Lecture 2 Interviewing in Qualitative Semi-Structured ...ikoskine/nova/02interviews.pdf ·...
7
1 Lecture 2 Semi-Structured Interviews Ilpo Koskinen Prof., Dr.Soc.Sci. (sociology) University of Art and Design Helsinki Nova: Qualitative methods for Business Studies Interviewing in Qualitative Research 1. Qualitative Interviewing in Business Studies ? methods in business studies became empirical in the sixties ? at first, methods came from quantitative disciplines (esp. accounting), and from case research tradition ? in the 1980s, these disciplines became "scientific": heavily quantitative ? ? still, during that decade, there was a qualitative undercurrent that became increasingly more important ? in the 1990s, qualitative methods became normal; in some disciplines, they even dominate the methods scene ? "qualitative research" in business studies usually means interviewing, esp. semi-structured interviewing ? today, it'd be important to widen the methods palette from interviewing to other methods such as observation and videotaping ? still, interviewing is a good instrument as long as you - know your method - don't use interviewing with topics better suited for other methods - proceed economically 2. Concepts ? forms of interviewing: - structured - semi-structured (so called "theme interviewing") - non-structured ? this lecture focuses on semi- structured interviewing (unless explained otherwise)
Transcript of Lecture 2 Interviewing in Qualitative Semi-Structured ...ikoskine/nova/02interviews.pdf ·...
1
Lecture 2Semi-Structured Interviews
Ilpo KoskinenProf., Dr.Soc.Sci. (sociology)
University of Art and Design Helsinki
Nova: Qualitative methods for Business Studies
Interviewing in QualitativeResearch
1. Qualitative Interviewing inBusiness Studies
? methods in business studies
became empirical in the sixties? at first, methods came from
quantitative disciplines (esp.accounting), and from caseresearch tradition
? in the 1980s, these disciplinesbecame "scientific": heavilyquantitative
?
? still, during that decade, there wasa qualitative undercurrent thatbecame increasingly moreimportant
? in the 1990s, qualitative methodsbecame normal; in somedisciplines, they even dominatethe methods scene
? "qualitative research" in businessstudies usually meansinterviewing, esp. semi-structuredinterviewing
? today, it'd be important to widenthe methods palette frominterviewing to other methodssuch as observation andvideotaping
? still, interviewing is a goodinstrument as long as you
- know your method
- don't use interviewing with topicsbetter suited for other methods
- proceed economically
2. Concepts? forms of interviewing:
- structured
- semi-structured (so called "themeinterviewing")
- non-structured? this lecture focuses on semi-
structured interviewing (unlessexplained otherwise)
2
? this is the most used qualitativeresearch method
? if used smartly, it is efficient
- you can direct it according to yourneeds
- still, it is open
3. Interviewing: When?
Always remember that there arealternative methods
? analyzing documents? historical research? analysis of interaction based on
tape recordings and video? visual data? even questionnaires
Before you decide to interview, thinkabout these alternatives, and justifyyour choice of interviewing!
Interviews are good, when? you're studying meanings (how
people understand, see,experience something)
? you want to produce quicklydescriptions of some company
- it is difficult to explain things withqualitative data
? interviewing may take yousurprisingly long back in history
? often, it is the only practicalmethod available
- time and budget constraints,ethics, etc.
Interviews should not be used, when? you're primarily interested in
quantitative information? in many types of contents
- for example, interaction? it is susceptible to biases:
embellishments, forgetting, evenlying
- esp. conflict situations!? if you are studying complex
processes, then documents anddescriptions written when thingstook place are a better source ofinformation
?
4. Types of Research ProcessesBased on Interviewing
? a typical research process based
on theme interviewing
- one interview guide used inevery interview
? inductive process possible, andeven recommended, though oftencautiously
- can you introduce new questionsduring the research process?
3
? choosing interviewees
- depends on the purposes of thestudy
- often sample is purposeful
- sampling can be formalized, esp.if the number of people studied islarge
- research economy is crucial!
? sampling methods
- statistical sampling
- snowball sampling
- theoretical sampling
5. Questions
Remember: research questions arenot questions done to interviewees!
The Research Guide
Open of closed questions?
The Order of Questions? warming up questions
- questions close to the intervieweefirst (his/her work...)
- sensitive questions at the end
- simple questions first, abstractquestions at the end
- time logic in process studies? closing questions? at the very end check whether
something else should have beenasked!
Order Matters!? crucial to the flow of the interview? good order improves data and the
validity of the study
- a good organization keeps theinterviewee on leash, s/he doesnot rush ahead too quickly
? it also minimizes the risk that theinterviewer gets confused
- helps to organize the interview
- the interviewer remembers whathas been talked about and what'sstill left to be done
Don't put in Too Many Questions!? a large amount of questions
- confuses the interviewer
- the risk of forgetting somethingincreases
- the conceptual logic of theinterview guide may be lost
- causes time shortage at the end,where more time would typicallybe needed
- leaves little time for elaboratinganswers
4
? thus, make only a few mainquestions
- perhaps 5-8 for one hour
- you can have sub-questionsunderneath the main questionsaid memory rather than are readto the interviewee
6. Interviewing
As such, interviewing is very simple? the interviewer asks, questions
are followed by the interviewee'sanswer
? sometimes answers are followedby an additional question basedon the answer
? you can plan these ahead to someextent; often they arespontaneous
Questions may become increasinglyspecific during the interview
Interviewing techniques? writing during interview? taping
Interviewing situations vary? one interviewer, one interviewee
the simplest case? often a group: then you ought to
think about group dynamics?
7. Transcribing
? to make data useful, it must be
transcribed? transcription is time-consuming
and fairly expensive? how accurately? - Enough but not
too much detail!?
8. Inference
Data as such is never enough for aresearcher
? at the end, it is the quality of theinference(s) drawn from data thatmatters
? two main cases must be keptseparate
A theoretically-driven inference? A good theory is simple, covers
the object well, and provides adynamic for how the mainconcepts are related
5
In an inductive process? you gather data, reformulate your
questions as you go further, andstop the process when you do notfind cases that do not fit yourevolving theory
? the result is a model thatdescribes your data accurately
? finally, this result ought to becontrasted with existing theories
- theoretical work does not guidethe process
Some techniques for data analysis? the scale:
- interpretation with no particularmethod apart from the researcher'sunderstanding
- matrixes, quasi-quantifying? data reduction essential? programs may aid
- NUD•IST and Atlas/ti are the onlycompetitive alternatives
9. Controlling Sources of Errors andReliability
In every study, there are errors. Toproduce reliable results, they oughtto be controlled
Systematic errors bias resultsseriously
? conceptual and theoretical errors? methodic
- badly processed data
- wrong method
- excessive reactivity
- bad inference (negative cases notanalyzed)
Non-systematic sources of error biasresults less seriously
? ill-fated questions, inaccuratetranscripts, insufficientlyelaborated answers, etc.
? these problems lead toinaccuracies rather than threatenthe validity of the whole study!
Expanding the Technique
6
• Or course, there are several variationsto this technique– in marketing in particular, it is customary to
bring in various “props” to interviews:• pictures, things, appliances• interviewing with photographs may be a good
way to keep reactivity at bay
– “Critical incident” technique in moldingquestions:
• Questions focus on critical junctures ofexperience (or whatever), with the aim ofidentifying themes
Interviewees vary: elite interviewing? the choice of interviewees among
elites:
- sampling based on access toinformation, judgment rather thanon sheer numbers
? getting an interview from elitesand their utility expectations
? cognition? remember your homework!? tandem interviewing? problem: price, esp. in
international studies
Group Interviewing/ Focus Groups
Term
• Originally, the term “focus groups” camefrom the sociologist Robert Merton(1940s)– today we often talk about “group
interviewing” instead
– interviewing several people simultaneously– economic, good data source, sometimes
more reliable than individual interviews
Things to remember
• Group dynamics– meaning...
– may be a great benefit…– ...but also destroy the interview: if
someone dominates the situation, datamay be seriously biased
7
Design principles
• To make sure group dynamics does notget our of hand, we ought to– keep groups no larger than 8 people (4-6
works best)– keep groups fairly “equal” -- this makes
sure that no one dominates discussion– the price is that we need typically several
groups
questions?
Sources
GeneralMorrill, Calvin 1995. The Executive Way. ConflictManagement in Corporations. Chicago: University ofChicago Press.Gorden, Raymond L. 1980. Interviewing: Strategy,Techniques, and Tactics. Homewood: Dorsey Press.Krueger, Richard A. ja Mary Anne Casey 2000. FocusGroups: A Practical Guide for Applied Research. ThousandOaks, CA: Sage.Kincaid, Harry ja Margaret Bright 1957. The TandemInterview: A Trial of the Two-Interviewer Team. PublicOpinion Quarterly 21: 304-312.Thomas, Robert J. 1993 Interviewing Important People inBig Companies. Journal of Contemporary Ethnography 22:80-96.
