaker. The examples that Licklider and Taylor provided in their article differ from the kind of collaborative activities that has been addressed in this thesis. Still, the modeling function is an important aspect of the computer support. The students in our studies constructed a shared representation of their digital discussion that remained visible for the whole time. The contributions that are put forward in the shared digital workspace had a sense of permanence. As a result, they could be revisited continuously. The students frequently reflected on what was written down digitally. Their reflections about what they read in the digital workspace were done digitally but also verbally.

Task-related communication

Results indicate that only the task-related communication was represented in the shared digital workspace. This part of the communication mainly addressed issues of uncertainty that are caused by ignorance or imprecision of a shared interpretation. The information that was shared digitally enabled students to form a more precise interpretation.

The students worked relatively undisturbed within the digital workspace. They experienced less influence from their group members when they expressed their thoughts into words. Working with the collaborative tools seems to stimulate cognitive freedom within the context of the group. Consequently, the contributions that were

197

Chapter 10

placed in the digital workspace diverged much more. This made the students aware of their differences in thinking. The group solved these differences verbally. One can conclude that “convergence of meaning” (Roschelle, 1992) happens verbally rather than digitally. The collaborative tool seems to be a less appropriate medium for the alignment of individual cognitive models into a group product. The tool stimulates differences in thinking but these differences are identified and discussed verbally.

Changing Group Dynamics

The different ideas that exist within a group form the basis of the collaborative learning. It creates a situation where elaboration, interpretation, explanation, and argumentation are integral to the activity of the group (Webb & Palincsar, 1996). The existence of different ideas does not necessarily mean that the group will assess their viewpoints critically. The tension between the individual and the group, between unhampered thinking and conformity may affect the learning potentials negatively. Beneath the surface of most groups flows an undercurrent that pushes the group together, towards greater consensus, uniformity, homogeneity or conformity (Forsyth, 1990). Janis (1972), for example, introduced the concept of “groupthink” for cohesive groups in a stressful situation who avoid disagreement and insufficiently use relevant information for decision-making. Groupthink promotes quick compromises in group decision-making (Hollingshead et al., 2005).

The collaborative tools keep the discussion open. The students become more aware of their differences in thinking. There is also a current in the opposite direction towards conformity that is visible in the verbal part of the communication. The group members align their cognitive models verbally.

A Concluding Remark

The design-based orientation of the research determines the process of fact-finding. It drew our attention to the communication problems that can be facilitated by a collaborative tool and that would improve learning. In that sense, we adopt a rather pragmatic approach. Theoretical insights were mainly used to set the direction for change. The quest for performance improvement led to adjustments of initial expectations. These adjustments made the tool more appropriate for its context of use. The research findings can only be placed within that perspective which makes any general conclusion conditional.

198

Conclusions

10.2 The Design-based Research ApproachIn chapter 2, we presented the design-based research approach that guided the research activities. The approach exploits the design process as a means to develop an understanding of small-group discussions in the classroom. It involves systematically engineering a learning environment in a way that allows the researcher to improve and generate evidence-based claims about learning (Barab & Squire, 2004).

Practical problems and expectations about solutions are the basis for the design-based research approach. Communication problems are the immediate cause for research. Subsequently, the solutions for these problems are based on clear expectations about how the collaborative tools will change the pattern of communication. These expectations were presented as a set of hypotheses that describe the direction of change. They were translated into a set of empirical verifiable design guidelines that lay down the basic properties of the collaborative tools. These guidelines define the changes that the tools will bring about. Therefore, an evaluation of the collaborative tools becomes an evaluation of the guidelines, and hence an evaluation of hypotheses and the theoretical framework that constitutes the design. In practice, it took us a number of cycles of research to make the envisioned solutions more applicable. Initial research led to a refinement of the design and a readjustment of the hypotheses.

To summarize, the design-based research approach could be characterized by a process of understanding and integration. Understanding increases because the approach continuously connects problems with solutions, practice with theory and design with usage.

Three Different Research and Design Cycles

The research approach as it has been described in chapter 2 has a strong theoretical basis. The theory-based research cycle is only one part of the story. In practice, there we two additional cycles that emphasize the close relationship between tool design and practice (Figure 10.1). The three cycles are:

1. usability studies, and

2. a practice-based research cycle, and

3. a theory-based research cycle.

199

Chapter 10

Usability studies

Several usability issues were addressed during the tool design. These issues referred to the ease by which the students used the collaborative tools. Usability included issues like the layout of the tools, information provided to the user, and the sequence of actions that users carried out. These issues can be very annoying for users. They may prevent the effective use of the tools. It meant that these issues were solved before any educational research could be carried out.

The usability studies took place before the collaborative tools were introduced in the classroom. Findings from these studies did not change the collaborative tools fundamentally. They led to improvements of the user-interface.

Practice-based Research and Design

The tool design was also guided by observations in the classroom that led directly to the specification of requirements. For example, the CoFFEE system has a Quick Communication Service to allow students to communicate with the teacher. The Quick communication service is a chat tool that enables students to ask questions to the teacher. This service mirrors a classroom activity: raising one’s hand to ask a question to the teacher. To conclude, some classroom activities were directly translated into an activity that is supported by the software system. This practice-based cycle is an

200

Figure 10.1: The three developmental cycles of the research approach.

Conclusions

adaptation of the initial research approach. It means that the robustness of the tools stems from: 1) theoretical framework that relates communication problems to solutions, and 2) empirical observations that directly fed back into the design.

Theory-based research and design

The design-based research approach can be typified as a modeling mode of inquiry in which theory, operationalization, and data patterns are treated simultaneously (Poole, McPhee & Canary, 2002). This mode can be contrasted with the hypothetical-deductive approach and grounded modes of inquiry. The hypothetical-deductive approach sets theory prior to data collection and hypothesis testing, and these steps are strictly separated. Within a grounded mode of inquiry concepts, hypotheses and theoretical propositions are generated directly from experience with the data (Poole, McPhee & Canary, 2002).

The role of theory

Scientific studies usually begin with a more precise specification of the problems that give rise to the research. These research problems can be internal to a theory and arise from criticism of that theory (Popper, 1999). This criticism leads to more fundamental research that aims to test the theory. The researcher formulates hypotheses or tentative answers to test relationships. These relationships are usually studied through experimentation where the researcher creates an artificial research environment. The researcher systematically manipulates some features of the environment and exercise great control on the other features. To test hypotheses, the researcher observes or measures the effects of the changes (Singleton & Straits, 2005).

In our case, the research problems were not associated with a particular theory but originated from practice. This had important implications for the research. We did not start with a clear theory from which we could generate research questions and that would guide the subsequent research activities. We did not opt for a normative approach that is based on a general theory about how students learn. Such an approach would deduce design guidelines from a general theoretical framework. Instead, we adopt a rather pragmatic approach that includes a search for theoretical insights that could help us to understand the communication problems that triggered the research. Theories, according to psychologist and philosopher William James (1907) become instruments, not answers. Theories operate at a micro-level, in the sense that they are used to create specific processes. They create the phenomena of interest, not simply

201

Chapter 10

understand or explain them (Argyris, 1997).

Hypotheses

Hypotheses in the design-based research have a specific objective. They have less to do with a logical statement that is tested through carefully controlled experiments. Instead, the hypotheses indicate an expectation about a direction of change that is only rudimentary expressed at the beginning of the research. Research enables us to come up with more accurate expectations that display a deeper understanding of the phenomenon under study.

Characteristic for the research approach was that we did not focus on the artifact as a whole. Rather, the impact of the artifact was reduced to a limited set of propositions and hypotheses that lay down the relationship between the basic properties of the tool and their intended effects. Narrowing down the artifact to a limited set of hypotheses had strong implications for research. We carried out research on a micro level. We did not come up with extensive cases descriptions that gave a comprehensive account of the new situation. Rather, we restricted the analysis mainly to the communication processes and focused on the relationship between the two media and the communication that occurred in these media.

A concluding remark

When we look back at the beginning of the research project, we may conclude that we only had a rudimentary notion of how a collaborative tool could change the face-to-face communication for the better. That notion, incomplete but appealing, served as an important objective for research. A further elaboration of the initial idea made us aware that the reality of groups is much more complex. Our research can be typified by a constant struggle to master that reality. A clearly formulated research approach helped us to overcome these hurdles.

202

References

Abrakam, R.H. & Shaw (1982). Dynamics: The geometry of behavior. Santa Cruz, CA: Aerial Press.

Ajzen, I. (1988). Attitudes, Personality, and Behavior. Milton Keynes, UK: Open University Press.

Alexander, C. (1964). Notes on the Synthesis of Form. Cambridge, MA: Harvard University Press.Alexander, C. (1979). The Timeless Way of Building. Oxford: Oxford University Press.Alexander, C., Ishikawa, S., Silverstein, M., Jacobson, M., Fiksdahl-King, I. & Angel, S. (1977).

A Pattern Language: Towns, Buildings, Construction. New York, NY: Oxford University Press.Allen, T.H. & Plax, T.G. (2002). Exploring consequences of group communication in the

classroom: Unraveling relational learning. In L.R. Frey (Ed.), New Directions in Group Communication (pp. 219-234). Thousand Oaks, CA: Sage Publications.

Anderson, J.R., Corbett, A.T., Koedinger, K.R. & Pelletier, R. (1995). Cognitive tutors: Lessons learned. The Journal of the Learning Sciences, 4(2), 167-207.

Anderson, T. (2008). The Theory and Practice of Online Learning. Edmonton, Canada: AU Press, Athabasca University.

Angyal, A. (1941). Foundations for a Science of Personality. Cambridge, MA: Harvard University Press.

Archer, M.S. (1995). Realist Social Theory: The Morphogenetic Approach. Cambridge, UK: Cambridge University Press.

Argyris, C. (1997). The Aston studies and design causality. In T. Clark (Ed), Advancement in Organizational Behaviour: Essays in Honour of Derek S. Pugh (pp. 73-87). Aldershot, England: Ashgate Publishing.

Arrow, H., McGrath, J.E. & Berdahl, J.L. (2000). Small Groups as Complex Systems: Formation, Coordination, Development, and Adaptation. Thousand Oaks, CA: Sage Publications.

Bailey, J.E. & Pearson, S.W. (1983). Development of a tool for measuring and analyzing computer user satisfaction. Management Science, 29(5), 530-545.

Baker, M. (2002). Forms of cooperation in dyadic problem-solving. Revue d’Intelligence Artificielle, 16(4-5), 587-620.

Baker, M. & Lund, K. (1997). Promoting reflective interactions in a computer-supported collaborative learning environment. Journal of Computer Assisted Learning, 13, 175-193.

Bales, R.F. (1950). A set of categories for the analysis of small group interaction. American Sociological Review, 15(2), 257-263.

Bales, R.F. (2002). Social Interaction Systems: Theory and Measurement. New Brunswick, NJ: Transaction Publishers.

Barab, S. & Squire, K. (2004). Design-based research: Putting a stake in the ground. The Journal of the Learning Sciences, 13(1), 1-14.

Barge, J.K. (2002). Enlarging the meaning of group deliberation: from discussion to dialogue. In L.R. Frey (Ed.), New Directions in Group Communication (pp. 159-177). Thousand Oaks,

203

References

CA: Sage Publications.Barron, B. (2003). When smart groups fail. The Journal of the Learning Sciences, 12(3), 307-359.

Beiser, A. (1981). Concepts of Modern Physics (3rd edition). Tokyo, Japan: McGRaw-Hill.Bereiter, C., Scardamalia, M., Cassells, C. & Hewitt, J. (1997). Postmodernism, knowledge

building, and elementary science. The Elementary School Journal, 97(4), 329-340.Bliss, J. & Säljö, R. (1999). The human-technological dialectic. In J. Bliss, R. Säljö & P. Light

(Eds), Learning Sites: Social and Technological Resources for Learning (pp. 1-11). Oxford, UK: Elsevier Science.

Boon, L. (1989). Geschiedenis van de Psychologie (in Dutch). Amsterdam, The Netherlands: Boom.

Bosman, A. (1977). Een Metatheorie over het Gedrag van Organisaties (in Dutch). Leiden, The Netherlands: Stenfert Kroese.

Boyle, J.T. & Nicol, D.J. (2003). Using classroom communication systems to support interaction and discussion in large class settings. Research in Learning Technology, 11(3), 43-57.

Bransford, J.D., Brown, A.L. & Cocking, R.R. (2000). How People Learn: Brain, Mind, Experience and School. Washington DC: National Academy Press.

Brauner, E. & Scholl, W. (2000). The information processing approach as a perspective for group research. Group Processes & Intergroup Relations, 3(2), 115-122.

Briggs, R.O. (2004). On theory-driven design of collaborative technology and process. In G.J. De Vreede, L.A. Guerrero & Raventós (Eds.), CRIWG 2004 Groupware: Design, Implementation and Use (pp. 1-15). Berlin, Germany: Springer-Verlag.

Briggs, R.O., de Vreede, G.J., Nunamaker, J.F. & Tobey, D. (2001). ThinkLets: Achieving predicatable, repeatable patterns of Group interaction with Group Support Systems (GSS). Proceedings of the Thirty-Fourth Hawaii International Conference on System Science (pp. 436-444). Los Alamitos, CA: IEEE Computer Society Press.

Brockbank, A. & McGill, I. (2000). Facilitating Reflective Learning in Higher Education. Buckingham, UK: Open University Press.

Brown, A.L. (1992). Design experiment: Theoretical and methodological challenges in creating complex interventions in classroom settings. The Journal of the Learning Sciences, 2(2), 141-178.

Bruner, J.S. (1996). The Culture of Education. Cambridge, MA: Harvard University Press.Brydon-Miller, M., Greenwood, D. & Maguire, P. (2003). Why action research. Action

Research, 1(1), 9-28.Bull, P. (2002). Communication under the Microscope: The Theory and Practice of Microanalysis.

New York, NY: Routledge.Bunge, M. (1979). A systems concept of society: beyond individualism and holism. Theory and

Decision, 10, 13-30.Bunge, M. (2004). How does it work? The search for explanatory mechanisms. Philosophy of the

Social Sciences, 34(2), 182-210.Bunge, M. (2000). Ten modes of individualism, none of which works, and their alternatives.

Philosophy of the Social Sciences, 30, 384-406.Burlingame, G.M. & Fuhriman, A. (1997). Small group research and dynamical systems

204

References

theory: Conceptual and methodological considerations, a reply to McGrath (1997). Group Dynamics: Theory, Research, and Practice, 1(1), 28-31.

Burr, V. (2003). Social Constructionism. East Sussex, UK: Routledge.Campbell, S.R. (2002). Constructivism and the limits of reason: Revisiting the Kantian

problematic. Studies in Philosophy and Education, 21, 421-445.Carr, W. & Kemmis, S. (1986). Becoming Critical: Education, Knowledge and Action Research.

London, UK: Deakin University Press. Cattell, R.B. (1965). The Scientific Analysis of Personality. Baltimore, MD: Penguin Books.Cazden, C.B. & Beck, S.W. (2003). Classroom discourse. In A.C. Graesser, M.A. Gernsbacher

& S.R. Goldman (Eds), Handbook of Discourse Processes (pp. 165-198). Mahwah, NJ: Lawrence Erlbaum.

Checkland, P. (1981). Systems Thinking, Systems Practice. New York, NY: John Wiley and Sons.Checkland, P. & Holwell, S. (2007). Action research: Its nature and validity. N.F. Kock (Eds.),

Information Systems Action Research: An Applied View of Emerging Concepts and Methods (pp 3-18). New York, NY: Springer Science.

Cheng, H.N.H., Chang, B., Deng, Y.C. & Chan, T.W. (2005). Matrixdesks: Interactive computing desks toward one-on-two educational computing environments. In T. Koschmann, D. Suthers & T.W. Chan (Eds.), Proceedings of the CSCL 2005: The Next 10 Years! (pp. 48-52) Mahwah, NJ: Lawrence Erlbaum Assoiates.

CHEPS scenarios (2004). The European Higher Education and Research Landscape 2020. Enschede, The Netherlands: Center for Higher Education Policy Studies, University of Twente.

Chernobilsky, E., Nagarajan, A. & Hmelo-Silver, C.E. (2005). Problem-based learning online: Multiple perspectives on collaborative knowledge construction. In D. Suthers, T. Koschmann & T.W. Chan (Eds.), Proceedings of the International Conference on Computer Supported Collaborative Learning (pp 53-62). Mahwah, NJ: Lawrence Erlbaum Associates.

Child, J. (1997). Strategic choice in the analysis of action, structure, organizations, and environment: Retrospect and prospect. Organizational Studies, 18, 43-76.

Cholowski, K.M. & Chan, L.K.S. (1992). Diagnostic reasoning among second-year nursing students. Journal of Advanced Nursing, 17, 1171-1181.

Churchman, C.W. (1971). The Design of Inquiring Systems: Basic Concepts of Systems and Organizations. New York, NY: Basic Books.

Clark, R.E. (1983). Reconsidering research on learning from media. Review of Educational Research, 53(4), 445-459.

Cobb, P., Confrey, J., diSessa, A., Lehrer, R. & Schauble, L. (2003). Design experiments in educational research. Educational Researcher, 32(1), 9-13.

Cohen, E.G. (1994). Restructuring the classroom: Conditions for productive small groups. Review of Educational Research, 64(1), 1-35.

Cohen, L., Manion, L. & Morrison, K. (2005). A Guide to Teaching Practice. New York: Routledge.

Coleman, E.B. (1998). Using explanatory knowledge during collaborative problem solving in science. The Journal of the Learning Sciences, 7(3&4), 387-427.

Collier, A. (1998). Language, practice and realism. In I. Parker (Ed), Social Constructionism, Discourse and Realism (pp. 47-58). London, UK: Sage Publications.

205

References

Collins, A. (1992). Towards a design science of education. In E. Scanlon & T. O’Shea (Eds), New Directions in Educational Technology (pp. 15-22). Berlin, Germany: Springer.

Collins, A., Joseph, D. & Bielaczyc, K. (2004). Design research: Theoretical and methodological issues. The Journal of the Learning Sciences, 13(1), 15-42.

Coskun, H., Paulus, P.B., Brown, V. and Sherwood, J.J. (2000). Cognitive stimulation and problem presentation in idea-generating groups. Group Dynamics: Theory, Research and Practice, 4(4), 307-329.

Coulson, A.J. (2006). The Cato Educational Market Index Full Technical Report. Retrieved July 16, 2009, from The Cato Institute Web site: http://www.cato.org/pub_display.php?pub_id=6851.

Crabtree, A. (2003). Designing Collaborative Systems: A Practical Guide to Ethnography. London: Springer-Verlag.

Crabtree, A., Hemmings, T. & Rodden, T. (2002). Pattern-support for interactive design in domestic settings. Proceedings of the Conference on Designing Interactive Systems: Processes, Practices, Methods, and Techniques (pp. 265-275), London, England.

Craig, R.T. & Tracy, K. (1983). Introduction. In R.T. Craig & K. Tracy (Eds.), Conversational Coherence (pp. 10-22). Beverly Hills, CA: Sage Publications.

Cummings, J.N. & Ancona, D.G. (2005). The functional perspective. In S.A. Wheelan (Ed), The Handbook of Group Research and Practice (pp. 107-118). Thousand Oaks, CA: Sage Publications.

Cushman, D.P. & Kovacic, B. (1994). Human communication: A rules perspective. In F.L. Casmir (Ed.), Building Communication Theories: A Socio-cultural Approach (pp 269-296). Hillsdale, NJ: Lawrence Erlbaum.

Cushman, D.P. & Whiting, G. (1972). An approach to communication theory: Towards consensus on rules. Journal of Communication, 22(3), 217-238.

Davis, A.M. (1990). Software Requirements: Analysis and Specification. Englewood Cliffs, NJ: Prentice Hall.

Davis, B. & Sumara, D. (2002). Constructivist discourses and the field of education: Problems and possibilities. Educational Theory, 52(4), 409-428.

de Jong, R. (1992). An Adaptive Trip Planning Support Environment. Delft, The Netherlands: Doctoral Dissertation, Delft University of Technology.

de Jong, T., Ainsworth, S., Dobson, M., van der Hulst, A., Levonen, J., Reimann, P., Sime, J.A., van Someren, M.W., Spada, H. & Swaak, J. (1998). Acquiring knowledge in science and mathematics: The use of multiple representations in technology-based learning environments. In: M.W. van Someren, P. Reimann, H.P.A. Boshuzien & T. de Jong (Eds), Learning with Multiple Representations (pp. 9-40). Amsterdam, The Netherlands: Elsevier Science.

De Lisi, R. & Golbeck, S.L. (1999). Implications of Piagetian theory for peer learning. In A.M. O’Donnell & A. King (Eds), Cognitive Perspectives on Peer Learning. Mahwah, NJ: Lawrence Erlbaum.

de Vries, E., Lund, K. & Baker, M. (2002). Computer-mediated epistemic dialogue: Explanation and argumentation as vehicles for understanding scientific notions. The Journal of the Learning Sciences, 11(1), 63-103.

206

References

Dewey, J. (1933). How we Think: A Restatement of the Relation of Reflective Thinking to the Educational Process. New York, NY: D.C. Heath and Company.

Diehl, M. & Stroebe, W. (1987). Productivity loss in brainstorming groups: Towards the solution of the riddle. Journal of Personality and Social psychology, 53, 497-509.

Dillenbourg, P. (1990). Introduction: What Do You Mean by “Collaborative Learning”? In P. Dillenbourg (Eds.), Collaborative Learning: Cognitive and Computational Approaches (pp. 1-29). Oxford, UK: Pergamon.

Dillenbourg, P., Baker, M., Blaye, A. & O’Malley, C. (1995). The evolution of research on collaborative learning. In P. Reimann & H. Spada (Eds.), Learning in Humans and Machines: Towards an Interdisciplinary Learning Science (pp. 189-211). Oxford, UK: Pergamon.

Draskovic, I., Holdrinet, R., Bulte, J., Bolhuis, S. and van Leeuwe, J. (2004). Modeling small group learning. Instructional Science, 32, 447-473.

Driscoll, M.P. (1994). Psychology of Learning for Instruction. Boston, MA: Allyn and Bacon.Dwyer, N. & Suthers, D. D. (2006). Consistent practices in arfifact-mediated collaboration.

International Journal of Computer-Supported Collaborative Learning, 1(4), 481-511.Ebbens, S. & Ettekoven, S. (2005). Effectief Leren (in Dutch). Groningen, The Netherlands:

Wolters-Noordhoff.Edelson, D.C. (2002). Design research: What we learn when we engage in design. The Journal

of the Learning Sciences, 11(1), 105-121.Emmer, E.T. & Stough, L.M. (2001). Classroom management: A critical part of educational

psychology, with implications for teacher education. Educational Psychologist, 36(2), 103-112.

Erkens, G. (2004). Dynamics of coordination in collaboration. In J. van der Linden & P. Renshaw (Eds.), Dialogic Learning: Shifting Perspectives to Larning, Instruction, and Teaching (pp. 191-216). Dordrecht, The Netherlands: Kluwer Academic Publishers.

Fisher, B.A. (1980). Small Group Decision Making: Communication and the Group Process. New York, NY: McGraw-Hill.

Fjermestad, J. (2004). An analysis of communication mode in group support systems research. Decision Support Systems, 37, 239-263.

Forsyth, D.R. (1990). Group Dynamcis (2nd edition). Pacific Grove, CA: Brooks/Cole Publishing.

Frensch, P.A. & Funke, J. (1995). Definitions, traditions, and a general framework for understanding complex problem solving. In P.A. Frensch & J. Funke (Eds.), Complex Problem Solving: The European Perspective (pp. 3-25). Hillsdale, NJ: Lawrence Erlbaum.

Frey, L.R. (1999). Introduction. In L.R. Frey, D.S. Gouran & M.S. Poole (Eds.), The Handbook of Group Communication Theory and Research. Thousand Oaks, CA: Sage Publications.

Gagné, R.M., Briggs, L.J. & Wager, W.W. (1992). Principles of Instructional Design, 4th Edition. Fort Worth, TX: Harcout Brace Jovanovich Publishers.

Galanes, G.J. & Adams, K. (2007). Effective Group Discussions: Theory and Practice. New York, NY: McGraw Hill.

Gallupe, R.B., Bastianutti, L.M. & Cooper, W.H. (1991), Unblocking brainstorms, Journal of Applied Psychology, 76, 137-142.

Garrison, D.R. (1992), Critical Thinking and Self-directed Learning in Adult Education: An

207

References

Analysis of Responsibility and Control Issues. Adult Education Quarterly, 42(3), 136-148. Gergen, K.J. (1995). Social construction and the educational process. In L.P. Steffe & J.E. Gale

(Eds), Constructivism in Education (pp 17-40). Hillsdale, NJ: Lawrence Erlbaum.Gick, M.L. & Holyoak, K.J. (1983). Schema induction and analogical transfer. Cognitive

Psychology, 15, 1-38.Gillies, R.M. (2004). The effects of cooperative learning on junior high school students during

small group learning. Learning and Instruction, 14, 197-213.Goffman, E. (1963). Behavior in Public Places: Notes on the Social Organization of Gatherings.

New York, NY: The Free Press.Goldberg, J.A. (1983). A move towards describing coherence. In R.T. Craig and T. Karen

(Eds.), Conversational Coherence. London: Sage.Goodman, N. (1976). Language of Art. Indianapolis, IN: Bobbs-Merrill.Goodyear, P. (2004). Patterns, Pattern Language and Educational Design. Paper presented at the

ASCILITE Conference Perth, Western Australia. Retrieved 10/10/2007 from http://ascilite.org.au/conferences/perth04/procs/pdf/goodyear.pdf.

Goodyear, P., Banks, S., Hodgson, V. & McConnell, D. (2004). Research on networked learning: An overview. In P. Goodyear, S. Banks, V. Hodgson & D. McConnell (Eds), Advances in Research on Networked Learning (pp. 1 -10). Dordrecht, Netherlands: Kluwer Academic Publishers.

Gordon, M. (1994). Nursing Diagnosis: Process and Application (3rd edition).St. Louis, MO: Mosby.

Gouran, D.S. & Hirokawa, R.Y. (1996). Functional theory and communication in decision-making and problem-solving groups. In R.Y. Hirokawa & M.S. Poole (Eds), Communication and Group Decision Making. Thousand Oaks, CA: Sage Publications.

Gravemeijer, K. (1994). Educational development and development research in mathematics education. Journal for Research in Mathematics Education, 25(5), 443-471.

Green, J.L. (1983). Research on teaching as a linguistic process: A state of the art. In E. Gordon (Ed.), Review of Research in Education (Vol 10, pp. 151-252). Washington DC: American Educational Research Association.

Greenfield, P.M. & Subrahmanyam, K. (2003). Online discourse in a teen chatroom: New codes and new modes of coherence in a visual medium. Applied Developmental Psychology, 24, 713-738.

Greeno, J.G. (1998). The situativity of knowing, learning and research. American Psychologist, 53(1), 5-26.

Guzzo, R.A. & Shea, G.P. (1992). Group performance and intergroup relations in organizations. In M.D. Dunnette, L.M. Hough & H.C. Triandis (Eds.), Handbook of Industrial and Organizational Psychology, Volume 3 (pp. 269-313. Palo Alto, CA: Consulting Psychologists Press.

Haberman, B. (2006). Pedagogical patterns: A means for communication within the CS teaching community of practice. Computer Science Education, 16(2), 87-103.

Hackman, J.R. (1987). The design of work teams. In J.W. Lorsch (Ed.), Handbook of Organizational Behavior (pp. 315-342). Englewood Cliffs, NJ: Prentice-Hall.

Hall, M.J.J., Hall, R. and Zeleznikow, J. (2003). A process for evaluating legal knowledge-based

208

References

systems based upon the Context Criteria Contingency-guidelines Framework. 9th

International Conference on Artificial Intelligence and LAW (ICAIL), Edinburgh, Scotland.Hare, A. P. (1960).The dimensions of social interaction. Behavioral Science. 5(3), 211-215.Hargie, O. and Dickson, D. (2004), Skilled Interpersonal Communication: Research, Theory and

Practice. London: Routledge.Harnack, R.V., Fest, T.B. & Schindler Jones, B. (1977). Group Discussion: Theory and Technique

(2nd edition). Englewood Cliffs, NJ: Prentice-Hall.Harrington, J. (1991). Organizational Structure and Information Technology. London, UK:

Prentice Hall.Hasenbein, M., Kopp, B. & Mandl, H. (2008). Collaborative problem solving with cases in a

virtual professional training. In J. Zumbach, N. Schwartz, T. Seufert & L. Kester (Eds.), Beyond Knowledge: The Legacy of Competence (pp. 59-62). New York, NY: Springer.

Heeren, E. (1996), Technology Support for Collaborative Distance Learning. Enschede, The Netherlands: University of Twente, Centre for Telematics and Information Technology.

Henri, F. (1992), Computer Conferencing and Content Analysis. In A.R. Kaye (ed.), Collaborative Learning through Computer Conferencing: The Najaden Papers. Springer-Verlag, Berlin.

Herring, S.C. (1999). Interactional Coherence in CMC. Proceedings of the 32nd Hawaii International Conference on System Science. Los Alamitos: IEEE Computer Society Press.

Hevner, A.R. & March, S.T. (2003). The Information systems research cycle. IEEE Computer, 26(11), 111-113.

Hevner, A.R., March, S.T., Park, J. & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75-105.

Hilgard, E.R. (1980). The trilogy of mind: Cognition, affection, and conation. Journal of the History of the Behavioral Sciences, 16, 107-117.

Hitchcock, D. (2002), The Practice of Argumentative Discussion. Argumentation, Vol. 16, 287-298.

Hoadley, C. (2002). Creating context: Design-based research in creating and understanding CSCL. In G. Stahl (Ed), Computer Support for Collaborative Learning 2002 (pp. 453-462). Mahwah, NJ: Lawrence Erlbaum.

Hogan, K., Nastasi, B.K. & Pressley, M. (2000). Discourse patterns and collaborative scientific reasoning in peer and teacher-guided discussions. Cognition and Instruction, 17(4), 379-432.

Hollingshead, A.B. & McGrath, J.E. (1995). Computer-assisted groups: A critical review of the empirical research. In R.A. Guzzo & E. Salas (Eds.), Team Effectiveness and Decision Making in organizations (pp 46-78). San Francisco, CA: Jossey-BassPublisher.

Hollingshead, A.B., Wittenbaum, G.M., Paulus, P.B., Hirokawa, R.Y., Ancona, D.G., Peterson, R.S., Jehn, K.A. & Yoon, K. (2005). A look at groups from the functional perspective. In Poole, M.S. & Hollingshead, A.B. (Eds), Theories of Small Groups: Interdisciplinary Perspective (pp. 21-62). Thousand Oaks, CA: Sage Publications.

Holmes, B. & Gardner, J. (2006). E-learning: Concepts and Practice. London: Sage Publications.Holsbrink-Engels, G.A. (1998), Computer-based Role Playing for Interpersonal Skill Training.

Enschede , The Netherlands: Doctoral Dissertation University Twente. Huff, A.S. (1990). Mapping Strategic Thought. New York, NY: John Willey.Isaksen, S.G. (1998). A Review of Brainstorming Research: Six Critical Issues for Inquiry. Creative

Problem Solving Group, Buffalo.

209

References

James, W. (1907). Pragmatism: A New Name for Some Old Ways of Thinking. Cambridge, MA: The Riverside Press.

Janis, I.L. (1972). Victims of Groupthink. Boston, MA: Houghton-Mifflin.Janis, I.L. (1982). Victims of Groupthink (2nd ed.).Boston, MA: Houghton-Mifflin.Janssen, J. (2008). Using Visializations to Support Collaboration and Coordination during

Computer-supported Collaborative Learning. Utrecht, The Netherlands: Doctoral Dissertation Utrecht University.

Janssen, M. (2001). Designing Electronic Intermediares: An Agent-based Approach for Designing Interorganizational Coordination Mechanisms. Delft, The Netherlands: Doctoral Dissertation, Delft University of Technology.

Johnson, D.W. (1980). Group processes: Influences of student-student interaction on school outcomes. In J. McMillan (Ed.), The Social Psychology of School Learning (pp. 123-168). New York, NY: Academic Press.

Jolly, A. (1999). Primate communication, lies, and ideas. In A. Lock & C.R. Peters (Eds.), Handbook of Symbolic Evolution (pp. 167-177). Oxford, UK: Blackwell Publsihers.

Jonassen, D.H. (1991). Objectivism versus constructivism: Do we need a new philosophical paradigm. Educational Technology Research & Development, 39(3), 5-14.

Jonassen, D.H. (1999). Designing constructivist learning environments. In C.M. Reigeluth (Ed), Instructional-design Theories and Models: A New Paradigm of Instructional Yheory (pp. 215-240). Mahwah, NJ: Lawrence Erlbaum Associates.

Jonassen, D.H. (2000). Computers as Mindtools for Schools: Engaging Critical Thinking. Upper Saddle River, NJ: Prentice Hall.

Jonassen, D.H. & Kwon, H.I. (2001). Communication patterns in computer mediated versus face-to-face group problem solving. Educational Technology Research and Development, 49(1), 35-51.

Jonassen, D.H., Peck, K.L. & Wilson, B.G. (1999). Learning with Technology: A Constructivist Perspective. Upper Saddle River, NJ: Prentice Hall.

Joyce, B. R., & Weil, M. (1980), Models of Teaching. Englewood Cliffs, NJ: Prentice Hall. Kanselaar, G., de Jong, T., Andriessen, J. & Goodyear, P. (2000). New technologies. In R.J.

Simins, J. van der der Linden & T. Duffy (Eds), New Learning. Dordrecht, The Netherlands: Kluwer Academic Publishers.

Kassirer, J.P. (1989). Diagnostic reasoning. Annals of Internal Medicine, 110(11), 893-900.Kearsley, G. P. (1976). Questions and question asking in verbal discourse: A cross-disciplinary

review. Journal of Psycholinguistic Research, 5, 355-375.Keefer, M.W., Zeitz, C.M. & Resnick, L.B. (2000). Judging the quality of peer-led student

dialogues. Cognition and Instruction, 18(1), 53-81.Keen, P. G. W. & Sol, H.G. (2008). Decision Enhancement Services: Rehearsing the Future for

Decisions that Matter. Amsterdam, The Netherlands: IOS Press.Keyton, J. & Frey, L.R. (2002). The state of traits: Predispositions and group communication.

In L.R. Frey (Ed.), New directions in Group Communication (pp. 99-120). Thousand Oaks, CA: Sage Publications.

Kiesler, S. & Cummings, J.N. (2002). What do we know about proximity and distance in work groups? In P. Hinds and S. Kiesler (Eds.), Distributed Work. Cambrige, MA: MIT Press.

210

References

Killen, R. (2007). Effective Teaching Strategies: Lessons from Research and Practice. Melbourne, Australia: Thomson.

King, A. (1994). Autonomy and question asking: The role of personal control in guided student-generated questioning. Learning and Individual Differences, 6(2), 163-185.

King, A. (1999). Discourse patterns for mediating peer learning. In A.M. O’Donnell & A. King (Eds.), Cognitive Perspectives on Peer Learning (pp. 87-115). Mahwah, NJ: Lawrence Erlbaum.

King, A., Staffieri, A. & Adelgais, A. (1998). Mutual peer tutoring: Effects of structural tutorial interaction to scaffold peer learning. Journal of Educational Psychology, 1, 134-152.

Kirschner, P.A., Sweller, J. & Clark, R.E. (2006). Why minimal guidance during instruction does not work: An analysis of the failures of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist, 41(2), 75-86.

Klahr, D. & Nigam, M. (2004). The equivalence of learning paths in early science instruction: Effects of direct instruction and discovery learning. Psychological Science, 15, 661-667.

Kock, N. (2003). The three treats of action research: A discussion of methodological antidotes in the context of an information systems study. Decision Support Systems, 37(2), 265-286.

Kozma, R.B. (1994). Will media influence learning? Reframing the debate. Educational Technology Research and Development, 42(2), 7-19.

Kozma, R.B. (1991). Learning with media. Review of Educational Research, 61(2), 179-211.Kneser, C. & Ploetzner, R. (2001). Collaboration on the basis of complementary domain

knowledge: Observed dialogue structures and their relation to learning success. Learning and Instruction, 11, 53-83.

Knott, A. & Sanders, T. (1998). The classification of coherence relations and their linguistic markers: An exploration of two languages. Journal of Pragmatics, 30, 135-175.

Kreijns, K., Kirschner, P.A. & Jochems, W. (2003), Identifying the pitfalls for social interaction in computer-supported collaborative learning environments: A review of research. Computers in Human Behavior, 19, 335-353.

Kyriacou, C. (2001). Essential Teaching Skills. Cheltenham, UK: Nelson Thornes.Lea, D. (1994). Christopher Alexander: An introduction for object-oriented designers. Software

Engineering Notes, 19(1), 39-46.Levine, R.L. & Fitzgerald, H.E. (1992). Systems and systems analysis: methods and

applications. In R.L. Levine & H.E. Fitzgerald (Eds.), Analysis of Dynamic Psychological Systems: Methods and Applications (pp. 1-16). New York, NY: Plenum Press.

Lewin, K. (1936). Principles of Topological Psychology. York, PA: McGraw-Hill.Lewin, K. (1947). Frontiers in group dynamics: Concept, method and reality in social science;

social equilibria and social change. Human Relations, 1, 5-41.Lewin, M.A. (2009). Kurt Lewin: His psychology and a daughter’s recollections. In G.A.

Kimble & M. Wertheimer (Eds), Portraits of Pioneers in Psychology, Volume III (pp. 105-118). Mahwah, NJ: Lawrence Erlbaum.

Licklider, J.C.R. & Taylor, R.W. (1968). The computer as communicative device. Science and Technology, April, 21-31.

Lindblom-Ylänne, S., Pihlajamäki, H. & Kotkas, T. (2003). What makes student group successful? Student-student and student-teacher interaction in a problem-based learning environment. Learning Environments Research, 6, 59-76.

211

References

Linn, M.C. & Burbules, N.C. (1993). Construction of knowledge and group learning. In K. Tobin (Ed.), The Practice of Constructivism in Science Education (pp. 91-119). Hillsdale, NJ: Lawrence Erlbaum Associates.

Littlejohn, S.W. (2002). Theories of Human Communication. Belmont, CA: Wadsworth/Thomson Learning.

Lkoundi, A. & van Woerden, W. (1997), Ontwerpen van Cases: Leren van Praktijkgevallen (in Dutch). Groningen, The Netherlands: Wolters-Noordhoff.

Mabry, E.A. (1999). The systems metaphor in group communication. In L.R. Frey, D.S. Gouran & M.S. Poole (Eds.), The Handbook of Group Communication Theory and Research (pp. 71-91). Thousand Oaks, CA: sage Publications.

March, S.T. & Smith, G.F. (1995). Design and natural science research on information technology. Decision Support Systems, 15, 251-266.

Marttunen, M. & Laurinen, L. (2001). Learning of argumentation skills in networked and face-to-face environments. Instructional Science, 29, 127-153.

Mason, L. (1998). Sharing cognition to construct scientific knowledge in school context: The role of oral and written discourse. Instructional Science, 26, 359-389.

Massey, A.P. & Wallace, W.A. (1996). Understanding and facilitating group problem structuring and formulation: mental representations, interaction, and representational aids. Decision Support Systems, 17, 253-274.

McGrath, J.E. (1984). Groups: Interaction and Performance.Englewood Cliffs, NJ: Prentice-Hall.McGrath, J.E. (1991). Time, interaction, and performance (TIP): A theory of groups. Small

Group Research, 22(2), 147-174.McGrath, J.E., Arrow, H. & Berdahl, J.L. (2000). The study of groups: Past, present, and

future. Personality and Social Psychology Review, 4(1), 95-105.van Meel, J.W. (1994). The Dynamics of Business Engineering: Reflections on Two Case Studies

within the Amsterdam Municipal Police Force. Delft, The Netherlands: Doctoral Dissertation, Delft University of Technology.

Mehan, H. (1979). Learning Lessons: Social Organization in the Classroom. Cambridge, MA: Harvard University Press.

Merton, R.K. (1967). On Theoretical Sociology. New York, NY: Free Press.Meyer, K.A. (2003). Face-to-face versus threaded discussions: The role of time and higher-order

thinking. Journal of Asynchronous Learning Networks, 7(3), 55-65Morgan, G. (1986). Images of Organization. Beverly Hills, CA: Sage publications.Moser, P.K. (1993). Philosophy after Objectivity: Making Sense in Perspective. New York, NY:

Oxford University Press.Munneke, L., Amelsvoort, M. & Andriessen, J. (2003). The role of diagrams in collaborative

argumentation-based learning. International Journal of Educational Research, 39, 113-131.Nardi, B.A. (1996). Studying context: A comparison of activity theory, situated action models,

and distributed cognition. In B.A. Nardi (Ed.), Context and Consciousness: Activity Theory and Human-computer Interaction (pp. 69-102). Cambridge, MA: The MIT Press.

Nemeth, C., Brown, K. & Rogers, J. (2001). Devil's advocate versus authentic dissent: Stimulating quantity and quality. European Journal of Social Psychology, 31, 707-720.

Newell, A. & Simon, H.A. (1972). Human Problem Solving. Englewood Cliffs, NJ: Prentice-

212

References

Hall.Newman, D.R., Johnson, C., Webb, B. and Cochrane, C. (1997), Evaluating the Quality of

Learning in Computer Supported Co-operative Learning. Journal of the American Society for Information Science, 48(6), 484-495.

Nijstad, B.A. (2000). How the Group Affects the Mind: Effects of Communication in Idea Generating Groups. Doctoral Dissertation Utrecht University, Utrecht.

Olson, G.M. & Olson, J.S. (2000). Distance matters. Human-Computer Interaction, 15, 139-178.

O'Neill, E.S., Dluhy, N.M. & Chin, E. (2005). Modelling novice clinical reasoning for a computerized decision support system. Journal of Advanced Nursing, 49(1), 68-77.

Overdijk, M. & van Diggelen, W. (2008). Appropriation of a shared workspace: Organizing principles and their application. Computer-Supported Collaborative Learning, 3, 165-192.

Packer, M.J. & Goicoechea, J. (2000). Sociocultural and constructivist theories of learning: ontology, not just epistemology. Educational Psychologist, 35(4), 227-241.

Palincsar, A.S. (1998). Social constructivist perspectives on teaching and learning. Annual Review of Psychology, 49, 345-375.

Pavitt, C. & Kline Johnson, K. (1999). An examination of the coherence of group discussions. Communication Research, 26(3), 303-321.

Pena-Shaff, J.B. and Nicholls, C. (2004), Analyzing Student Interactions and Meaning Construction in Computer Bulletin Board Discussion. Computers and Education, 42, 243-265.

Phillips, D.C. (1995). The good, the bad, and the ugly: The many faces of constructivism. Educational Researcher, 24(7), 5-12.

Pickel, A. (2007). Rethinking systems theory: A programmatic introduction. Philosophy of the Social Sciences, 37, 391-407.

Poole, M.S. (1999). Group communication theory. In L.R. Frey, D.S. Gouran & M.S. Poole (Eds.), The Handbook of Group Communication Theory and Research (pp. 37-70). Thousand Oaks, CA: Sage publications.

Poole, M.S., Keyton, J. & Frey, L.R. (1999). Group communication methodology. In L.R. Frey, D.S. Gouran & M.S. Poole (Eds), The Handbook of Group Communication Theory and Research (pp. 92-112). Thousand Oaks, CA: Sage Publications.

Poole, M.S., McPhee, R.D. & Canary, D.J. (2002). Hypothesis testing and modeling perspectives on inquiry. In M.L. Knapp & J.A. Daly (Eds), Handbook of Interpersonal Communication (pp. 23-72). Thousand Oaks, CA: Sage Publications.

Popper, K. (1999). All Life is Problem Solving. London, UK: Routledge.Propp, K.M. (1999). Collective information processing in groups. In L.R. Frey, D.S. Gouran

&M.S. Poole (Eds.), The Handbook of Group Communication Theory and Research (pp. 225-250). Thousand Oaks, CA: Sage Publications,.

Raizman, D. (2003). History of Modern Design. London, UK: Laurence King Publishing.Ratto, M., Shapiro, R.B., Minh Truong, T. & Griswold, W.G. (2003). The ActiveClass project:

Experiments in encouraging classroom participation. In B. Wasson, S. Ludvigsen & U. Hoppe (Eds), Designing for Change in Networked Learning Environments: Proceedings of the International Conference on Computer Support for Collaborative Learning. Dordrecht, The Netherlands: Kluwer.

213

References

Reason, P. & Bradbury, H. (2006). Handbook of Action Research. London, UK: Sage Publications.

Reigeluth, C.M. & Moore, J. (1999). Cognitive education and the cognitive domain. In C.M. Reigeluth (Ed.), Instructional-design Theories and Models: A new Paradigm of Instructional Theory (pp. 51-68). Mahwah, NJ: Lawrence Erlbaum.

Robertson, S.I. (2001). Problem Solving. East Sussex, UK: Psychology Press.Rogelberg, S.G., Barnes-Farrell, J.L. & Lowe, C.A. (1992). The stepladder technique: An

alternative group structure facilitating effective group decision making. Journal of Applied Psychology, 77(5), 730-737.

Romiszowski, A. (1984) Designing Instructional Systems: Decision Making in Course Planning. London: Hogan Page.

Romiszowski, A. J., Mulder, M., & Pieters, J. (1990). The Case-Study Methodology and Instructional Development. In D. Eastcott, B. Farmer, & B. Lantz (Eds.), Aspects of Educational and Training Technology London: Kogan Page.

Roschelle, J. (1992). Learning by collaborating: Convergent conceptual change. The Journal of the Learning Sciences, 2(3), 235-276.

Roschelle, J. & Teasley, S.D. (1995). The construction of shared knwoledge in collaborative problem solving. In C. O'Malley (Ed), Computer Supported Collaborative Learning (pp. 69-97). Berlin, Germany: Springer Verlag.

Rosenshine, B. (1987). Explicit teaching and teacher training. Journal of Teacher Education, 38(3), 34-36.

Sacks, H., Schegloff, E.A. & Jefferson, G. (1974). A simplest systematics for the organization of turn-taking for conversation. Language, 50(4), 696-735.

Salomon, G. (1979). Interaction of Media, Cognition, and Learning. San Franscico: Jossey-Bass.Sawyer, R.K. (2002). Unresolved tensions in sociocultural theory: Analogies with contemporary

sociological debates. Culture and Psychology, 8(3), 283-305.Sawyer, R.K. (2003). The mechanisms of emergence. Philosophy of the Social Sciences, 34(2),

260-282.Sayer, R.A. (1992). Method in social science: A Realist Approach. London, UK: Routledge.Schegloff, E.A. (2000). Overlapping talk and the organization of turn-taking for conversation.

Language and Society, 29, 1-63.Schegloff, E.A. (2007). Sequence organization of interaction: A primer in Conversation Analysis I.

Cambridge, UK: Cambridge University Press.Schrage, M. (1990). Shared Minds: The New Technology of Collaboration. New York, NY:

Random House.Schwartz, D.L. (1995). The emergence of abstract representations in dyad problem solving. The

Journal of the Learning Sciences, 4(3), 321-354.Schwartzman, H.B. (1989). The Meeting: Gatherings in Organizations and Communities. New

York, NY: Plenum Press.Schwarz, B. & De Groot, R. (2007). Argumentation in a changing world. Computer-Supported

Collaborative Learning, 2, 297-313.Scribner, S. (1997). A sociocultural approach to the study of mind. In E. Tobach, R.J.

Falmagne, M. Brown Parlee, L.M.W. Martin & A. Scribner Kapelman (Eds), Mind and

214

References

Social Practice: Selected Writings of Sylvia Scribner. Cambridge, UK: Cambridge University Press.

Sfard, A. & Kieran, C. (2001). Cognition as communication: Rethinking learning-by-talking through multi-faceted analysis of students’ mathematical interactions. Mind, Culture, and Activity, 8(1), 42-76.

Simon, H.A. (1996). The Sciences of the Artificial. Cambridge, MA: The MIT Press.Simons, R.J., van der Linden, J. & Duffy, T. (2000). New learning: Three ways to learn in a

new balance. In R.J. Simons, J. van der Linden & T. Duffy (Eds.), New Learning (pp. 1-20). Dordrecht, The Netherlands: Kluwer Academic Publisher.

Singleton, R.A. & Straits, B.C. (2005). Approaches to Social Research (4th ed.). New York, NY: Oxford University Press.

Slavin, R.E. (2000). Educational Psychology:Theory and Practice. Needham Heights, MA: Allyn & Bacon.

Slavin, R.E., Hurley, E.A. & Chamberlain, A. (2003). Cooperative learning and achievement: Theory and research. In W.M. Reynolds, G.E. Miller & I.B. Weiner (Eds.), Handbook of Psychology (pp. 177-198). Hoboken, NJ: John Wiley & Sons.

Smith, G.F. (1988). Towards a heuristic theory of problem structuring. Management Science, 34(12), 1489-1506.

Sol, H.G. (1982). Simulation in Information Systems Development. Groningen, The Netherlands: Doctoral Dissertation, University of Groningen.

Stahl, G., Koschmann, T. & Suthers, D. (2006). Computer-supported collaborative learning: An historical perspective. In R. K. Sawyer (Ed), Cambridge Handbook of the Learning Sciences (pp. 409-426). Cambridge, UK: Cambridge University Press.

Stangor, C. (2004). Social Groups in Action and Interaction. Hove, UK: Psychology Press.Stefik, M., Foster, G., Bobrow, D.G., Kahn, K., Lanning, S. & Suchman, L. (1987). Beyond

the chalkboard: Computer support for collaboration and problem solving in meetings. Communications of ACM, 30(1), 32-47.

Stein, N.L. & Albro, E.R. (2001). The origins and nature of arguments: Studies in conflict understanding, emotion, and negotiation. Discourse Processes, 32(2&3), 113-133.

Steiner, I.D. (1972). Group process and productivity. New York, NY: Academic Press.Stenning, K. & Oberlander, J. (1995). A cognitive theory of graphical and linguistic reasoning:

Logic and Implementation. Cognitive Science, 19, 97-140.Stewart, J., Bederson, B.B. & Druin, A. (1999). Single Display Groupware: A model for co-

present collaboration. Proceedings of the SIGCHI Conference on Human factors in Computing Systems (pp. 286-293) New York, NY: ACM.

Sugimoto, M., Hosoi, K. & Hashizume, H. (2004). Caretta: A system for supporting face-to-face collaboration by integrating personal and shared spaces. Proceedings of the SIGCHI conference on Human Factors in Computing Systems (pp. 41-48) New York, NY: ACM.

Susman, G.I. & Evered, R. (1978). An assessment of the scientific merits of action research. Administrative Science Quarterly, 23(4), 582-603.

Suthers, D.D., Hundhausen, C.D. & Girardeau, L.E. (2003). Comparing the roles of representations in face-to-face and online computer supported collaborative learning. Computers & Education, 41, 335-351.

Tang, J.C. (1991). Findings from observational studies of collaborative work. International

215

References

Journal of Man-Machine Studies, 34, 143-160.Tanner, C.A., Padrick, K.P., Westfall, U.E. & Putzier, D.J. (1987). Diagnostic reasoning

strategies of nurses and nursing students. Nursing Research, 36(6), 358-363.Taylor, C. (2000). Clinical problem-solving in nursing: Insights from the literature. Journal of

Advanced Nursing, 31(4), 842-849.Taylor, J.C. & Felten, D.F. (1993). Performance by Design: Sociotechnical Systems in North

America. Englewoord Cliffs, NJ: Prentice-Hall.Teurlings, C., van Wolput, B. & Vermeulen, M. (2006). Nieuw Leren Waarderen: Een

Literatuuronderzoek naar Effecten van Nieuwe Vormen van Leren in het Voortgezet Onderwijs (in Dutch). Utrecht, The Netherlands: Schoolmanagers_VO.

Twomey Fosnot, C. (1996). Constructivism: A psychological theory of learning. In C. Twomey Fosnot (Ed.), Constructivism: Theory, Perspectives and Practice. New York, NY: Teachers College Press.

Vallacher, R.R. & Nowak, A. (1994). The chaos in social psychology. In R.R. Vallacher & A. Nowak (Eds.), Dynamical Systems in Social Psychology (pp.1-16). San Diego, CA: Academic Press.

van Amelsvoort, M. (2006). A Space for Debate: How Diagrams Support Collaborative Argumentation-based Learning. Utrecht, The Netherlands: Doctoral Dissertation Utrecht University.

van Boxtel, C., van der Linden, J. & Kanselaar, G. (2000). Collaborative learning tasks and the elaboration of conceptual knowledge. Learning and Instruction, 10, 311-330.

van Diggelen, W. (2005). Annex I - Description of Work. The LEAD project (IST-2005-028027).Van Diggelen, W. (2008). Conceptual Models. The LEAD project (IST-2005-028027).van Diggelen, W. & Overdijk, M. (2006). The LEAD project: Computer support for face-to-

face collaborative learning. In W. Nejdl & K. Tochtermann (Eds.), Innovative Approaches for Learning and Knowledge Sharing: Proceedings of the First European Conference on Technology Enhanced Learning (pp. 579-584). Berlin, Germany: Springer.

van Diggelen, W. & Overdijk, M. (2007). Small-group face-to-face discussions in the classroom: A new direction of CSCL research. In C. A. Chinn, G. Erkens, & S. Puntambekar (Eds.), Mice, Minds, and Society: The 8th Conference on Computer Supported Collaborative Learning (pp. 726-735). New Brunswick, NJ: International Society of the Learning Sciences.

van Diggelen, W., Overdijk, M. & Andriessen, J. (2004). Constructing an Argumentative Map Together: Organizing Principles and their Application. Paper presented at the 2004 Dutch Educational Research Conference (ORD). Retrieved January 15, 2010 from http://edu.fss.uu.nl/ord/homepage.htm.

van Diggelen, W. & Scarano, V. (2006). Exploring the Potentials of Networked-computing Support for Face-to-Face Collaborative Learning: Workshop Proceedings EC-TEL 2006. Utrecht, The Netherlands: LEAD project.

van Dijk, T.A. (1985). Semantic discourse analysis. In T.A. Van Dijk (Ed.), Handbook of Dscourse Analysis (pp. 103-136). London: Academic Press.

van Rees, M.A. (1999), We komen er wel uit: Over rationaliteit in Probleemoplossende Discussies (In Dutch). Amsterdam: Boom.

216

References

Veldhuis-Diermanse, A.E. (2002), CSCLearning? Participation, Learning Activities and Knowledge Construction in Computer-Supported Collaborative Learning in Higher Education. PhD dissertation Wageningen University, Wageningen.

von Bertalanffy, L. (1950). The theory of open systems in physics and biology. Science, 111, 23-32.

von Glaserfield, E. (1991). Radical Constructivism in Mathematics Education. Dordrecht, Netherlands: Kluwer Academic Publishers.

von Glasersfeld, E. (1995). Radical Constructivism: A Way of Knowing and Learning. London: Falmer.

Vygotsky, L. (1986). Thought and Language. Cambridge, MA: The MIT Press.Waldeck, J.H., Shepard, C.A., Teitelbaum, J., Farrar, W.J. & Seibold, D.R. (2002). new

directions for functional, symbolic, convergence, structuration, and bona fide group perspectives on group communication. In L.R. Frey (Ed.), New Directions in Group Communication (pp. 3-23). Thousand Oaks, CA: Sage Publications.

Webb, N.M. & Palincsar, A.S. (1996). Group processes in the classroom. In D.C. Berliner & R.C. Calfee (Eds.), Handbook of Educational Psychology (pp. 841-873). New York, NY: Simon & Schuster Macmillan.

Weick, K.E. (1995). Sensemaking in Organizations. Thousand Oaks, CA: Sage Publications.Weick, K.E. (1979). The Social Psychology of Organizing. New York, NY: Random House.Wells, G. & Arauz, R.M. (2006). Dialogue in the classroom. The Journal of the Learning

Sciences, 15(3), 379-428.Wertsch (1991). Voices of the Mind: A Sociocultural Approach to Mediated Action. Cambridge,

MA: Harvard University Press.Wertsch, J.V., del Rio, P. & Alvarez, A. (1995). Sociocultural studies: history, action, and

mediation. In J.V. Wertsch, P. del Rio, & A. Alvarez (Eds), Sociocultural Studies of Mind. Cambridge, UK: Cambridge University Press.

Wenger, E. (1998). Communities of Practice: Learning, Meaning, and Identity. Cambridge, UK: Cambridge University Press.

Winn, W. & Windschitl, M. (2001). Towards an explanatory famework for learning in artificial environments. Cybernetics & Human Knowing, 8(4), 5-23.

Winograd, T. & Flores, F. (1986). Understanding Computers and Cognition: A New Foundation for Design. Norwood, NJ: Ablex publishing.

Wittenbaum, G.M., Hollingshead, A.B., Paulus, P.B., Hirokawa, R.Y., Ancona, D.G., Peterson, R.S., Jehn, K.A. & Yoon, K. (2004). The functional perspective as a lens for understanding groups. Small Group Research, 35(1), 17-43.

Wubbels, T. & Brekelmans, M. (2005). Two decades of research on teacher-student relationships in class. International Journal of Educational Research, 43, 6-24.

Wyndhamm, J. & Säljö, R. (1999). Quantifying time as a discursive practice: Arithmetics, calandars, fingers and group discussion as structuring resources. In J. Bliss, R. Säljö & P. Light (Eds), Learning Sites: Social and Technological Resources for Learning (pp. 80-96). Oxford, UK: Elsevier Science.

Zimmerman, D.H. & West, C. (1975). Sex roles, interruptions and silences in conversation. In B. Thorne & N. Henley (Eds.), Language and Sex: Differences and Dominance (pp. 105-129). Rowley, MA: Newbury House.

217

References

Zuriff, G. (1998). Against metaphysical social constructivism in psychology. Behavior and Philosophy, 26, 5-28.

218

Appendix A: A Scenario for the Threaded Discussion Tool

We developed a scenario that describes how the students work with the envisioned Threaded-discussion tool. We used the data of a study that has been carried out at a secondary school during a Dutch language course. The experiment included three lessons that prepared students for a final examination where they have to write an argumentative text about “healthy school canteens”. The students formed groups of three students to discuss the central claim: ‘schools should make their canteens healthy.” This discussion consists of two parts: 1) the students had to discuss the claim by using a formal argumentation scheme, and 2) the students individually reflect on the discussing by identifying the relevant arguments.

The scenario describes how the Digalo-supported activities was implemented in the Threaded-discussion tool. The data that is displayed stems from the original study.

General user-interface

Figure 1 displays how the User interface of the Threaded discussion tool may look like. A user selects a category in the “category” menu by selecting a category in the drop-down menu. The notation system is displayed in the menu beneath the shared workspace. The notation system consists of the same four labels as in the Digalo: 1) main argument, 2) subargument, 3) counterargument, and 4) rebuttal. We use colours to differentiate between the various labels.

Submit a contribution

In our example, the user wants to comment on contribution 12. First, the user selects that contribution (contribution 12 changes of colour). Next, the user selects the kind of respond that he or she wants to make. In our example, the user wants to react by giving a counterargument (Figure 2). A new window appears on the screen. We use the distinction between a ‘title’ and ‘comment’. The user types in the counterargument in the contribution window (Figure 3). The counterargument is displayed in the shared workspace when the user double clicks on the ‘submit’ button (Figure 4).

219

Appendix A

220

Figure 1: A user interface for the threaded discussion tool.

Figure 2: The user selects a contribution type.

A Scenario for the Threaded Discussion Tool

221

Figure 3: Write a contribution.

Figure 4: Contribution 12.1 has been submitted.

Appendix A

Move a contribution to another category

During the next activity, the user reflects on the digital discussing and identifies the relevant arguments. The user performs this activity individually. The user selects the arguments and places them in a new category. Figure 5 displays a possible solution for this type of action (this User interface has a separate category window). In the group session “moving a contribution” will be carried out by an individual group member.

The user selects contribution 12 (and all the contributions that are connected with this contribution, i.e. in our example contribution 12.1) and moves this contribution to the category “arguments” (Figure 5). The contribution is put in the new category “arguments” when the cursor has selected that category (Figure 6).

There are, of course, various other User interface options for moving a contribution to another category that we need to examine.

222

Figure 5: Move a contribution to a new category.

A Scenario for the Threaded Discussion Tool

223

Figure 6: Submit the contribution to the new category.

Figure 7: The contribution has been moved.

Appendix A

224

Appendix B: A Functional Description of the Graphical Tool

The Graphical tool has several structural features that helps the group to share and organize ideas in a meaningful way. The features can be considered as representation aids that are relevant for the task performance. The representational aids that are characteristics for the Graphical tool are:

– A notation system that elicits specific communicative actions in the shared workspace of the tool.

– The possibility to link related concepts so that the external representation can be organized in a logical order that reflects students’ reasoning.

– The possibility to divide shared workspace into functional spaces that have a specific meaning.

Notation SystemA notation system encourages students to display specific communication action. A notation system consists of a set of labels or contribution cards that represents pre-defined communicative actions. The notation system primes the the users to display

225

Figure 1: Different means to visualize contributions types.

Appendix B

certain communicative actions in the shared workspace. Users choose a label before they start to type a text.

The notation system must be configurable. The exact configuration depends on the pedagogical objectives, the learning task, and the learning activity. For example, a notation system encourages argumentation by providing the students with a system for formal argumentation.

Recognizable contribution types

The different contribution types must be easily recognizable within the shared workspace. There are several ways to mark differences between contribution types (Figure 1):

– the use of different shapes,

– a textual label attached to the contributions, or

– the use of different colours.

The different types of contributions displayed as different kinds of shapes must be visually discriminative. Digalo uses clearly distinguishable shapes. Figure 2 displays the different shapes that are used by the Digalo tool. All the shapes have the same height and width and contain the same text. There is a large variety in the actual amount of text that is displayed by the different shapes.

226

Figure 2: The different Digalo shapes.

A Functional Description of the Graphical tool.

Comment Window

The different ways for representing a lengthy text determines to a large extend the appearance of a diagram. Only a small amount of text can be placed within the borders of a shape. Otherwise a shape takes up a very large area of the shared workspace. Limiting the size of the shape and the use of a note window could overcome this problem. Users write down additional text in a Note window to give a more detailed description of their idea. This additional Note window belongs to a specific shape in the shared workspace. The shape is always visible, while the additional Note window is not directly visible. First of all, the distinction between labels and notes is crucial:

– A label is a generic classification of the different types of contributions or links between two contributions

– A note is an additional description of a specific shape or link in the shared workspace.

There are several options for adding additional text in a Note window. Options that are used by other tools are:

– Move with the mouse cursor over the shape to display the associated comment and double click on the shape to open an edit window (e.g. DREW).

– Right-clicking on shape opens a menu that allows adding a note by opening the note editor. When the user closes the note editor a “note” icon is attached to the shape (see Inspiration).

– By right-clicking on the shape, an edit window appears where the user adds a title in the title box and a lengthy textual comment in the comment box. Only the title is directly visible in the shared workspace (upper window). To read the text in the comment box the users have to select and double click on a shape in the shared workspace, and the edit window (re)appears (see Digalo).

The Inspiration option seems to be the most sensible solution because users are directly aware of the additional comment. A “note label” indicates when is shape has an additional text window. A user can double click on the shape to add a note or right-click on the “note label” icon to open the Note window and to read or edit the text (Figure 3).

227

Appendix B

The use of the Note window to write a lengthy text leads to a less “crowded” diagram. We evaluated this hypothesis in a study where groups were stimulated to use a Note window. The study has been carried out at a secondary school, with 4 th grades students. Nineteen groups participated in the study. They had to carry out an argumentative discussion with the support of the Digalo. The software did not force users to use the Note window. Beforehand students received an instruction that made them aware of this option. Not all groups did use the note window frequently, some groups did not use the option at all. An evaluation of the diagrams indicate that users typed less text for the title when they frequently use the comment window of the Digalo (Table 1).

Connect related concepts

A concept-mapping interface enables users to link related contributions so that users can organize their discussion in a logical order that reflects their reasoning. The links are configurable. The following types can be distinguished:

– association (e.g. concept map),

– causal (e.g. causal map),

– support (e.g. argumentative diagram), and

– opposition (e.g. argumentative diagram).

A Notation editor allows the teacher to determine the semantics of shapes and connectors. The teacher configures seven different kinds of shapes or connectors beforehand.

228

Figure 3: Adding a text in the note window.

A Functional Description of the Graphical tool.

Functional spaces

A diagram becomes very complex in a relatively short time when the users put forward contributions simultaneously. The study that has been discussed in chapter 7 showed that users spend a lot of time organizing their contributions in the shared workspace into an understandable whole. Functional spaces that divide the shared workspace into meaningful areas helps users to organise their diagram. These areas serve as a representational aid that helps users to organize their contributions in the shared workspace.

Meaningful areas will be pre-configured by the teacher in the session designer. The different areas have a distinct meaning, e.g. different solutions . A View window quickly displays these areas. It makes users aware of the different aspects or views of the diagram. Users can also quickly navigate towards the specific area that is associated with a specific view.

229

Group No. of words in title

No. of contributions

Average no. of words in title

Use of comments

1 180 31 6 A lot

2 585 46 13 Very little

3 615 42 15 No

4 580 48 12 Very little

5 376 36 10 Very little

6 508 45 11 Regularly

7 264 44 6 A lot

8 337 36 9 A lot

10 88 15 6 A lot

11 873 73 12 No

12 300 30 10 Regularly

13 351 27 13 No

14 138 17 8 Regularly

15 195 31 6 A lot

16 281 31 9 A lot

Table 1: Average number of words in a title.

Appendix B

230

Figure 4: : Meaningful areas and an overview window.

Appendix C: CoFFEE: A Brief Description and Download Information

The two collaborative tools that have been discussed in this thesis are part of the CoFFEE software system. This software system has been developed by the LEAD project team. At the end of the project a functional model of CoFFEE could be identified that made a distinction between two different users who have their own application: the Discusser for the students and the Controller for the teacher.

Discusser

The Discusser has four types of tools or services (Figure 1): 1) collaborative tools, 2) shared tools, 3) personal tools, and 4) communication tools.

231

Figure 1: A functional model of CoFFEE.

Appendix C

The collaborative tools support particular collaborative learning activities within the context of a group discussion. These tools are discussed in this thesis.

The shared tools enable students to share information that may be relevant for their task performance. With the Repository the teacher makes certain documents available to the students or students share documents within their group. The Positionometer enables students to display their preference or opinion with regard to a certain topic or question. The Presence tool displays personal information about the student.

Personal tools are tools for private use; they cannot be accessed by other group members, that is, the information that students put forward is not shared. With the Note tool students write down personal notes for their own use.

A fourth set of tools can be typified as communication services. These tools allow students or the teacher to communicate with each other. The Quick communication service is equivalent to raising one's finger in a traditional classroom situation. For example, students can ask a question to the teacher without disturbing other groups.

Controller

The Controller enables the teacher to manage the activities that students carry out with CoFFEE. The teacher has some additional services to manage the ongoing sequence of learning activities. These services are part of the Controller application. The teacher can load a session from a previous lesson so that students can continue to work on an assignment that covers several lessons.

The teacher schedules several group activities (steps) during one session, for example the groups start to discuss a problem with the Threaded-discussion tool after which they construct a concept map with the Graphical tool. The teacher can change group composition during a session with the Controller. With this service the teacher assigns a latecomer to a group. The Controller also displays the activities of the groups so that teacher may observe these groups.

How to obtain CoFFEE

CoFFEE is freely available and can be downloaded from the following web site: www.coffee-soft.org

232

Summary

Collaborative learning is a valuable means for learning that can be characterized by active involvement of the learners. Collaborative learning – as it is presented in this thesis – consists of a group discussion between a small number of students. These discussions are directed towards the exploration of a particular subject or the resolution of a problem. Learning from this perspective can be seen as a continuing effort to improve on existing knowledge through an engagement in a discourse that advances mutual understanding (Bereiter, Scardamalia, Cassells & Hewitt, 1997).

Bringing students together does not automatically lead to a productive discussion that promotes learning. Sfard and Kieran (2001), for example, concluded that the merits of collaborative learning could not be taken for granted due to ineffective communication patterns. These communication patterns inhibit the free expression of ideas and the further exploration of these ideas. In this thesis, we argue that these ineffective patterns are closely related to the medium within which they occur. Such a medium for communication provides specific opportunities for expressing ideas and for the coupling of these ideas into a meaningful whole. The medium prescribes the content and sequence of communication in a specific manner (Cushman & Kovacic, 1994).

The objective of the research that is described in this thesis is to develop two collaborative tools to support face-to-face group discussions. The computer tools offer the students an alternative medium for communication. It is worth to notice that the computer-mediated communication is not contrasted with verbal communication. Rather both forms occur simultaneously in a face-to-face setting where students are co-located and communicate directly with each other. The collaborative tools provide the groups with a digital medium that stimulates a constructive dialogue between students. This brings us to the following four research questions:

Q1: What are the criteria for a constructive dialogue in terms of requisite patterns of communication?

Q2: Which communication patterns during a verbal face-to-face discussion prevent students from carrying out a constructive dialogue?

233

Summary

Q3: How do the structural features of the medium relate to the ineffective communication patterns?

Q4: How do the envisioned collaborative tools change the group communication for the better?

Understanding through Design

The four research questions reflect a problem-solving paradigm that considers the design process as a means to advance scientific understanding. One of the most noticeable outcomes of such a process is a concrete artifact – a product developed by humans – that offers a solution for a problem. Another result of the design process – less tangible but of equal importance – is increased understanding, or what Winograd and Flores (1986) referred to as “the interaction between understanding and creation”. Exactly this latter outcome comprises the essence of the Design-based research approach that guides the research. This approach can be typified as a modeling mode of inquiry in which theory, operationalization, and data patterns are treated simultaneously (Poole, McPhee & Canary, 2002).

Hypotheses in the Design-based research approach serve a specific goal. We follow the philosopher Popper (1999) who stated that the solution of a problem might be compared to an expectation, and hence to a hypothesis or a theory. The hypotheses that we formulated offer a solution for a communication problem, they are anchored in a theoretical model, and will be implemented in a collaborative tool. The approach can be further characterized by a regular evaluation of the tool and, if necessary, a refinement of hypotheses so that the tool better fits the context of use.

Rationale for a Design: Systems and Functions.

In this thesis, we argue that the Design-based research approach relates to a distinct social ontology, and hence with a particular explanatory framework. We adopt a Systems perspective that sees reality as consisting of functional entities or systems that are related. The group is viewed as a purposeful system that can be described on various levels of complexity. This opens up explanations in terms of coordination processes and function. These functional descriptions call for a “concrete and detailed account of the mechanisms, which operate to perform a designated function” (Merton, 1967). In our case, we focus on the communication and examine:

234

Summary

– which functions the communication fulfills for the group, and

– how the group organizes individual utterances into a coherent and meaningful whole.

A Constructive Dialogue

Several studies about collaborative learning looked at the group communication (e.g. Gillies, 2004; Barron, 2003; Kneser & Ploetzner, 2001; Hogan, Nastasi & Pressley, 2000; Keefer, Zeitz & Resnick, 2000). These studies make it clear that the quality of interaction has implications for learning (Barron, 2003). In the thesis, we further elaborated on what quality exactly means and present four criteria for a constructive dialogue in terms of requisite communication. First, the communication must be oriented at effective task performance because learning achievement is associated with how well students communicate about the task. Task-related communication leads to cognitive activities often referred to as knowledge elaboration, which, in turn, facilitates the acquisition of that knowledge (Draskovic, Holdrinet, Bulte, Bolhuis and Leeuwe, 2004). Secondly, the group has to maintain acceptable levels of participation so that all its members are able to share their knowledge. Equal participation is a fundamental element of well-performing student groups (Lindblom-Ylänne, Pihlajamäki & Kotkas, 2003). Thirdly, the group members must organize their individual talk into a coherent whole. Research indicated that successful groups carried out more coherent discussions where they linked proposals to the prior conversation (Barron, 2003; Kneser & Ploetzner, 2001). Finally, discussions of successful groups can be characterized by frequent knowledge elaborations (Barron, 2003; Kneser & Ploetzner, 2001; Hogan, Nastasi & Pressley, 2000). These elaborations resemble a coherent sequence in the sense that several speakers make substantive statements that build on or clarify prior statement.

Ineffective Communication Patterns

The teachers that participated in our studies reported that a discussion was sometimes controlled by one or more dominant students. These students talked a lot while the other students hardly got the opportunity to share their ideas with the group. This does not only work out badly for individual students, it also hampers the learning potentials of the group. Bales (2002) considered this form of interpersonal dominance as one of the basic problems of a group that manifests itself by behaviors like repeatedly taking

235

Summary

the turn, interruptions and simultaneous speech. These disruptive behaviors are an intentional violation of the turn-taking rule that rests on the rule that only one group member talks at the same time.

Another consequence of the turn-taking principle is product blocking. Product blocking occurs when individuals need to wait to verbalize an idea because of some procedural limitation (Isakesen, 1998). Group members, who are prohibited to verbalize their ideas on the moment they occur, forget or suppress them so that the process of idea generation stagnates or it is disrupted because of blocking (Nijstad, 2000).

Design Guidelines

The two ineffective communication patterns – interpersonal dominance and product blocking – formed the rationale for the tool design. It is expected that these behaviors do not occur when the group members organize their communicative exchanges according to the principle of “parallel access” instead of turn taking.

With parallel access, all the group members can access a shared workspace simultaneously. They do not have to wait for their turn but they can immediately share an idea with the group. The transition from turn taking to parallel access is a basic property of the collaborative tools. The users of these tools cannot interfere directly when other users express their thought into words. They type an idea in a private window rather than expressing it verbally.

The general guideline of parallel access does not suffice. That guideline describes the process of digital communication only in rudimentary terms. The implementation of parallel access triggered further research that aimed to come up with a tool that is specifically suited to support face-to-face group discussions. This resulted in eight design guidelines:

1. The support closely matches the characteristics of effective task performance.2. Associate a medium with a specific communicative function.3. Ensure joint attention by concentrating the digital communication in one

collaborative tool. 4. Use parallel access as floor-control mechanism for access to the digital discussion.5. Use functional spaces to stimulate overview and coherence.

6. Associate the functional spaces with a relevant macro structure that is related to the problem or the task.

236

Summary

7. Establish global coherence by using connections that enables users to respond to a contribution in the shared workspace.

8. Use a notation system to stimulate elaboration.

Functional design

The eight design guidelines were translated into a number of services. These services describe a set of tangible actions that the user can perform with the collaborative tool. They lay down the external behavior of a tool from the user's point of view (Davis, 1990).

The tool design resulted in two collaborative tools – a Graphical tool and a Threaded-discussion tool – to support particular collaborative learning activities. Both tools are based on the eight design guidelines and offer the users the ability:

– To divide the shared workspace into functional spaces, i.e. categories for the threaded-discussion tool and meaningful areas for the graphical tool.

– To arrange a discussion according to different levels of complexity: categories and threads for the Threaded-discussion tool and meaningful areas and sequences of related text objects for the Graphical tool.

– Graphically link ideas that are related.

– To label contributions according to a pre-defined notation system.

The Graphical tool allows students to make abstract representations of a problem. The tool restricts the communication to a distinct and limited number of classes (Stenning & Oberlander, 1995). The tool has a user interface that represents concepts and their relationships in a two-dimensional workspace. The tool displays small text objects (shapes) that represent a concept or an idea and lines or arrows (links) to represent a relationship between two concepts. The tool has a number of representational aids that help students to represent their ideas in a meaningful and coherent manner.

The Graphical tool seems to be less useful to represent the richness of a problem-solving discussion that addresses various topics. Conversely, the Threaded-discussion tool has a workspace where students can discuss a problem extensively. The tool enables students to share, discuss and organize their ideas in a systematic manner. The tool has a stratified structure to represent different levels of complexity that represent different levels of elaboration.

237

Summary

Parallel Access

The first study that is discussed in the thesis, examines the effects of parallel access. Parallel access was implemented in a Graphical tool. The tool supported an argumentative discussion between a small group of students. Initial findings revealed that the number of contributions expanded quickly. The students spent a lot of time organizing these contributions into a meaningful whole. Some groups came up with comprehensible diagrams that looked organized while other groups failed to achieve this.

An analysis of the diagrams revealed that spatial grouping led to more coherence between contributions. The groups who did not apply the spatial character of the workspace came up with less structured diagrams. This insight served as inspiration for the development of design guidelines 5 and 6. These guidelines aimed to solve the coordination problem that users experience when they access a digital workspace simultaneously.

Design guideline 5 is grounded in the hypothesis that states that students are better able to coordinate their actions when the shared workspace is divided into functional spaces. Design guideline 6 indicates that these functional spaces must have a distinct meaning.

We evaluated the two design guidelines in a study with two conditions. In one condition, the shared workspace was divided into a number of meaningful areas while in the other condition the groups did not have such a representational aid. Results indicate that the use of meaningful areas that relate to the learning task reduced the coordination problem: students spent less time organizing their contributions and they put forward substantial more contributions in the digital workspace.

Task Performance

In the second study, we elaborated on design guideline 7 and 8 that further define effective task performance. The aim of the study is to establish a proper fit between the use of collaborative tool and the communicative demands set by the learning task. It was expected that the representational aids encourage students to elaborate on the knowledge that they share in the digital workspace.

An analyses of the content of the digital discussion revealed that the communication in the digital workspace was purely about the task. Students share new

238

Summary

information to form more precise interpretations. The communication removed the uncertainty that was caused by ignorance or imprecision of a shared interpretation.

An analyses of the interaction sequences showed that several discussion lines existed in parallel and remained active for a long period. These discussion lines can be considered as focused discussions that deal with a specific topic. These discussion lines are situated in different areas of the shared workspace. An analysis of students’ pattern of participation revealed that they constantly switched between discussions lines. Students “jumped” from one focused discussion to another; added a contribution to the discussion and moved forward on to the next topic.

A Diagnostic reasoning discussion

The third study investigates the effects of a collaborative tool on the communication and learning. The students that participated in the study worked together in small groups. They carried out a diagnostic reasoning discussion during which they analyzed a patient's health condition to formulate a care plan. Three aspects of the support were addressed by the study:

1. The instructional strategy that aims to establish a good match between the learning task and the computer support.

2. If the digital part of the communication is really a discussion.

3. How a face-to-face discussion that has a digital and verbal component looks like in practice.

While the collaborative tools stay more or less the same over different contexts, its use will vary. Teachers can use a collaborative tool in many ways that can be more or less appropriate to the learning task. Just introducing the collaborative tools in the classroom is not sufficient. In our case, the introduction of the collaborative tools is accompanied with an instructional strategy that makes the computer support suitable for a specific learning situation. An instructional strategy specifies how the students learn effectively. It consists of a lesson plan that directs student's learning towards the achievement of the learning goals. After each lesson, the instructional strategy was evaluated with the teacher. This led to a number of adaptations that mainly had to do with improving the coherence between the various activities.

The second topic that was addressed in the study had to do with the utility of the collaborative tools. Utility refers to how well the students are able to carry out a

239

Summary

discussion digitally. The students reported that they could express their ideas in writing and were aware of what other group members shared. The students also mentioned that there were frequent interactions within the shared digital workspace.

The third topic had to do with the two media for communication: verbal speech that is paralleled by digital communication. Participation during the verbal part of the discussion varied substantially: in almost every group, there was at least one student who hardy said anything. The communication in the digital workspace revealed a different kind of pattern: all the group members contributed substantially to the digital discussion. There was an even pattern of participation in the digital workspace. The willingness to participate was irrespectively of the behavior during the verbal discussion. Group members who hardly said anything, frequently communicated digitally.

There is a complex interplay between the verbal and computer-mediated part of the communication. The permanence of the contributions in the digital workspace made constant reflection possible. Students frequently reacted on the statements that they read in the digital workspace. This was done digitally but also verbally. Verbal communication helped the group members to align the views with regard to the differences that became visible in the digital workspace.

Changing communication

Group discussions can be characterized by multiparty talk that is episodic in nature (Schwartzman, 1989). The episodic character implies that the utterances of individual group members are only loosely coupled. Coherence, order and meaning between successive statements cannot be taken for granted. Groups have to follow certain rules for coupling to make their individual expressions into a coherent and meaningful whole. A central premise of this thesis is that each medium for communication provides specific rules for coupling. These rules do not only determine what can be expressed and how, it also affects the behavior of the group.

In our case, we compared parallel access as floor control mechanism for digital communication with turn taking that can be associated with verbal speech. Findings indicate that parallel access leads to participation that is more or less equal: dominant group members have fewer opportunities to influence the digital part of the discussion so that the other group members are better able to share their thoughts with the group. Parallel access changed the discussion in several ways: students worked at their own

240

Summary

pace, they could freely choose a topic of immediate interest, and they expressed their thoughts without being interrupted.

Turn taking as floor control mechanism is closely associated with adjacency pair or “nextness” as unit for sequence construction (Schegloff, 2007). Adjacency pair means that a statement receives its meaning in relation to a statement that directly precedes in time. Adjacency pair as a rule for coherence works less well for parallel access because there is no clear time frame. Consequently, the shared digital workspace became crowded with contributions that lacked a clear relationship. To solve this coordination problem we developed some additional guidelines:

– a shared digital workspace that is divided into functional spaces,

– functional spaces that are based on a macro structure that represents relevant aspects of the problem or the task, and

– the possibility to link a new contribution to a contribution that is already present in the shared workspace.

The three guidelines enable users to organize their digital discussion based on a global model of coherence. Global coherence refers to a macro structure that is based on topics or themes in a discourse (van Dijk, 1985). Such a structure was implemented in the collaborative tools as a representational aid, i.e. categories for the Threaded- discussion tool or meaningful areas for the Graphical tool. These macro structures changed the digital part of the discussion. Several discussion lines occurred in parallel that remained active during the whole time. These discussion lines addressed specific topics that had a direct relation with the macro structure, i.e. with the problem or task. It seems that the macro structure broadens up the discussion, i.e. they gave the students more freedom to follow their own lines of thinking and to share their thoughts with the group.

The collaborative tools keep the discussion open. As a result, the students become more aware of their differences in thinking. The tension between the individual and the group, between unhampered thinking and conformity to the group becomes more explicit. There is also a current in the opposite direction towards conformity that is visible in the verbal part of the communication. Differences in thinking are discussed verbally. The group members align their cognitive models verbally.

241

Summary

The medium for communication does not only determine how thoughts are expressed and joined together into a meaningful story. The medium also stimulates certain behaviors while it suppresses other behaviors. This becomes clearly visible in a face-to-face setting where a digital medium for communication is places next to human speech. The research that is described in this thesis shows that the digital medium changes the dynamics of a group in such a way that students are better able to collaborate, to share knowledge and to elaborate on the knowledge that is shared.

242

Samenvatting

Samenwerkend leren is een waardevolle manier van leren die zich kenmerkt door actieve betrokkenheid van de studenten. Samenwerkend leren – zoals het gepresenteerd wordt in dit proefschrift – bestaat uit een discussie tussen een klein aantal studenten. Tijdens zo'n discussie verkennen de studenten een bepaald onderwerp of lossen zij een probleem op. Leren wordt vanuit dat oogpunt opgevat als het verdiepen van bestaande kennis door deelname aan een dialoog die zich richt op het ontwikkelen van wederzijds begrip (Bereiter, Scardamalia, Cassells & Hewitt, 1997).

Het bijeenbrengen van studenten betekent echter niet automatisch dat er een productieve discussie ontstaat die het leren bevordert. Sfard en Kieran (2001) concluderen dat de voordelen van samenwerkend leren niet als vanzelfsprekend zijn te beschouwen vanwege ineffectieve communicatie patronen. Deze communicatiepatronen verhinderen dat studenten hun ideeën vrij kunnen delen. In dit proefschrift betogen wij dat deze ineffectieve patronen nauw verbonden zijn aan het medium waarbinnen zij optreden. Zo'n communicatiemedium biedt specifieke mogelijkheden voor het uiten van ideeën en het verbinden van deze ideeën tot een betekenisvol geheel. Het medium dicteert als het ware hoe de inhoud en de volgorde van communicatie eruit ziet (Cushman & Kovacic, 1994).

Het onderzoek dat in het proefschrift wordt beschreven heeft tot doel om een tweetal samenwerkingstools te ontwikkelen voor het ondersteunen van face-to-face groepsdiscussies. Deze computertools bieden de studenten een alternatief medium voor communicatie. De computer-gemedieerde communicatie staat dus niet tegenover mondelinge communicatie. Beide vormen vinden gelijktijdig plaats in een face-to-face omgeving waar studenten bij elkaar zitten en direct met elkaar communiceren. De samenwerkingstools dienen een constructieve discussie te bevorderen die een meerwaarde biedt ten opzichte van de “traditionele” mondelinge discussie. Dit uitgangspunt brengt ons tot de volgende onderzoeksvragen:

Q1: Wat zijn de criteria voor een constructieve dialoog in termen van noodzakelijke communicatie patronen?

Q2: Welke communicatie patronen tijdens een mondelinge discussie verhinderen het voeren van een constructieve dialoog?

243

Samenvatting

Q3: Hoe relateren de onderliggende structurele eigenschappen van het medium zich tot de ineffectieve communicatie patronen?

Q4: Hoe veranderen de te ontwerpen samenwerkingstools de communicatie zodanig dat de groep beter in staat is om een constructieve dialoog te voeren?

Begrijpen door te Ontwerpen

De vier onderzoeksvragen weerspiegelen een probleemoplossend paradigma dat het ontwerpproces ziet als een middel voor het ontwikkelen van wetenschappelijke kennis. Het ontwerpproces leidt tot een artefact – een door mensen gemaakt product – dat een oplossing biedt voor een probleem. Zo'n proces kan ook nieuwe kennis opleveren.

Een van de meest zichtbare uitkomsten van het ontwerpproces is een “tastbaar” voorwerp, in ons geval een samenwerkingstool voor het ondersteunen van groepsdiscussies. Een ander resultaat – minder zichtbaar maar van even groot belang – is kennis, of wat Winograd en Flores (1986) omschrijven als “de interactie tussen begrijpen en creëren”. Juist dit resultaat van toenemende kennis omvat de essentie van de Ontwerpgebaseerde onderzoeksbenadering die tijdens het onderzoek werd gevolgd. Deze onderzoeksbenadering kan worden getypeerd als een modelmatige wijze van onderzoek waarbij theorie, operationalisaties en datapatronen gelijktijdig aan bod komen (Poole, McPhee & Canary, 2002).

Hypotheses vervullen een specifiek doel in de Ontwerpgebaseerde onderzoeksbenadering. Wij volgen de filosoof Karl Popper (1999) die stelt dat de oplossing van een probleem vergeleken kan worden met een verwachting en derhalve met een hypothese of een theorie. De hypotheses die wij formuleren bieden een oplossing voor een praktisch probleem, zij zijn verankerd in een theoretisch model en worden uitgedrukt in het ontwerp van de tools. Deze benadering benadrukt dat het ontwerp regelmatig getoetst wordt en dat de hypotheses, indien nodig, worden aangepast zodat het ontwerp beter aansluit bij het gebruik.

Grondgedachte van Ontwerpen: Systemen en Functies

In het proefschrift relateren wij de Ontwerpgebaseerde onderzoeksbenadering aan een bepaalde sociale ontologie en een daarbij behorend verklarend raamwerk. Wij gaan uit van het Systeemdenken waarbij de werkelijkheid bestaat uit functionele entiteiten of systemen die aan elkaar gerelateerd zijn. De groep wordt opgevat als een doelbewust

244

Samenvatting

systeem die op verschillende niveaus van complexiteit kan worden beschreven. Dit maakt het mogelijk om de groep te beschrijven in termen van coördinatieprocessen en functies. Dit resulteert in functionele beschrijvingen die zich richten op de mechanismen voor het uitvoeren van een bepaalde functie (Merton, 1967). In ons geval richten wij ons op de communicatie en gaan na:

– welke functies de communicatie vervult voor de groep, en

– hoe de groep individuele communicatieve uitingen organiseert tot een coherent en betekenisvol geheel.

Een constructieve dialoog

Diverse studies naar samenwerkend leren richtten zich op de groepscommunicatie (zie bijvoorbeeld Gillies, 2004; Barron, 2003; Kneser & Ploetzner, 2001; Hogan, Nastasi & Pressley, 2000; Keefer, Zeitz & Resnick, 2000). Deze studies maken duidelijk dat de kwaliteit van de interacties gevolgen heeft voor het leren (Barron, 2003). In het proefschrift gaan we verder in op de betekenis van het woord kwaliteit en komen wij met vier criteria voor een constructieve dialoog. Ten eerste, dient de communicatie gericht te zijn op het effectief uitvoeren van een leertaak. Dit omdat de leerprestaties nauw verbonden zijn aan hoe goed studenten inhoudelijk praten over de taak. De Taakgerelateerde communicatie leidt tot cognitieve activiteiten die worden getypeerd als kennis elaboraties. Kennis elaboraties leiden vervolgens weer tot het verwerven van die kennis (Draskovic, Holdrinet, Bulte, Bolhuis and Leeuwe, 2004). Ten tweede, dient een groep een acceptabel niveau van participatie te handhaven zodat alle leden in staat zijn om hun kennis te delen met de groep. Volgens Lindblom-Ylänne, Pihlajamäki & Kotkas (2003) is gelijkwaardige participatie een fundamentele eigenschap van goed presterende groepen. Ten derde, moeten de groepsleden hun individuele communicatieve uitingen organiseren to een coherent geheel. Onderzoek liet zien dat de discussies van succesvolle groepen meer samenhang vertoonden. Deze groepen relateerden voorstellen aan de direct voorafgaande discussie (Barron, 2003; Kneser & Ploetzner, 2001). Als laatste, vertonen discussies van succesvolle groepen regelmatige kennis elaboraties (Barron, 2003; Kneser & Ploetzner, 2001; Hogan, Nastasi & Pressley, 2000). Kennis elaboraties bestaan uit een samenhangende reeks van inhoudelijke bijdragen die voortbouwen op wat er eerder gezegd is. Essentieel daarbij is dat meerdere sprekers substantiële bijdragen leveren aan de discussie.

245

Samenvatting

Ineffectieve communicatie patronen

De docenten die deelnamen aan de studies gaven aan dat een discussie soms gecontroleerd werd door één of meerdere dominante studenten. Deze studenten praatten erg veel terwijl andere studenten nauwelijks de gelegenheid kregen om hun ideeën te delen met de groep. Dit werkte niet alleen negatief uit voor de individuele studenten, het belemmerde ook het leerpotentieel van de groep. Bales (2002) beschouwde deze vorm van interpersoonlijke dominantie als een van de fundamentele problemen van een groep wat zich manifesteert in gedrag zoals herhaaldelijk voor de beurt praten, interrupties en door elkaar heen praten. Deze verstoringen zijn een bewuste schending van het “bij-beurten” praten regel dat gebaseerd is op de regel dat slechts een persoon op het zelfde moment spreekt.

Een andere consequentie van het bij-beurten principe is “product blokkering”. Product blokkering treedt op wanneer iemand moet wachten om iets te zeggen vanwege een procedurele beperking (Isakesen, 1998). Bij product blokkering vergeten of onderdrukken personen hun gedachten wanneer zij deze niet gelijk onder woorden kunnen brengen op het moment dat zij zich voordoen. Het proces van het delen van kennis stagneert dan of wordt verstoord (Nijstad, 2000).

Ontwerprichtlijnen

De twee ineffectieve communicatiepatronen – interpersoonlijke dominantie en product blokkering – vormden het uitgangspunt van het ontwerpen. De verwachting was dat deze gedragingen zich niet zouden voordoen wanneer de groep haar communicatie zou organiseren volgens het principe van “gelijktijdige toegang”.

Gelijktijdige toegang betekent dat alle groepsleden tegelijkertijd toegang hebben tot een gedeelde werkruimte. Zij hoeven niet op hun beurt te wachten maar kunnen een idee onmiddellijk delen met de groep. De overgang van “bij beurten” naar “gelijktijdige toegang” is een van de fundamentele eigenschappen van de digitale werkruimte van de samenwerkingstools. De gebruikers van deze tools kunnen elkaar niet direct storen wanneer zij een idee onder woorden brengen. De ideeën worden namelijk in een afgezonderd scherm ingetypt in plaats van mondeling uitgedrukt.

Eén algemene ontwerprichtlijn is onvoldoende voor een goed ontwerp. De richtlijn “gelijktijdige toegang” beschrijft het proces van digitale communicatie slechts in elementaire termen. De implementatie van “gelijktijdige toegang” gaf aanleiding tot

246

Samenvatting

verder onderzoek dat tot doel had om een tool te ontwikkelen die precies geschikt was voor het ondersteunen van face-to-face groepsdiscussies. Dit resulteerde uiteindelijk in acht ontwerprichtlijnen:

1. Relateer een medium met een specifieke communicatieve functie.

2. De ondersteuning dient overeen te stemmen met kenmerken die horen bij het effectief uitvoeren van de leertaak.

3. Garandeer gedeelde aandacht door de digitale communicatie te laten plaats vinden in één samenwerkingstool.

4. Gebruik “gelijktijdige toegang” als mechanisme voor toegang tot de digitale discussie.

5. Verdeel de digitale werkruimte in functionele ruimtes om overzicht en samenhang te bevorderen.

6. Verbind de functionele ruimtes met relevante macro structuur die wijst naar het probleem of de taak.

7. Creëer globale coherentie door de mogelijkheid om nieuwe bijdrage te relateren aan een bijdrage in de gedeelde werkruimte.

8. Gebruik een notatie systeem om elaboraties te bevorderen.

Functioneel Ontwerp

De acht ontwerp richtlijnen werden vertaald in een aantal “services”. Deze services beschrijven de concrete handelingen die de gebruiker kan uitvoeren met de tool. Zij leggen het externe gedrag vast vanuit het oogpunt van de gebruiker (Davis, 1990).

Het ontwerpproces resulteerde in twee samenwerkingstools – een Grafische tool en een Threaded-discussion tool – voor het ondersteunen van specifieke leeractiviteiten. Beide tools zijn gebaseerd op de acht ontwerprichtlijnen en bieden de gebruikers de mogelijkheid om:

– De gedeelde digitale werkruimte te verdelen in functionele ruimtes: categorieën voor de Threaded-discussion tool en betekenisvolle gebieden voor de Grafische tool.

– De discussie te ordenen volgens verschillende niveaus van complexiteit: categorieën en een reeks van samenhangende bijdragen (thread) voor de Threaded-discussion tool én betekenisvolle gebieden en samenhangende tekstobjecten voor de Grafische tool.

247

Samenvatting

– Ideeën die bij elkaar horen grafisch met elkaar te verbinden.

– Bijdragen te labelen volgens een van te voren bepaald notatiesysteem.

De Grafische tool stelt studenten in staat om abstracte representaties te maken van een probleem. De tool begrenst de communicatie tot een beperkt aantal afzonderlijke categorieën (Stenning & Oberlander, 1995). De Grafische tool heeft een gebruikers­interface dat concepten en de relatie tussen de concepten weergeeft in een twee-dimensionale ruimte. Deze ruimte bevat tekstobjecten (vormen) voor het weergeven van ideeën en lijnen of pijlen (links) die de relatie tussen twee ideeën aangeeft. Daarnaast heeft de tool een aantal weergave-hulpmiddelen (representational aids) voor het duidelijk weergeven van een digitale discussie.

De Grafische tool lijkt minder geschikt voor het weergeven van een discussie die diverse onderwerpen behandelt. De Threaded-discussion tool daarentegen voorziet de groep van een digitale werkruimte waar zij een probleem uitgebreid kunnen behandelen. Met behulp van deze tool delen de studenten ideeën, reageren zij op elkaars ideeën en organiseren zij hun bijdragen op een systematische manier. De Threaded-discussion tool heeft een gelaagde structuur die de studenten in staat stelt hun discussie te organiseren volgens verschillende niveaus van abstractie.

Gelijktijdige Toegang

De eerste studie die besproken wordt in het proefschrift beschrijft het effect van gelijktijdige toegang. Gelijktijdige toegang werd in een Grafische tool geïmplementeerd. Deze tool ondersteunde een argumentatieve discussie tussen een beperkt aantal studenten. De eerste bevindingen lieten zien dat het aantal bijdragen in de digitale werkruimte snel toenam. De studenten hadden veel tijd nodig om die bijdragen te organiseren tot een betekenisvol geheel. Sommige groepen kwamen met een duidelijk diagram dat gestructureerd oogde. Bij andere groepen lukte dit niet.

Een analyse van de diagrammen liet zien dat het ruimtelijk groeperen van bijdragen leidde tot meer samenhang. De groepen die het twee-dimensionale karakter van de werkruimte niet gebruikten tijdens het ordenen hadden meer ongestructureerde diagrammen. Dit inzicht diende als inspiratie voor het ontwikkelen van ontwerp richtlijnen 5 en 6. Deze twee richtlijnen hadden tot doel om het coördinatieprobleem op te lossen dat veroorzaakt wordt door “gelijktijdige toegang”.

248

Samenvatting

Aan ontwerprichtlijn 5 ligt de hypothese ten grondslag dat studenten hun activiteiten beter zouden coördineren wanneer de gedeelde digitale werkruimte van te voren wordt opgedeeld in functionele ruimtes. Ontwerprichtlijn 6 geeft vervolgens aan dat deze ruimtes een specifieke betekenis moeten hebben.

De twee ontwerprichtlijnen werden geëvalueerd in een studie waarbij de groepen verdeeld werden over twee condities. In de ene conditie werd de digitale werkruimte opgedeeld in een aantal betekenisvolle gebieden. De groepen in de andere conditie konden geen gebruik maken van zo'n weergave-hulpmiddel. De resultaten laten zien dat het gebruik van functionele ruimtes leidde tot vermindering van het coördinatie probleem: de tijd die groepen nodig hadden voor het organiseren van hun bijdragen nam significant af. Er werden ook substantieel meer bijdragen geplaatst in de digitale werkruimte.

Taakuitvoering

De tweede studie gaat nader in op ontwerprichtlijnen 7 en 8 die effectieve taakuitvoering verder preciseren. Het doel van deze studie is het realiseren van een juiste afstemming tussen het gebruik van de samenwerkingstool én de eisen voor communicatie zoals deze door de taak worden bepaald. De verwachting was dat de weergave-hulpmiddelen kennis elaboratie zou stimuleren.

Een inhoudelijke analyse van de digitale discussie liet zien dat de communicatie in de digitale werkruimte zich richtte op de taak. Het ging daarbij met name om het delen van nieuwe informatie ter verduidelijking. De communicatie had tot doel onbegrip te verminderen die veroorzaakt werd door onwetendheid of een onnauwkeurige uitleg.

Een analyse van de interactie sequenties liet zien dat er meerdere discussie lijnen gelijktijdig actief waren. Deze discussie lijnen zijn gerichte discussies die een bepaald onderwerp behandelen. Deze discussie lijnen zijn te situeren binnen de verschillende functionele gebieden van de digitale werkruimte. De analyse laat zien dat studenten constant veranderen van discussie lijn. Zij “springen” van de ene gerichte discussie naar de andere, plaatsen een bijdrage en gaan dan verder naar een ander onderwerp.

Diagnostische Discussie

De derde studie onderzoekt het effect van de samenwerkingstools op de communicatie, de samenwerking en het leren. De studenten uit deze studie werkten samen in kleine groepjes. Zij namen deel aan een diagnostische discussie waarbij de studenten de

249

Samenvatting

gezondheidstoestand van een patiënt analyseerden om vervolgens een behandelingsplan op te stellen. Drie aspecten van de ondersteuning werden behandeld in de studie:

1. De instructie strategie die zorgt voor een goede afstemming tussen de leertaak en de computer ondersteuning.

2. Of het digitale gedeelte van discussie voldoet aan de minimale eisen voor interactie.

3. Hoe een face-to-face discussie met een digitale en mondelinge component er in de praktijk uitziet.

Terwijl de samenwerkingstools min of meer hetzelfde blijven zal het gebruik van situatie tot situatie variëren. Docenten kunnen de samenwerkingstool op verschillende wijzen inzetten waarbij sommige manieren van gebruik betere passen bij de leertaak dan anderen. De introductie van de tools gaat dan ook gepaard met een instructie strategie die aangeeft hoe studenten de tools effectief kunnen inzetten. Zo'n strategie bestaat uit een lesplan die de leeractiviteiten stuurt in de richting van de leerdoelen. Er werd een speciale instructie strategie ontwikkeld die na elke les werd geëvalueerd met de docent. Dit leidde tot een aantal aanpassingen die zorgden voor een betere samenhang tussen de activiteiten.

Het tweede vraagstuk dat in de studie werd behandeld, had te maken met de bruikbaarheid van de samenwerkingstools. Bruikbaarheid verwijst naar hoe goed de studenten in staat zijn om digitaal met elkaar te discussiëren. De studenten gaven aan dat zij hun ideeën goed naar voren konden brengen in de digitale werkruimte. De studenten waren zich bewust van wat andere groepsleden naar voren brachten. Ook vonden er regelmatig interacties plaats in de digitale werkruimte.

Het derde onderwerp richtte zich op de twee vormen van communicatie. De resultaten lieten zien dat de participatie tijdens de mondelinge discussie uiteenliep. Voor bijna alle groepen gold dat ten minste één student bijna niets zei. De communicatie in de digitale werkruimte had een heel ander patroon. Alle studenten leverden een substantiële bijdrage aan het digitale deel van de discussie. Er was sprake van een evenwichtige participatie. Deze bereidheid tot deelname stond los van het gedrag tijdens de mondelinge discussie. Studenten die weinig mondeling communiceerden, communiceerden regelmatig digitaal.

Er ontstond een complexe wisselwerking tussen het mondelinge en digitale deel van de discussie. Omdat de bijdragen in de digitale werkruimte zichtbaar bleven, was er

250

Samenvatting

sprake van constante reflectie op het digitale deel van de discussie. Studenten reageerden regelmatig op bijdragen die zij lazen in de digitale werkruimte. Dit gebeurde via de computer maar ook mondeling. De studenten discussieerden mondeling over uiteenlopende denkbeelden die zichtbaar werden in de digitale werkruimte.

Verandering van de Communicatie

Groepsdiscussies zijn een gesprek tussen meerdere personen dat episodisch van aard is (Schwartzman, 1989). Het episodische karakter impliceert dat individuele uitspraken slechts los gekoppeld zijn. Samenhang, ordening en betekenis tussen opeenvolgende uitspraken zijn niet vanzelfsprekend. Voor het aanbrengen van samenhangen en betekenis dienen de deelnemers bepaalde regels te volgen. Een vooronderstelling in dit proefschrift is dat elk communicatie medium specifieke regels voor “het koppelen” heeft. Deze regels bepalen niet alleen wat er wordt uitgedrukt en hoe, zij beïnvloeden ook het gedrag van de groep.

In ons geval keken wij naar hoe de toegang tot de discussie geregeld wordt waarbij een vergelijking werd gemaakt tussen “gelijktijdige toegang” en “om de beurt praten”. De eerstgenoemde regel is kenmerkend voor het digitale deel van de discussie terwijl het tweede principe aangeeft hoe dit bij mondelinge communicatie wordt geregeld. Resultaten laten zien dat “gelijktijdige toegang” tot een evenredige participatie leidt. Dominante groepsleden hebben minder mogelijkheden om het digitale deel van de digitale discussie naar hun hand te zetten. De andere groepsleden kunnen daardoor hun kennis beter delen met de groep. Gelijktijdige toegang veranderde de discussie op meerdere manieren: de studenten werkten in hun eigen tempo, zij konden een onderwerp vrij kiezen en hun gedachten direct verwoorden zonder gestoord te worden door anderen.

Om de beurt praten als regel voor de toegang tot een discussie is nauw verbonden met aangrenzende paren (adjacency pair) of nabijheid als eenheid voor het construeren van sequenties (Schegloff, 2007). Nabijheid houdt in dat een uitdrukking haar betekenis ontleent aan wat er direct voorafgaand wordt gezegd. Nabijheid als regel voor het creëren van samenhang werkt minder goed voor het principe van “gelijktijdige toegang”. Er ontbreekt een duidelijk tijdskader. Dit brengt het gevaar met zich mee dat de bijdragen in de digitale werkruimte weinig samenhang vertonen. Om dit coördinatieprobleem op te lossen, werden een aantal nieuwe ontwerp richtlijnen ontwikkeld:

251

Samenvatting

– een gedeelde digitale werkruimte die opgedeeld is in functionele ruimtes.

– Het gebruik van functionele ruimtes die gebaseerd zijn op een macro structuur die belangrijke aspecten van het probleem of de taak weergeeft, en

– de mogelijkheid om nieuwe bijdragen direct te verbinden met een bijdrage in de gedeelde werkruimte.

De drie ontwerprichtlijnen stelden gebruikers in staat om hun digitale discussie te organiseren volgens een globale structuur van coherentie. Globale coherentie verwijst naar een macro structuur die gebaseerd is op onderwerpen of thema's (van Dijk, 1985). Zo'n structuur werd in de samenwerkingstools geïmplementeerd als een weergave- hulpmiddel: categorieën voor de Theraded-discussion tool en betekenisvolle gebieden voor de Grafische tool. Deze macro structuren veranderden het digitale deel van de discussie. Meerdere discussielijnen ontstonden parallel aan elkaar en bleven gedurende de gehele tijd actief. Die discussielijnen behandelden een specifiek onderwerp dat een directe relatie had met de macro structuur. Deze structuur verbreedde de discussie: het gaf de studenten meer vrijheid om hun eigen gedachtelijn naar voren te brengen.

De samenwerkingstools houden de discussie open. Dat had tot gevolg dat de studenten zich beter bewust werden van hun verschillen in denken. De spanning tussen het individu en de groep, tussen onafhankelijk denken en aanpassen aan de groep werd meer expliciet. Door de digitale discussie werd de groep bewust van de verschillen in denken. Dat had een tegenstroom tot gevolg die zichtbaar werd in het mondelinge deel van de discussie. De verschillen in denken werden mondeling besproken. Zij weerspiegelden de intentie van de groepen om te komen tot een gemeenschappelijk inzicht.

Het medium voor communicatie bepaalt niet alleen hoe gedachten onder woorden wordt gebracht en vervolgens worden samengesmeed tot een samenhangend verhaal. Het medium stimuleert ook bepaald gedrag terwijl ander gedrag onderdrukt wordt. Dit wordt duidelijk zichtbaar wanneer het digitale medium naast mondelinge communicatie wordt geplaatst. Het onderzoek in dit proefschrift laat zien dat gebruik van een digitaal medium de groepsdynamiek kan veranderen zodat studenten beter samenwerken, eenvoudig kennis delen en die gedeelde kennis verder uitdiepen.

252

Curriculum Vitae

Wouter van Diggelen studied Electrotechnical engineering at the Hogere Technische School (HTS) at 's Hertogenbosch. He also graduated as MSc in Work- and Organizational psychology at the Tilburg University . During his master he specialized in Organization psychology and carried out research into telework.

Wouter's multidisciplinary background is also visible in his work. His professional interest as a psychologist focuses on the performance of individuals and groups within an educational or a professional context. He is especially interested in the use of new technologies to improve performance.

Wouter has seven years of experience as a researcher and project manager for European Research and Development projects. He took the initiative for the LEAD project: an European funded project that included nine universities, research institutes, and small businesses. He was responsible for the management of the project and the coordination of the technical design activities. Wouter made the conceptual and functional design of collaborative tools that are part of CoFFEE, the award winning software that was delivered by the project.

As a researcher, he carried out applied research in educational institutes and governmental organizations. He did several studies into user experience, collaborated with teachers to implement and evaluate learning with the support of ICT, and examined the role of ICT to support the policy making process. Wouter has published articles about his research in scientific journals and conference proceedings. Furthermore, he coordinated a number of workshops at scientific conferences.

As a group facilitator, he has worked with multidisciplinary design teams in various organizational settings. In that role, he applied various software applications to support team meetings.

As a lecturer, Wouter taught courses in Information Science at the Faculty of Technology, Policy and Management of the Delft University of Technology. He also gave guest lectures at the Tilburg University and Utrecht University about digital collaboration.

253