POE: Understanding Innovative Learning Places and Their ... · ted at a diff ngagement le g’s...
Transcript of POE: Understanding Innovative Learning Places and Their ... · ted at a diff ngagement le g’s...
-
Journal of Education and Learning; Vol. 8, No. 5; 2019 ISSN 1927-5250 E-ISSN 1927-5269
Published by Canadian Center of Science and Education
31
POE: Understanding Innovative Learning Places and Their Impact on Student Academic Engagement—Index 6–8 ‘Alpha’ Survey
Developments Lennie Scott-Webber1, Roger Konyndyk2 & Marilyn Denison3
1 INSYNC: Education Research + Design, Estero, Florida, USA 2 Statistical Consulting, Grand Rapids, Michigan, USA 3 DLR Group, K12 Education Practice, Dallas, Texas, USA Correspondence: Lennie Scott-Webber, PhD, INSYNC: Education Research + Design, Estero, Florida, 33929, USA. E-mail: [email protected] Received: August 1, 2019 Accepted: August 28, 2019 Online Published: September 19, 2019 doi:10.5539/jel.v8n5p31 URL: https://doi.org/10.5539/jel.v8n5p31 Abstract New evidence builds upon the Student Engagement IndexTM and Teacher Engagement IndexTM research (Scott-Webber, Konyndyk, & French, 2019; Scott-Webber, Konyndyk, French, & French, 2018; Scott-Webber, Konyndyk, French, Lembke, & Kinney, 2017) determining post-occupancy answers to, “Can we demonstrate that the design of the built environment for grades 6–8 impacts student academic engagement levels post-occupancy?” The early studies used respondents from grades 9–12. This one is from users in grades 6–8 (‘alpha’ pilot). All studies were conducted in the USA as convenience samples. Engagement performance is a high predictor of student success across multiple domains and learning/work experiences. Specifically, “Research that shows that engagement, the time and energy students devote to educationally purposeful activities, is the best single predictor of their learning and personal development” (Anonymous, NSSE, 2010, p. 2), and thus our research focus. From both the students and educators perspectives, design of the built space impacts engagement performance (p < .0001). Keywords: academic engagement, architecture, learning place design, post-occupancy evaluation, survey development 1. Introduction 1.1 Does Design Make a Difference in Learning and Teaching? Why survey students and educators to see if design makes a difference in their everyday learning or teaching situations? Because, evidence indicates engagement performance is a high predictor of success across multiple domains and learning/work experiences. Specifically, “Research shows that engagement, the time and energy students devote to educationally purposeful activities, is the best single predictor of their learning and personal development” (Anonymous, NSSE, 2010, p. 2), and thus our research focus. Furthermore, and perhaps just as importantly, most of our human experiences are inside this built ‘box’ called school; the USA average is 6.5 hrs per day for approximately 180 days = 1,170 hrs per year; added up = 15,210 hours. Therefore, it stands to reason that where we spend our time and how these places are designed to support individual needs is critical to understand. This current work builds on a career effort and the questions used are framed from multiple researchers in several domains in a holistic approach titled the Users Environmental Interaction Framework.v2© (UEIF.v2) (Scott-Webber, 1999). This study builds upon that work in trying to understanding how the deign of the built environment impacts student academic engagement levels (Scott-Webber, Konyndyk, French, & French, 2018; Scott-Webber, Konyndyk, French, Lembke, & Kinney, 2017; Scott-Webber, 2004; Scott-Webber, Marini, & Abraham, Spring, 2000). This report differs as it studies a new age cohort - students in grades 6 to 8. The research question for this study continues to be the same as for the higher grades of 9 to 12. It is, “Can we demonstrate that the design of the built environment for grades 6-8 impacts student academic engagement levels post-occupancy?” The research design is explained next.
-
jel.ccsenet.o
This reseapost-occupimpacts huof the desqualitativeuses a mixand qualita
1) Qu2) Qunaturinqui23).
The researinquiry relthe contex1.2 ExplorAs we resenew, we sAlthough mis that “…they have Our researdesign of their respounderstandstudents, (supportinglearning sp
• Red
• P• A
asen
• C
org
arch team utipancy. The fouman behaviorsign solution. Me and quantitatxed-method (Jative) for compuantitative techualitative techre, and may iiry, cultural pr
rch question alating to how
xt of the educatring the Imporearch new dessurvey post-ocmany definitio
…POE can be dbeen built and
rch over multithe school ma
onses differ slid: (1) the leve(2) whether sg the educationpaces. This stu
Research-grounductional levelerceptions rec
Awareness of ts the individungagement
Comparison to
lized a Humacus for humanr, importantly Multiple methtive. To ensureJohnson, Onwuparative purpohniques = liter
hniques = “obsinclude particirobes, and oth
Figure 1. A h
and the researthe built envirtional setting ftance of Studyign concepts fccupancy at aons of Post-Ocdefined as ‘thed occupied foriple years has akes a significightly. Engageel of academicstudents at granal efforts, andudy followed anded questionls across multieived from bohe building’s
ual learning sp
instructional s
Journal of Ed
an-Centered Rn-centered reshow users beh
hods are often e a rich set of uegbuzie, & T
oses, and thus lrature review, serving and talipant observat
her methods d
human-centere
rch design are ronment impacfor grade levelsying the Problefor education, aa minimum ofccupancy Evalu process of evr some time’ (indicated that cant differencement is a critc engagement ades 6–8, a nd then (4) howa rigorous sciennnaire proven iple years th user groups(macro or ove
paces’ (micro o
strategies and h
ducation and Le
32
Research Desiearch understahave and if tha
used to answf data is generaTurner, 2007) rlimits research“questionnaireking to peopletion, artifact esigned to sam
ed research des
in harmony acts human behs 6 to 8 and usem from a Postand in order to
f at least threeuation (POE) h
valuating build(Preiser, Harvfrom both the
e relative to aical piece of tas perceived b
new cohort lew this knowledntific research
to be both
s, students and erall/rest of theor classrooms
how these buil
earning
ign (HCRD) ands how the at behavior wa
wer research quated, the HcRDresearch desigh bias. The HcRes and surveyse…these methanalysis, photmple human e
sign protocol (
as the HcRD havior. Our ressed post-occupt-Occupancy Po mitigate the e months afterhave been prop
dings in a systevey, & Edwarde students’ andacademic engathis research. Tby both studenvel can disce
dge may be useprotocol and preliable and
educators, aboe built environ) design and h
lt spaces suppo
protocol to stdesign of the
as anticipated iuestions, but fD (see figure n (inclusive oRD method us” (Hanington, ods are typicato and diary experience” (H
(HcRD)
is human-censearch questio
pancy. Perspective effect of experr occupation iposed, a usefuematic and rigd, 1988,” in, Ad the educatoragement perfoThis work hasnts and by edu
ern this impaced to impact dprovided the fovalid, pre-tes
out academic enment) buildinhow they cont
ort these endea
Vol. 8, No. 5;
tudy this que built environin the developfall into two a1) protocol al
of both quantitses the followin
2010, p. 22), aally ethnographstudies, conte
Hanington, 201
ntered and supon asks the sam
riencing somein a new buil
ul, classic definorous manner
Anonymous, 2rs’ perspectiveormance. How the opportuniucators about
ct, (3) if desigdesigns for theollowing: sted at a diff
engagement leng’s design, astribute to leve
avors.
2019
estion ment ment
areas, ways tative ng: and hic in xtual
10, p.
ports me in
thing ding.
nition after
008). es the
wever, ity to their
gn is next
ferent
vels well
els of
-
jel.ccsenet.o
Overall, thknowledge(overall/rework to bu1.3 RelevaThe Stude
builda hol1999)interathe dDimeZoneEnvir(Hall2001;
The curreStricklandfor gradesdesigned tenvironmeasks how wof choice environmebe comfor[the situatithe valuesquestions. The Teacheducator tperspectiveducator hdigs in deesupport theThis reseawith each
org
his 6–8 ‘alphe that the desiest of built learuild reliable anant Scholarshipnt Engagemens on a career eistic approach). A review action/engagemdesign of the ensions of Vals at the microronment Behav, 1966; Somm; Scott-Webbe
Fig
ent study build, & Kapitula, 2s 9–12 survey to ask a series ent) (2) the clawell one can mand control
ental qualities rtable and be cion of each scs placesd on
her Engagemeto rate his/her
ve is critical ahow the designeper in the eduem.
arch was attemother, with th
a’ pilot (3 scign and use ofrning spaces) and valid instrump nt IndexTM (SEeffort and the qh titled the Useof the framew
ment understanbuilt spaces—
lue and Enviroo level. This frvior Theorists
mer, 1959; Maer, 2000 & 200
gure 2. Users e
lt upon early 2013; Scott-Winstruments.
of questions reassroom (a micmove about, naover where aare touched upconnected to ochool, and its odifferent type
nt IndexTM (Tr student’s en
and the teachen of the built ucators perspec
mpting to ‘provheir teachers, a
Journal of Ed
chools, studenf the built spaareas impacts ments for whic
EI) questionnaiquestions useders Environmework is nextndings with mu—the micro lonment, (3) twframework has, in an effort t
aslow, 1943; B04; Scott-Webb
environmental
work done Webber, Marini
This newest related to the Ucro environmeavigate from inand how studpon; accessibliothers; and moorganization’ses of learning
TEI) questionaingagement lever responses aspace is suppoctive to unpack
ve’ that the deand with the ac
ducation and Le
33
nts n = 2,007ace at both thestudent academch to study this
ire is focused od are framed frental Interactio. The graphicultiple facets ievel, or class
wo Responses s built on the rto more fully Bloom, Krathwber, Abraham,
l interaction fra
in higher edui, & Abraham,research effort
UEIF.v2 relativent), or learninnstructional strents may wisity to ‘tools’ wotivational facs culture] of thg experiences.
ire follows thevels using theare compared orting his/her k the instructio
sign of space cademic conte
earning
; teachers n =e micro (classrmic engagemes important qu
on the studentfrom multiple ron Frameworkc in Figure 2including thesesroom, and mof Internal an
research of mexamine spacewohol, & Har& Marini, 200
amework.v2©
ucation (Scott Spring, 2000)t studies gradeve to two spating spacebuilt erategy to instrush to learn anwithin their leactors. It then fhe school and
Finally, it en
e SEI with twe same questito the studentneeds, again a
onal strategies
makes a diffeent. This work
= 210) continroom/learning ent levels. Thisuestion.
’s perspective.researchers in k.v2© (UEIF.v2 represents te specific segm
macro level, ond Behavioral,
many others, pae and its relatirrow, 1956; E00) (see Figure
© (UEIF.v2)
t-Webber, 20), and the relates 6 through 8ial areas (1) thenvironment ofuctional strategnd with whomarning spaces; focuses on thetheir perceivends with a se
wo specific focions the studet responses. Sat the macro aand how the d
erence in how tried to under
Vol. 8, No. 5;
nues to extendplace), and m
s team continu
. This current several domai
v2) (Scott-Wethe complexitments: (1) layeor overall, (2), and (4) Proxarticularly clasionship to its lliot & Covine 2).
14; Scott-Wetively new rese8. The surveye overall (or mf the school. Itgy; the opportm; whether inability to see, situational cud understandinet of demogra
ci. First, it askents’ receive. Second, it askand micro levedesign of the sp
individuals enrstand and mea
2019
d the macro ues to
work ins in bber, ty of ers of
two xemic ssical users gton,
bber, earch s are
macro t also unity
ndoor hear,
ulture ng of aphic
ks the This
s the els. It paces
ngage asure
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
34
what was impacting interactions or engagements. We believed it was important to not just answer the research question, but try and provide a tool, or index and measurable awareness levels to use as gauges of engagement and environmental fit. This document reports on a survey of three middle schools, grades 6–8, and is the first time we have attempted to survey students in this age group, as an ‘alpha’ test, and expected to learn some things to do differently the next time around. Our earlier surveys were of high schools, grades 9–12. Surveying middle school students raised several questions:
1) Would middle school students understand a survey of the type we need to do? 2) How much language would need to change to ensure these cohorts would comprehend our questions from
high school? 3) Would middle school students be willing to respond to a survey? 4) How much would the survey need to be simplified and/or shortened for them? 5) Would we get results similar to the high school surveys, and how would results differ? 6) How might middle school teachers answer differently than high school teachers?
We are pleased that overall, students responded well to the survey, and the results were similar to those of the high schools’ surveys. To address our concern stated in #2, we asked a former assistant superintendent and some of her educators to review the 9–12 text and help ensure the vocabulary and meaning would fit with this age cohort. Some slight changes were made to the original surveys (three pilots for high school) as a result of their reviews. While we encountered a couple issues along the way, there was nothing that would call the validity of the survey into question. As one would expect, there were some differences between the middle school and the high school survey responses. 2. Method A short definition for each step in this Human-Centered Research Design protocol (refer back to Figure 1) is shared:
• Discover [D]: Develop a research question/hypothesis and understand what will be the best research design, methods and techniques to find answers, and use them to gather data. It’s best to use three techniques to ensure bias is reduced. Once gathered the researcher(s) puts this information into appropriate format(s) for analysis. Whether using quantitative methods or qualitative methods, all data will be worked to produce some numerical findings. Once this latter stage is done, these become research ‘instruments’ or tools. (NOTE: a human subject’s protocol has been reviewed by a third party prior to beginning work with a client; all consent forms approved and received).
• Analysis [A]: Take the data from the research techniques and use appropriate methods to break down the information. By using multiple discovery techniques to avoid bias ensures the comparisons generates consistent and reliable findings. Use statistical methods when appropriate. Pilot test and test again to vet the data for reliability and validity.
• Synthesis [Sy]: Recognize the analysis phase of this work only generates facts. What these facts ‘are saying,’ how each is connected to the next is revealed by generating meaning and understanding relative to the original question. This segment takes time and expertise to clarify and built a ‘truthful’ and unbiased consensus from the data.
• Share [S]: Be prepared to share information to multiple audiences and for multiple purposes—clients, designers, conferences, and research manuscripts.
• Plan [P]: Know all data reveal a truth—not always the ones we’re looking for or expecting. Be prepared to plan for next steps (ex, go back and address an issue found in a design and/or adjust the design solution for the next time it is used).
The sample information and response rates are shared next. 2.1 Sample and Response Rates and Analysis of Missing Data There were three schools used as convenience samples with purposeful user groups (students and teachers), each school designed by a particular architectural firm, with 1,381 total responses to the student survey. Student response to the survey was very good, with over 60% of the students in each school responding to the survey. Response rates from the teachers were much lower at schools A and B; the reason is unknown (see Table 1). The small number of teachers responding made it difficult to draw statistically valid inferences about how the
-
jel.ccsenet.o
teachers vhere are fr Table 1. R
We had a Figure 3).
Similarly, more than
Of the “Otlittle probland anothefatigue” in5).
org
viewed the diffrom the 6-8 stu
Response rates School
A B C
good mix of s
we had a gooone grade; see
ther” group, thlem with non-rer 12% skippen the student su
ferent schoolsudy.
by school No. Students
848 479 680
students by gr
Fi
od mix of teace self-identifie
Fig
here was one aresponse to qued only one quurvey, as the n
Journal of Ed
. To ensure an
Respondents 545 403 433
rade level, whi
igure 3. Grade
chers by gradeed other grade
gure 4. Grade l
answer each ouestions on theuestion (but nonumber of skip
ducation and Le
35
nonymity, sch
% No64.3 84.1 63.7
ich proved to
level indicato
level; focus olevels taught (
level indicator
f “Office”, “Ge survey. Amonot the same qupped questions
earning
ool names are
o. Teachers R120 45 45
be an importa
rs/students
on grades 6-8.(see Figure 4).
rs / teachers
Grade 5”, and “ng the studentuestion). Hows rose toward t
e coded, and a
Respondents 31 20 29
ant factor in th
. Of course, m
“Counseling”. ts, 68% answer
wever, we did the end of the
Vol. 8, No. 5;
all data represe
% 25.8 44.4 64.4
he student data
many teachers t
Overall, therered every quessee a little “susurvey (see F
2019
ented
a (see
teach
e was stion, urvey igure
-
jel.ccsenet.o
Eighty teaLooking anumeric-raquestions w95% or beshown in tsupported
3. Results3.1 ReliabThe objectquestions Alpha is away. We hsurvey, onrating valuwith only and the stuThe idea oAnother mare strongevidence othese survdifference surveys, an
org
achers in all tooat the questionating item, andwas not a probetter on those the graph belothe teaching m
ility & Validitytive behind meon the survey
a number betwehave very goodnly question 4 ues were both gquestion 1 (im
udent surveys aof validity in a
method to detergly correlated of validity in beys are similais that the re
n intuitively p
Figure 5.
ok the survey, n items with d 4 teachers skblem on this suquestions. Fivw. The greates
method the teac
Figure 6.
y easuring reliaby. Consistent ween 0 and 1, wd reliability nu(classroom angreater than .7
mportance of vare reliable. a survey is to rmine whetheras expected)
both the studenar to, but not idesponse by grapleasing result,
Journal of Ed
Survey respon
although one numeric answkipped only onurvey; for eachve teachers quist non-responscher used (see
Survey respon
bility of a survewith the earlie
with a value of umbers overallnd building rat70. On the studvarious items)
ask whether r the survey is
and divergennt and the teachdentical with, ade level is of, given the yo
ducation and Le
36
nse rate—surve
person answewers (1–4 and
ne item (see fh question grouit after questiose was for the Figure 6).
nse rate—surve
ey is to assess er surveys, wef 1 indicating thl for both the ttings) had Alp
dent survey, allbeing less tha
the study meas providing thent (weak or nher surveys. Athe surveys d
f greater impoounger ages of
earning
ey fatigue/stud
red only the d6–9), 70 of
figure 6). Non-up with numer
on 9, so there second part of
ey fatigue/teac
whether peope used Cronbahat questions ateacher and thepha below .80,l the values of
an .80. Thus, w
asures what the results intendnegative corre
A strong indicadone of high scortance in thisf the students.
dents
demographic quthe 80 teache-Response of ric answers thewas a bit of “f question 10,
chers
le give similarach’s Alpha toare answered ie student surve, and the classf Alpha were awe conclude th
he researcher dded is throughelation) validitator is the fact chool studentss survey than Teachers also
Vol. 8, No. 5;
uestions at theers answered eindividual nume response rate
“survey fatiguewhether the de
r answers to simo assess reliabin exactly the eys. On the teasroom and builso greater tha
hat both the tea
desired to meah convergent (ity. We have that the findin
s and teachers.in the high sc
o seem to sens
2019
end. every meric e was e”, as esign
milar bility. same acher lding n .70, acher
asure. items good gs of One chool e the
-
jel.ccsenet.o
building asConvergenvariables; theoreticaland low coA similar pand questiinterest, al3.2 StatistiClassroom1 as beingteachers anseems to bstudents wtemperaturon our 5-pwere mino“mentoringsociability(testing, cr
org
s having a greant and divergequestions 2–9l question 1 (Iorrelations whepattern occurs ons 2, 3, 4, 6, long with a comics and Data A
m ratings overag more importnd students. Tbe universal buwere slightly mre “Excellent”point scale. Thor. The most dg” shows less
y, i.e., inter-perritical thinking
F
Fig
ater impact onent validity in, about how thImportance ofere one wouldin the teacher 8, and 9 have
mparison of stAnalysis all were fairly tant than did t
The environmeut the studentmore critical
” or “Very goohe average studdivergent varias divergence. rsonal relation
g, creativity, et
Figure 7. Impor
gure 8. Environ
Journal of Ed
n them than at tn the student hings actually w…) has lower
d expect them tr survey. Quest
good correlatitudents and tea
similar (see Fithe students. Q
ental quality qus rated the nothan the teac
od”. The overadent rating for ables are “beloThese three m
nships and engtc.).
rtance of items
nmental quality
ducation and Le
37
the high schoosurvey may b
work and are wcorrelations w
to be. tion 1 has relations with eachachers are offer
igure 7). Not sQuestion 1 hauestions generise level as behers. For both
all average ratir Noise Level wonging to the more than the gagements, rat
s for engageme
y ratings from
earning
ol level. be seen in thwell correlatedwith all the oth
tively low corrh other. Some ired next.
surprisingly, tead the greatestrated complaineing just as bah students aning by studentswas 2.74 (see school’s commother six var
ther than indiv
ent/students vs
both student a
STU
DEN
TS
STU
DEN
TS
he correlationsd with each othher variables.
relations with individual que
eachers saw tht divergence onts about the tad (see Figure nd teachers, ons was 2.78, anFigure 9). Difmunity” and “riables are relavidual experien
s. teachers
and teachers
TEAC
HERS
TEAC
HERS
Vol. 8, No. 5;
s of the compher, while the Thus, we see
the other variaestions of parti
he items in queof opinion betwtemperature, w8). In genera
nly 26% ratednd 2.74 for teacfferences by sc“collaboration,ated to inter-ances and outco
2019
posite more high
ables, cular
estion ween
which l, the d the chers chool ” yet
active omes
-
jel.ccsenet.o
On questio
3.3 The RoSome demtheir gradethey held. French, Lschool levlevel prodschool. In the student3.4 PrincipBefore coquestion gquestion gFor each nwhole grouset to a miThe compo
org
on 6, the value
ole of Demogramographic infoe level, and wWe knew whiembke, & Kin
vel, the schoolsduced some int
the regressiont grade level mpal Componen
onstructing thegroup, and a pgroup would benumeric questup. If more thaissing value. Tosite variable f
Figure 9. Per
es of the school
Figure
aphic Factors rmation was c
we asked the teich school theynney, 2017; Ss in this surveyteresting diffens described bmattered. nt Analysis & Ce composite vprincipal compe a reasonable tion group, thean one questiohe composite vfor question 7
Journal of Ed
rceptions of no
l, teachers wer
e 10. Perceptio
collected on theachers what gy were at withoScott-Webber, y did not becorences, perhap
below, demogr
Composite Varvariables, descponents analysproxy for the g
e mean of the on in the groupvariables were(At the end of
ducation and Le
38
oise and tempe
re sometimes m
ons of what the
he surveys. Forgrades they tauout asking. UnKonyndyk, F
ome a factor inps a reflectionraphic factors w
riables cribed below,sis was perforgroup. group was us
p was skipped, e used in the ref a school day,
earning
erature comfor
more positive t
e school values
r the students ught, their gennlike the “OmeFrench, & Frenn most of the qn of the changewere checked
, Cronbach’s rmed in order
sed as the comthen the comp
egression analy how often do
STU
DEN
TS
rt levels
than the studen
s
we asked abounder, and whatega’s” (Scott-Wnch, 2018), su
questions. For es a child expfor relevance,
Alpha was cto verify that
mposite variabposite variableyses and in theyou feel that y
TEAC
HERS
Vol. 8, No. 5;
nts (see Figure
ut their gendert academic degWebber, Konyurvey for the students, the g
periences in m, but typically
computed for t the mean of
ble to represene for that groupe cluster analysyou…) was us
2019
e 10).
r and grees
yndyk, high
grade middle
only
each each
nt the p was sis. ed as
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
39
the engagement index for both teachers and students. The overall average values for both the teacher engagement index (TEI) and the student engagement index (SEI) were similar to the corresponding values on the “Omega” survey, and like that survey, the student values showed greater variability (see Figure 11). The results of the analysis are next.
Figure 11. Distributions of the engagement indexes
3.5 Analysis Results 3.5.1 Regression Analysis/Students In general, the school played little or no role in the regression results in this survey, while gender had some impact, and grade level had even more impact. Consistent with the surveys of high school students, girls tended to have slightly higher engagement than boys, and those who opted for the “prefer not to say” sector having lower engagement. Engagement also tends to drop as the grade level rises. Results of the regressions in this survey were very similar to those of the “Omega” at the high school level. While there were a couple of statistically significant interactions, they were of little practical importance. The impact of the various questions on student engagement is basically the same across school, gender, and grade level. These are shared next:
• SEI and Question 1: (The importance of various items for engagement). Of all the numeric questions (1–4, 6, 8, and 9), question 1 has the weakest relationship with student engagement. Regressing the SEI on just question 1 yields a small r2 = .13. Students who perceive that the items in question 1 are important show only a slight tendency to have a higher level of engagement.
• SEI and Question 2: (How well do the classrooms provide you with the ability to…?). Regressing the SEI on question 2 by itself gives a very strong r2 = .29 with p < .0001. The school is not a factor, but adding in grade level and gender to the regression raises r2 to .34. Also, there is no interaction of Q2 with either grade level or gender, and we see that students who believe that the classrooms provide the ability to see and hear well and give access to appropriate items are also likely to be more engaged. The values of r2 here are quite similar to those of the “Omega” survey (see Figure 12).
Student Survey
Mean: 3.73
Std Dev: 0.82
Teacher Survey
Mean: 3.98
Std Dev: 0.46
STUDENT TEACHER
-
jel.ccsenet.o
• SE
qleinstQ
• SE
asraanleef
org
SEI and Questquestion, the scevel, though nnteraction betwtatistically sign
Q3 to predict th
Figu
SEI and Questisked to rate thate the buildinnd engagemenevel, though stffects (see Fig
Figu
tion 3, (How mchool figures innone with schoween question nificant, only he SEI gives r2
ure 13. Regress
ion 4: (Pleaseheir classroomsng overall, a cnt was fairly sttatistically sigure 14).
1
1.5
2
2.5
3
3.5
4
4.5
5
Q
Journal of Ed
ure 12. Regres
much impact dnto the regressool or gender.
3 and grade lincreases r2 fr
2 = .26, with p
sion plot of the
e rate the follos on several aschange from thtrong, with r2 =nificant, raised
Q2 How well c
ducation and Le
40
sion plot of th
does the desigsion, and there Thus, using qlevel yields r2
om .30 to .31,< .0001, show
e SEI on Q3 (im
owing aspects spects (noise, the grades 9–1= .29, very simd the value of
classrooms prto...
earning
he SEI vs. Q2
gn of the class is an interacti
question 3, gen2 = .31, with p, resulting in li
ws a good assoc
mpact of class
of the classrotemperature, li2 survey. The
milar to the “Of r2 by only .0
rovide the abi
sroom have oion between qunder, grade levp < .0001. Theittle practical iciation by itsel
sroom design)
oom environmeighting, etc.), te association b
Omega” survey3, indicating t
ility
Vol. 8, No. 5;
on you…). Foruestion 3 and gvel, school, ane interaction, wimpact. Using lf (see Figure 1
ent). Students though they dibetween the ray. Gender and gthat they are m
2019
r this grade d the while only
13).
were d not
atings grade minor
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
41
Figure 14. Regression plot: SEI regressed on Q4 (classroom ratings)
• SEI and Question 6: (What do you think your school values…). Consistent with the “Omega” survey of
grades 9–12, we see a strong relationship between student engagement and whether the student believes that the school values items such as creativity, critical thinking, and mentally challenging work. Regressing the SEI on question 6 alone gives r2 = .39. (The corresponding regression on the “Omega” survey had a similar r2 of .45). This question and question 8 (impact of the design of the school on access to peers and teachers and on the ability to move in the classroom) had the strongest association with student engagement, of all the questions.
Once again, school is not statistically significant, even at the p = .05 level. Adding grade level and gender into the regression gives only a small improvement in r2, to .42. We may conclude that the effect of the values of the school is the same across gender, grade level, and school, and the effect of the perceived values on engagement is quite strong (see Figure 15).
Figure 15. Regression plot: SEI vs. values of the school
• SEI and Question 8: (How much to you believe the design of the school overall impacts your ability)
to….). Once again, school is not significant, nor is any interaction terms. Grade and gender add only a little to the information provided by question 8, improving r2 from .42 to .45. We have another very
1
1.5
2
2.5
3
3.5
4
4.5
5
Q4 Ratings of Classrooms
1
1.5
2
2.5
3
3.5
4
4.5
5
Q6 Values of the school
-
jel.ccsenet.o
stte“Osi
• SE
reinst
In summarbuilding an3.5.2 Regr3.5.2.1 ThLooking a(TEI). Thea lower mone-way Alevel of edthe p = .00(see Figure
org
trong associatieacher and proOmega” (9–12imilar r2 = .39
SEI and Questelationship betn the grade levtatistically signry, these regrend the student ression Analys
he TEI and Demat gender by itse means of the
mean of 3.58, bANOVA for thducation appea012 level, givine 17).
ion here; studeovides the ab2) survey, the(see Figure 16
tion 9: (Leveltween Q9 and vel improves nificant, but ofessions show aengagement in
sis/Teachers mographic Varself, there weree TEI were an but with only the TEI on levar to have no ng r2 = .166, w
Figure 17.
1
1.5
2
2.5
3
3.5
4
4.5
5
Q
Journal of Ed
ents who believility to move e regression o6).
Figure 16. Re
l of impact ofthe SEI, with r2 to .33. Genf little practicaa strong statisndex.
riables e no detectableall-but-identicthree people in
vel of educatioimpact on the
with school C h
Boxplot of tea
Q8 Impact of
ducation and Le
42
ve that the desin the classro
of the SEI on
egression plot
f the design or2 = .30 with
nder, school, aal importance, tically signific
e differences ical 3.98 for won that group, t
on yielded no e TEI. Differenhaving the high
acher engagem
design of the
earning
sign of the schooom tend to hthe correspon
: SEI vs. Q8
of the buildingthe SEI regres
and the interacas each of themcant relationsh
in the values oomen and 4.01there is no stastatistically signces by schoolhest mean (4.1
ment index by s
e school over
ool facilitates have higher ending question
g on your…).ssed on questioction of grade m improves r2 hip between sa
of the Teacher for men. “Pre
atistical signifignificant diffel were statistic
17) and school
school
rall
Vol. 8, No. 5;
access to peerengagement. Inn produced a
. We see a ston 9 alone. Ad level with Qby only .01.
atisfaction with
Engagement Iefer not to say”icance. Similarerence. Gendercally significan A the lowest
2019
s and n the very
trong dding 9 are
h the
Index ” had rly, a r and nt, at (3.75)
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
43
3.5.2.2 The TEI and the Numeric Questions (1–4 and 6–9) Perhaps because of the smaller sample size in this survey, the relationships between the numeric questions and the TEI lack statistical significance. As we saw with the “Omega” 9–12 survey, the school is a better predictor of the TEI than are any of the numeric questions. A summary of this information is next:
• Question 1, the second part of question 1 (Design supports), and question 2 do not come close to statistical significance for the TEI, whether with or without the demographic variables
• Question 3 hovers near statistical significance, but only at the p = .05 level. The school is of greater importance to the TEI than is question 3. Regressing the TEI on just question 3 yields a measly r2 = .055, with p = .0425, while including the school in the model gives r2 = .22.
• Question 4 Classroom Ratings are statistically significant, with r2 = .10 and p = .0052. Again, school is more important; adding it to the model raises r2 to .25. This actually a bit stronger relationship than we saw in the “Omega” survey.
These results are similar to those of the “Omega” from grades 9–12 survey, which also found only weak relationships between teacher engagement and the other numeric questions. Cluster analysis is described next. 3.6 Custer Analysis 3.6.1 Clusters for the Students For the students, four clusters worked well. As with the previous surveys, Ward’s Method was used for the clustering. Average engagement in the clusters varied widely, from a high of 4.70 in cluster 2 to a low of 2.91 in cluster 4. Alas, cluster 2, with the highest average engagement, is the smallest cluster. Except for question 1, part 2, a higher number means “good” and a lower number means “bad” for the questions. Question 1, part 2 asked whether the design helped the items in the first part of Q1 succeed, and the scale of question 1, part 2 was reversed, with 1 = Yes and 2 = No. That explains the strange look in the graph at that question. Thus, the four clusters are quite neatly stratified, with cluster 2 having the highest means for each question, followed by cluster 1, and then cluster 3, with cluster 4 having the lowest means for each question (see Table 2). Table 2. Student cluster means
Cluster Count Q7 Student Engagement
Index
Q1 Importance of various items
Q1 part 2 Is the design helpful? (1
= Yes, 2 = No)
Q2 How well classrooms provide
the ability to...
Q3 Impact of classroom
design
Q4 Ratings of
Classrooms1 356 4.19 3.57 1.06 4.15 3.50 3.62 2 133 4.70 3.97 1.03 4.51 4.58 4.08 3 499 3.46 3.20 1.14 3.43 2.73 2.91 4 182 2.91 2.66 1.55 2.85 2.16 2.38
Cluster Q6 Values of the school Q8 Impact of design of the
school overall Q9 Impact of design of building's physical spaces
1 3.97 4.14 3.70 2 4.57 4.76 4.81 3 3.22 3.17 2.82 4 2.48 2.40 2.10
A corresponding graph of the cluster means is in Figure 18 (see Figure 18).
-
jel.ccsenet.o
A look at tFigure 19)
Consistentand 2, with
org
Figure 18
the clusters by).
t with the regrh a bigger drop
8. Correspond
y school does n
ression results,p-off in “prefe
Journal of Ed
ing cluster gra
not show stron
Figure 19. Cl
, one can see aer not to say” w
ducation and Le
44
aph of interpret
ng differences,
uster means by
a little higher when viewing c
earning
tation of desig
, a change from
y school
percentage of cluster by gend
Studen
gn support/stud
m the “Omega
f females than der (see Figure
nt
Vol. 8, No. 5;
dent
a” 9–12 survey
males in cluste 20).
2019
y (see
ters 1
-
jel.ccsenet.o
The graphdirection!’
3.6.2 ClustThe results3 (blue) anphysical eengaged tvarious qu
org
h of clusters by’ Next, is the c
ters for the Tes of the clusternd 4 (orange) benvironment, teachers helps
uestions (see Fi
y grade level clustering for th
achers r analysis of thboth show higthe 9 teachers
explains whyigure 22).
Journal of Ed
Figure 20.
(Figure 21) ishe teachers.
Figure 21. C
he teachers aregh engagements of cluster 3 y the regressi
ducation and Le
45
. Cluster by ge
s more dramat
Cluster by grad
e reminiscent ot, while the 26are quite unh
ons showed s
earning
ender
tic, and the ind
de level
of the surveys teachers of cl
happy with it.so little conne
dicators are go
of high schooluster 4 are ver This diverge
ection between
Vol. 8, No. 5;
oing ‘in the w
ol teachers. Clury happy with
ence among hn the TEI and
2019
wrong
usters their ighly d the
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
46
Figure 22. Corresponding cluster graph of interpretation of design support/teachers
Table 3. Cluster analysis
Cluster Count Q7 Teacher Engagement Index
Q1 Importance of various items
Q1 Design supports for activities
Q2 How well classrooms provide the ability to…
Q3 Impact of classroom design
1 23 3.908 4.061 1.087 3.555 3.065 2 16 3.758 2.900 1.238 3.392 2.292 3 9 4.097 4.333 1.956 2.374 2.870 4 26 4.144 4.046 1.015 4.024 3.494
Cluster Q4 Ratings of
Classrooms Q4 Rate the building overall
Q6 Values of the school
Q8 Impact of design of the school overall
Q9 Impact of design of building's physical spaces
1 3.060 3.207 3.588 3.661 3.435 2 3.102 3.188 3.083 3.125 2.696 3 3.083 3.208 3.370 2.667 3.079 4 3.928 4.024 4.175 4.208 4.203
The cluster means are in the Table 3 (refer back to Table 3). As with the student survey, the second part of question 1 (“Design supports?”) had 1 = Yes and 2 = No. Therefore, low values are desirable rather than high values for that question. Note that cluster 4, with the highest TEI, also had the “best” mean for each question other than the first part of question 1. Cluster 3, which also had a high mean TEI, had the “worst” ratings for the “Design Supports” part of Q1, and questions 2 and 8, along with nearly bottom ratings for questions 4 (both parts) and 6. Clusters 1 (red) and 2 (green) had similar average TEI values, but cluster 1 assigned very low importance to the items in question 1, while saying that they were very well supported, and cluster 1 saw much more impact from the building than cluster 2 (questions 3, 8, and 9) Comparing the teacher cluster by school, results show that School B has the highest percentage of teachers in cluster 4, the “happy” group, with the highest means for TEI, and no teachers in cluster 3, the group that seems quite dissatisfied with their classrooms, despite having a high TEI (see Figure 23). What we found regarding the impact the physical surroundings has on the teachers and students follows.
Teacher Cluster 1
Cluster 2 Cluster 3
Cluster 4
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
47
Figure 23. Clusters by School
4. Discussion 4.1 Impact of the Physical Surroundings 4.1.1 Teachers The table below shows the results of t-tests for H0: Mean = 3 vs. Mean > 3 for the questions about the impact of the building (see Table 4). Meanings of the numbers:
• Q3: 3 = Moderate impact - 4 and 5 are Makes a noticeable (or big) difference. • Q8: 3 = Acceptable - 4 and 5 are Easy or Very easy. • Q9: 3 = Moderate impact - 4 and 5 are Makes a noticeable (or big) difference.
Different question items elicited distinctly different average values of the perceived impact of the built environment. Items significant at the p = .01 level are underlined. The overall view seems to be that the impact of the design of the school on teachers’ ability to do things is better than merely “Acceptable” (Q8), but that the impact of the building on teachers (Q9) is greater than the impact on students (Q3). Average values in question 3 are slightly lower than in the “Omega,” while answers to Question 9 tended to be a little higher than for the corresponding questions in the “Omega.” Table 4. Impact of built environments/teachers (selected items)
Question Item N Mean T-Test statistic P
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
48
4.1.2 Students Students were asked most of the same questions as the teachers about the impact of the built environment. Question 3 asked about the impact of classroom design on the student’s attitudes. Question 8 asked about the impact of design on the student’s ability to do certain activities, and question 9 asked about the impact of the physical spaces on various things. The table below shows the results of t-tests for means of 3 or higher (see Table 5). For questions 3 and 9, this means at least a moderate impact vs. a little or no impact. For question 8, this means the impact is acceptable or better. Overall, the averages here are a little higher than those on the “Omega” survey. The descriptors by grade level are next. Table 5. Impact of built environment/students (selected items)
Question Item N Mean T-Test statistic P
-
jel.ccsenet.o
4.1.4 “Mov4.1.4.1 StuComparingreported ahand-in-hamove, is aquestion 1actually be
org
Figu
vement” Quesudents/Movemg “movementabove, but theand with highea bit theoretica1 are less stroeing able to mo
ure 24. Grade l
Figure 25.
stions ment t” questions we point is woer student eng
al, and not surpongly related tove, as shown
Journal of Ed
level of studen
. Averages Q6
with student orth repeating gagement. For prising that quto engagemenin the table be
ducation and Le
49
nt vs. mean val
: what the scho
engagement sfor some indstudents, quesestion items ab
nt than are theelow (see Tabl
Y
earning
lues of compos
ool values/stud
somewhat dupdividual questistion 1, about bout the percee questions itele 6).
YR 6
YR 7
YR 8
site variables
dents
plicates the reions. The abithe importanc
eived importanems in questio
Vol. 8, No. 5;
egression anaility to move ce of being ab
nce of movemeons 8 and 9 a
2019
alyses goes
ble to ent in about
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
50
Table 6. Questions related to movement and engagement Question Item Correlation with the SEI Q1b Move about classroom to be actively engaged - Importance of… 0.248 Q1d Have the choice to use different parts of the room to work with others - Importance of… 0.208 Q8c Ability to move to engage in classroom activities 0.573 Q8d Choose either to sit/lounge/stand in order to be active in the classroom 0.496 Q9c Ability to move around to become deeply engaged in my learning 0.507
The overall message is that for these middle school students, the ability to move correlates strongly with student engagement. A look at the “ability to move” question items in question groups 8 and 9, and engagement shows that engagement is higher for students who believe that movement is easier (see Figure 26). (A note about boxplots: the line inside the box indicates the median value, and the box itself contains the middle 50% of the values, giving an idea of the “spread” of the values.)
Figure 26. Ability to move corresponding to level of engagement / students
4.1.4.2 Teachers Teachers were also asked some questions directly relating to movement, specifically:
• Q1a Transition in and out of small groups—Importance of… • Q1b Move about classroom to be actively engaged—Importance of... • Q1d Have the choice to use different parts of the room to work with others—Importance of... • Q2i Move around to keep students engaged (How well do classrooms provide this ability) • Q2j Have your students move around to keep themselves engaged • Q8c Impact of the design of the school on: Ability to have your students move to engage in classroom
activities • Q9c Impact of the physical space on: Ability for you to move around to get your students deeply
engaged in their learning. While the student answers to the analogous questions in their survey were positively correlated with the student engagement index, the teachers’ answers to their questions showed little relationship to teacher engagement. The questions listed above are all positively correlated with each other, and they fall into two basic groups: the three parts of question 1 are well-correlated, indicating that teachers who viewed one type of movement as important tended to view the other types of movement as important. The second group consists of items 2i, 2j, and 8c,
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Q8c Ability to move to engage inclassroom activities
Q8d Choose either to sit/lounge/stand in order to be active in the
classroom
Q9c Ability to move around tobecome deeply engaged in my
learning
-
jel.ccsenet.o
which havitems in qgrouped wthey believability to m
Table 7. M
4.1.5 TeacAccordingdone outsistudents dschool A. teachers (s
Students restudents ge
org
ve to do with question groupwith the items ive that movemmove (see Tab
Mean answers t Q1a TransitionQ1b Move aboQ1c ExperiencQ1d Have the Q2i Move arouQ2j Have yourQ8c Ability toQ9c Ability fo
ching Outside tg to the teacheride the classroo not completStill, student
see Figure 27).
eport little diffet outside of th
the ability to p 1. Question n question 1. T
ment is importle 7).
to “movement”
n in and out of smout classroom to bce hands-on activichoice to use diffund to keep studer students move a
o have your studenor you to move aro
the Classroomrs, very little t
oom at school tely agree withs in school B.
ference in nonhe classroom m
Journal of Ed
move, and theitem 9c is on The means of ttant (question
” / teachers
mall groups (of 2 tobe actively engageities during class tferent parts of the nts engaged round to keep them
nts move to engagound to get your s
m teaching is donB; no teacher
h the teachers, B seem to repo
Figure 27. Te
-classroom teamore (see Figu
ducation and Le
51
ey are stronglyits own, thou
the teachers’ a1), but are pe
o 4) - Importance ed - Importance oftime - Importanceroom to work wit
mselves engagedge in classroom actstudents deeply en
ne outside the r at school B rin that they re
ort the most t
eaching outside
aching by gradure 28).
earning
y correlated wugh a factor ananswers to the rhaps a little l
of... f... e of... th others - Importa
tivities ngaged in their lear
classroom at sreported alwayeported much teaching outsid
e the classroom
de level. Intere
with each othenalysis sugges“movement” qless enthusiast
Me3.73.93.8
ance of... 3.63.43.43.6
rning 3.7
school A, whilys staying in tmore non-cla
de the classro
m
stingly, engag
Vol. 8, No. 5;
er, but not withsts that it coulquestions showtic about the a
ean 78 96 88 67 49 41 62 76
le more teachithe classroom.ssroom teachin
oom, similar to
gement rises a b
2019
h the ld be
w that actual
ing is . The ng in o the
bit as
-
jel.ccsenet.o
A one-wayalways taufor the “mmethods aperspectiv4.1.6 Teac4.1.6.1 StuThe studenprojects sestudents re
org
y ANOVA shught in the clamore than halfare often a prves. ching Methodsudent Responsnts reported oneem to decreaseporting each m
Figure 2
hows that the dssroom, the mf” group. Thesredictor of act
and Strategieses nly minor diffese, while lecturmethod as one
Figure 3
Journal of Ed
28. Teaching o
differences in mean engagemese results are tive engageme
s Used
erences by schre and discuss of the top two
30. Teaching m
ducation and Le
52
outside the clas
means are reent is 3.58, cosimilar to wh
ent by the stu
hool, but as theion increase (s
o methods used
methods by gra
earning
ssroom / studen
al (F = 8.0289ompared to 3.9hat we saw inudents. These
e grade level rsee Figure 30)d.
ade level/stude
nts
9, with p < .091 for the 6–25n the “Omega”
data are shar
rises, hands-on. The graph sh
nts
Vol. 8, No. 5;
0001). For stud5% group and ” survey. Teacred from both
n learning and hows the perce
2019
dents 3.94
ching user
team ent of
-
jel.ccsenet.o
Another im
4.1.6.2 TeaIn questionso only theput it neartake-homedifference
There werQuestion 1of the stud
org
mportant way t
acher Responsn 10, which ase total groupinr the top; an obe messages; te
to student out
re too few resp12 (Teaching sdents (see Figu
to illustrate the
F
ses sked about teacng is shown hebvious and imeaching stratetcomes. Perhap
F
ponses to detestrategies used
ure 33).
Figu
Journal of Ed
ese perception
Figure 31. Tea
ching methodsere. The teache
mportant percepgies and pracps this data sho
Figure 32. Teac
ect differencesd), differences
ure 33. Teachin
ducation and Le
53
s from student
aching methods
s used, differeners put lecture/ption to considctices to ensuows a best prac
ching methods
in the “Desigby school wer
ng strategies u
earning
ts is in the figu
s/students
nces between /presentation nder (see Figureure a variety octice scenario
/ teachers
gn Supports?” re not signific
used/teachers
ure below (see
schools were nnear the bottome 32). One of of activities tfor considerati
answers by teant. This ques
Vol. 8, No. 5;
Figure 31).
not significantm, while the stuthe most impotake place maion.
eaching methostion was not a
2019
t, and udent ortant ake a
od. In asked
-
jel.ccsenet.o
In the “Deof the 37 f
5. LimitatAction resresponses,determinindesign of tFirst, whilstudent enexample, twith the bsample of 6. ConclusOverall, thsurveying seemed ot high schomiddle schinteraction
•
•
•
•
•
•
org
esign Supportsfor “Project-ba
tions search at the p and therefore
ng the types othe learning plle the survey
ngagement, it mthat an overalbuilding and cthree schools tsion his survey “wo
middle schoounderstand th
ol students, shool students, n. Some specif
Students aengagemenTeaching tstudents’ alearning s(BehavioraMoving torecognizedConnectingoutside walevels of enWhen respengagemenlearning ouAs with tengagemenwas much
s?” portion of qased, and 2 of t
Figure 34
post-occupance controlling foof teaching strlace afforded ths showed higmust be noted l positive attitclassrooms. Seto all schools i
orked”, althouols, the surveyhe questions asatisfaction wiand the corre
fic findings areand teachers bnt levels at a hitechniques chattitudes seemtrategies, the
al Responses) o learn was imd its importancg to nature anas not done verngagement (Inpondents recognt, the perceptutcomes, and mthe high schont and satisfactweaker (Envi
Journal of Ed
question 12, othe 29 answers
4. Teaching str
cy level recognor variables is rategies used vhose activities
gh correlationsthat correlatio
tude could be econd, one muin the country.
gh some tweay went well. Ts they answereith the physicespondence is e shared in termboth perceivedigh rate of sign
hanged with gm to become
more their
mportant as ie (Environmen
nd being outsiry often. Whenternal Respons
gnized the valuion was there
motivation to aool surveys, tion with the pironmental Dim
ducation and Le
54
nly 7 of the 13s of “Personal
rategy & does
nizes that therchallenging. T
versus the lev. s between saton was not the
behind both bust be very ca.
aks would be hThe students sed like those oal facilities gthe same acro
ms of how theyd the design onificance (Envgrade level, amore negativstudents elici
t increased enntal Dimensionide to learn wn it was, studeses) ues of the schwere higher l
attend classes (we found strphysical enviromensions). Fo
earning
33 respondentLearning” (see
design suppor
re are multiplThis survey wvel of student
isfaction withe same as causbeing more enautious in gen
helpful. Given showed only af the high schooes hand-in-hoss gender, gry connect to thof the built spvironmental Diand along wite. Thus, the ited higher le
ngagement levns)
was viewed as ents responded
hool, or its situlevels of stude(Value Dimensrong correlationment, howevr teachers, the
s gave an answe Figure 34).
rt it?
le factors that worked to inclu
activity perm
h the physicalsation. Someonngaged and beneralizing from
that this was a little ‘surveyool students. A
hand with engade level, and
he UEIF.v2 arepace impactedimensions) th these moremore teacherevels of acad
vels, and both
important, bud positively tha
uational culturent academic esions) ons between ver, for teachee cultural/situa
Vol. 8, No. 5;
wer of “Not w
contribute to ude some contr
mitted, and how
l surroundingsne might argueeing more satim our non-ran
the first attemy fatigue’ and As was the casgagement for d school, with e: d student acad
e didactic chas embraced ademic engage
h response gr
ut actually moat it increased
re relative to aengagement, h
student acaders, that conneational climate
2019
ell, 5
user rol in w the
s and e, for isfied ndom
mpt at they
se for these little
demic
anges active ment
roups
oving their
active igher
demic ction
e was
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
55
more important (Value Dimensions). Both student and educator perspectives are important to provide a holistic understanding of how the design of a space for learning impacts student outcomes. A high level of statistical significance has been reached across all schools, geographies participating, grade levels, demographics and user groups that YES the design of space makes a difference for student academic engagement levels at p < .0001. The knowledge gained across the level of cohorts we measured, and will continue to measure, impacts the way one should consider developing design solutions from the macro level to the micro level. The UEIF.v2 continues to ensure questions connect to this grounding framework. The 21st century learning goals, and teaching strategies to support them are well articulated, and through post-occupancy analysis we bridged the connection between design and performance. References Abonyi, J., & Feil, B. (2007). Cluster analysis for data mining and system identification. Boston, MA:
Birkhauser Basel. Anonymous. (2010). Working with NSSEdData: A facilitator’s guide. National Survey of Student Engagement.
Retrieved from http://www.nsse.iub.edu/links/facilitators_guide Anonymous. (2008, April). Implementation of post-occupancy evaluation: A potential tool for building asset
management and creating more productive, cost-effective and sustainable buildings at MSU (pp. 1–7). White Paper # C2P2AI, SPDC. Michigan State University.
French, J., Scott-Webber, L., Ferking, T., & Fulton, D. (2015). Early childhood space design: Nurturing every student’s potential (p. 20). EDspaces Conference. LA: education Market Association.
Hanington, B. M. (2010). Massachusetts Institute of Technology Design Issues, 26(3), 22. Retrieved August 13, 2019 from https://s3.amazonaws.com/academia.edu.documents/31490953/DESI2603_Hanington_pp018-pp026_v4.pdf?response-content-disposition=inline%3B%20filename%3DRelevant_and_Rigorous_Human-Centered_Res.pdf&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAIWOWYYGZ2Y53UL3A%2F20190813%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20190813T185400Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&X-Amz-Signature=73057dc18edb4d8951f04d7c8722947fae5f39e25bbe6020811d68c9eff48b56
Johnson, BR, Onwuegbuzie, AJ, & Turner, LA. (2007). Toward a definition of mixed methods research. Journal of Mixed Methods Research. Vol. 1, pp. 112–133. doi: 10.1177/1558689806298224.
Kilbourne, J., Scott-Webber, L., & Kapitula, L. R. (2017). An activity-permissible classroom: Impacts of an evidence-based design solution on student engagement and movement in an elementary school classroom. Children, Youth and Environments, 27(1), 112–134. https://doi.org/10.7721/chilyoutenvi.27.1.0112
Lever, J., Krzywinski, M., & Altman, N. (2017). Points of significance: Principal component analysis. Nature Methods, 14, 641–642. https://doi.org/10.1038/nmeth.4346
Murtagh, F., & Legendre, P. (2014). Ward’s hierarchical clustering method: Clustering criterion and agglomerative algorithm. Journal of Classification, 31(3), 274–295. https://doi.org/10.1007/s00357-014-9161-z
Preiser, W. F. E., Harvey, Z. R., & Edward, T. W. (1988). Post-Occupancy Evaluations. Van Nostrand Reinhold. Scott-Webber, L. (2000, December 10–13). Characteristics of a conceptual framework addressing the
environment of instructional space holistically: design, users, pedagogy and technology. In H. Thomas (Ed.), American Vocational Education Research Association Conference Proceedings (p. 54). New Orleans, LA.
Scott-Webber, L. (2004). In-sync—Environmental behavior research and the design of learning spaces. MI: The Society for College and University Planning. Retrieved from http://books.google.com/books/about/In_Sync.html?id=snRPAAAAMAAJ
Scott-Webber, L., Konyndyk, R., French, R., & French, J. (2018). Significant results: Space makes a difference for student academic engagement levels. European Scientific Journal, 14(16), 61. https://doi.org/10.19044/esj.2018.v14n16p61
Scott-Webber, L., Konyndyk, R., French, R., Lembke, J., & Kinney, T. (2017). Spatial design makes a difference in student academic engagement levels: A pilot study for grades 9–12. European Scientific Journal, 13(16), 5. https://doi.org/10.19044/esj.2017.v13n16p5
-
jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019
56
Scott-Webber, L., Marini, M., & Abraham, J. (Spring, 2000). Higher education classrooms fail to meet needs of faculty and students. Journal of Interior Design, 26(1), 16–34. https://doi.org/10.1111/j.1939-1668.2000.tb00356.x
Song, M. K., Lin, F. C., & Ward, S. E. (2013). Composite variables: When and how. US National Library of Medicine National Institutes of Health, 62(1), 45–49. https://doi.org/10.1097/NNR.0b013e3182741948
Tavakol, M., & Dennick, R. (2011), Making sense of Cronbach’s alpha. International Journal of Medical Education, 2, 53–55. https://doi.org/10.5116/ijme.4dfb.8dfd
Trochim, W. M. K. (2006). Correlation. WEB Center for Social Research Methods. Retrieved from https://socialresearchmethods.net/kb/statcorr.php
Trochim, W. M. K. (2006a). Convergent & Discriminant Validity. WEB Center for Social Research Methods. Retrieved from https://socialresearchmethods.net/kb/convdisc.php
Zimring, C. M., & Reizenstein, J. E. (1980). Post-Occupancy evaluation: An overview. Environment and Behavior, 12(4), 429–450. https://doi.org/10.1177/0013916580124002
Notes Research sponsored by DLR Group’s K12 Education Studio. With thanks to all of our respondents. Copyrights Copyright for this article is retained by the author, with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).