expectation, feedback and performance amongst … thu 1520 b1.pdf · 2016-12-16 · first year...
TRANSCRIPT
Australian International Education Conference 2007 www.idp.com/aiec 1
Expectation, Feedback and Performance Expectation, Feedback and Performance Amongst Students Speaking Amongst Students Speaking
Languages Other than English in Languages Other than English in First Year Human Biology Units in First Year Human Biology Units in
Western AustraliaWestern Australia
Kayty Plastow, Jan Meyer, Georgina Fyfe, Kayty Plastow, Jan Meyer, Georgina Fyfe, Sue Fyfe, Mel Sue Fyfe, Mel ZimanZiman
Background Background –– the projectthe project
►► Produce feedback for online tests in first year Human Produce feedback for online tests in first year Human BiologyBiology
►► Three WA universities (mainly Three WA universities (mainly BScBSc students) students) –– A, B, CA, B, C
►► Large (>2000 per year) and diverse student cohortLarge (>2000 per year) and diverse student cohort
►► Importance of feedback Importance of feedback –– meaningful, immediate, meaningful, immediate, individualindividual
►► Benefits of online feedbackBenefits of online feedback
Australian International Education Conference 2007 www.idp.com/aiec 2
The Project The Project -- to Produce Feedback for Online Tests & to Produce Feedback for Online Tests & Construct an Online Instrument of Reflective PracticeConstruct an Online Instrument of Reflective Practice
Stage 1 Stage 1 –– the studentsthe studentsWho were we writing feedback for?
How did they view feedback?
Experience of Feedback?
What did they do with feedback?
What feedback were they expecting?
15 Focus Groups
Glover (2004)-based Paper Questionnaire
Trialled on level 3 HB students
Reduced, Refined, Administered
AnalysedEnrolment Mode - 98% full-time
LOTE - 22% of students spoke a language other than English at home – 2/3rds exclusively so. 42 different languages other than English
LOTE & Bilingual at Home by Institution
27%
18%
11%0
150
300
450
UniA UniB UniC
Native EnglishLOTE & bilingual
Average Age by Language Status
17.5
18
18.5
19
19.5
20
20.5
Native English LOTE Bilingual
Aver
age
Age
(yea
rs +
/- s.
e)
a a
b ANOVA (Institution as blocks)
F=3.28, df 2 & 1068, p=0.038
Proportion LOTE and Bilingual Speakers by Institution
0
5
10
15
20
25
UniA (70) UniB (84) UniC(79)
Institution (average TER cutoff 2003-2007)
% o
f Res
pond
ents
LOTE
Bilingual
Distribution of Language Groups (%)
African
W European
Indian subcontinent
E European
Aramaic BaseOtherAsian
ChineseEnglish LOTE
Language GroupsLanguage Groups
Australian International Education Conference 2007 www.idp.com/aiec 3
Work & Language StatusWork & Language Status►► Fewer speaking LOTE at home in paid work (and more Fewer speaking LOTE at home in paid work (and more
declining to answer the question) (22% sample LOTE)declining to answer the question) (22% sample LOTE)
Krause et al 20051999 LOTE 37.5% (12.58 hrs/wk)
Engl 54.9% (12.64 hrs/wk)2004 LOTE 45.9% (11.46 hrs/wk)
Engl 58.1% (12.79 hrs/wk)
Work
0
10
20
30
40
50
60
70
80
No work Work Unanswered
Perc
enta
ge o
f Lan
guag
e G
roup
EnglishLOTE & bilingual
χ2= 85.10, 2 d.f, p < 0.001Answering vs not χ2=31.01, 1 d.f. p < 0.001 Even including not answering as working = 49% LOTE vs 71%
none 1-10hrs 11-20hrs>20hrs
ECU
UWACURTIN
020
40
60
80
100
120
140
160
Over 65% in paid work – of those working49% worked <10hrs/week41% work 11-20 hrs/week
9% (30% at UniA) worked >20hrs/ week
UniC
UniB
UniA
Hours of Paid Work by Language Status
02468
101214161820
ECU UWA CURTINA
v. H
rs W
orke
d pe
r Wee
k
English onlyLOTE onlyBiling with English
UniA UniB UniCF native English vs any LOTE=53.46,df 1&1065, p<0.001
English Av 9.3
LOTE 4.9 hrs
Language Group and ExpectationLanguage Group and Expectation
Distribution of Language Groups (%)
African
W European
Indian subcontinent
E European
Aramaic BaseOtherAsian
ChineseEnglish LOTE
Anticipated Mark by Language Group
60
62
64
66
68
70
72
74
76
78
80
82
Chi
nese
Engl
ish
Oth
er A
sian
Aram
aic
Base
Afric
an
E Eu
rope
an
W E
urop
ean
Indi
ansu
bcon
tinen
t
Ant
icip
ated
Mar
k (%
)
ANOVA (Institutions as blocks & Age as covariate)F=4.65, df 7 & 1060, p<0.001
International students have different International students have different learning experiences and expectations learning experiences and expectations ((RamburthRamburth and McCormick, 2001)and McCormick, 2001)
Australian International Education Conference 2007 www.idp.com/aiec 4
Factors influencing expectationFactors influencing expectation
Expected Mark by Level of Paid Work
64
65
66
67
68
69
70
71
72
none 1-10hrs 11-20hrs >20hrs
Hours per week
Expe
cted
Mar
k %
ANOVA (age as covariate) F LOTE =11.72, df1 &1066, p<0.001F Institution =14.69, df1 &1066, p<0.001
ANOVA (Institutions as blocks, Age as covariate)F Work =4.6, df 1 & 1036, p= 0.032F Language =6.17, df 1 & 1036, p= 0.013
Anticipated Mark by Age
65
66
67
68
69
70
71
72
73
74
16-18 19-21 >21
Age Category
Ant
icip
ated
Mar
k (%
)
Native English
LOTE or bilingual
Anticipated Course Mark by Work and Language Status
66
67
68
69
70
71
72
73
74
Nonworker Worker
Ant
icip
ated
Mar
k (%
)
LOTE & bilingualNative English
ANOVA (Institutions as blocks) F LOTE =8.36, df1 &1065, p = 0.004
Anticipated Mark by Institution
54
56
58
60
62
64
66
68
70
72
74
UniA (70) UniB (84) UniC(79)
Institution (av TER cutoff 2003-2007)
Ant
icip
ated
Mar
k (%
)Native EnglishLOTE or bilingual
LOTE students:
•Were more optimistic about value of feedback
•Valued most sorts of feedback more
•Valued explanations of correct answers significantly more than non-LOTE students
Attitudes to FeedbackAttitudes to Feedback
Australian International Education Conference 2007 www.idp.com/aiec 5
The Project The Project -- to Produce Feedback for Online Tests & to Produce Feedback for Online Tests & Construct an Online Instrument of Reflective PracticeConstruct an Online Instrument of Reflective PracticeStage 2 Stage 2 –– The feedbackThe feedback
Who were we writing feedback for?
How did they view feedback?
Experience of Feedback?
What did they do with feedback?
What feedback were they expecting?
What sort of feedback was needed?
What sort of errors were being made?
What was the point of our questions?
Wrote 6072 feedback comments for MCQ tests for modules at each institution
Focus Groups
Glover (2004)-based Paper Questionnaire
Trialled on level 3 students
Reduced, Refined, Administered
Analysed
Examined past MCQ responses with analyses of who made errors, why? discriminating power etc
Administered tests with feedback
Guiding Principles for FeedbackGuiding Principles for Feedback►► Address to Address to ‘‘frameworkframework’’ rather than specific level rather than specific level ––
““whywhy””, not , not ““whatwhat””►► Meaning of key terms in parentheses as Meaning of key terms in parentheses as
encountered, otherwise plain languageencountered, otherwise plain language►► Affirmation for correct responsesAffirmation for correct responses►► Where need to choose incorrect statement reWhere need to choose incorrect statement re--
iterate in affirmation (iterate in affirmation (““Yes! That statement was Yes! That statement was untrue.untrue.””) and as a general comment) and as a general comment
►► Personal voice for correct responses, Personal voice for correct responses, impersonal for pointing out error impersonal for pointing out error
►► Give mental images to remember by,Give mental images to remember by,avoid meaningless mnemonic aidsavoid meaningless mnemonic aids
►► 11--2 sentences2 sentences
Having correct and incorrect. Makes it very easy to misread what the question is asking especially if u glance at the question quickly.(UniB)
Excellent use of brackets within the feedback. (UniA)
Australian International Education Conference 2007 www.idp.com/aiec 6
The Project The Project -- to Produce Feedback for Online Tests & to Produce Feedback for Online Tests & Construct an Online Instrument of Reflective PracticeConstruct an Online Instrument of Reflective PracticeStage 3 Stage 3 –– The impactThe impact
Who were we writing feedback for?
How did they view feedback?
Experience of Feedback?
What did they do with feedback?
What feedback were they expecting?
What sort of feedback was needed?
What sort of errors were being made?
What was the point of our questions?
Wrote 1605 feedback (4467 more)
comments for MCQ tests for one module at each institution
Focus Groups
Glover (2004)-based Paper Questionnaire
Trialled on level 3 students
Reduced, Refined, Administered
Analysed
Examined past MCQ responses with analyses of who made errors, why? discriminating power etc
Administered tests with feedback
Was the feedback seen as helpful?
What did they learn from it?
Did it improve performance?
Online Reflective Practiceinstrument
Analysed Test/ Retest & Final Examination Marks for Effects of Test with Feedback
Directed to online survey
Test Evaluation
Reflective Practice Reflective Practice
Expectations
(approx 10 options plus “other”)
Problem Areas
Technique
Strategies
About 5 minutes with test evaluation
where do you find out feedback of your test technique? All I viewed was the results and information about why it was wrong/right etc. (UWA)
Australian International Education Conference 2007 www.idp.com/aiec 7
Response to Test Response to Test
Errors of Test Technique
0
5
10
15
20
25
30
35
40
45
50
[Non
e Give
n]
misrea
d
noprobs
timeman
age
toinc
orrect
leftuna
ns
LOTE & bilingual
Native English
Problem vs No Problem χ2 =1.33, df=1, p=0.248Misread vs Changed to Incorrect χ2 =4.90, df=1,
Helped me to understand questions I got right
0
10
20
30
40
50
60
70
80
90
100
agreement <agreement agreement <agreement
read right didn't read right%
of l
angu
age
x re
adin
g gr
oup
LOTE & bilingual
Native English
Overall χ2 =9.18, df=1, p=0.002Amongst non-readers χ2 =5.11, df=1,
No difference from Native English speakers in judgments that feedback helped them to understand
• what they got wrong (88.7% vs 87.5%, χ2<1, NS)
• other questions in similar area (64.5%, 63.9%, χ2<1, NS
helpfulness of reflective practice (64.5% vs57.7%, χ2<1, NS
Type Read by Language Background
0 10 20 30 40 50 60
scroll
right
w rong
all
Type
Rea
d
Percent of Language Category
Native English
LOTE & Bilingual
Language, Expectation & AchievementLanguage, Expectation & AchievementTest Mark by Language Status
50
52
54
56
58
60
62
LOTE & bilingual Native English
Test
Mar
k (%
)
Test mark
Age Adjusted Test Mark
ANOVA (Institutions as blocks, Age as covariate)F=0.32 NS
Perceived Accuracy of Test as Reflection of Knowledge by Language Status
0
10
20
30
40
50
60
70
LOTE & bilingual Native English
% o
f Lan
guag
e G
roup
accurateinaccurate
χ2 =4.82, df 1, p=0.028
Test Score in Relation to Expectations by Language Group
0
10
20
30
40
50
60
70
LOTE & bilingual Native English
% o
f Lan
guag
e G
roup
higherexpectedlower
χ2 =14.05, df 2, p<0.001
Expected Mark by Language Status
67
68
69
70
71
72
73
LOTE & bilingual Native English
Expe
cted
Mar
k (%
)
ANOVA (Institutions as blocks, Age as covariate)F=8.74, df1 &1068, p=0.003
Australian International Education Conference 2007 www.idp.com/aiec 8
Test mark and work statusTest mark and work statusTest Mark by Work Status
LOTE
& b
ilingu
al
Nat
ive
Eng
lish
59
60
61
54
56
58
55
57
Age
-Adjus
ted
Test
Mar
k (%
)
worker
SE of Difference of Means
nonworker
ANOVA
Work or not F 7.321, p= .008
Expectation F =3.178, p=.077
Interaction F =5.512, p=.02
workernonworker
Paid Work Status
80
75
70
65
60
55
50
Tes
t Sco
re (
per
cen
t)
LowerHigherExpected
Expectation
Expectation, Test Scores and Paid Work
Test an Accurate Reflection of Knowledge
0
10
20
30
40
50
60
70
LOTE & bilingual Native English
% o
f Wor
k-La
ngua
ge G
roup
nonw orker
w orker
InteractionInteraction
high
er th
anex
pect
ed
low
er th
anex
pect
ed
Nat
ive
Eng
lish
wor
ker
LOTE
& b
iling
wor
ker
01020304050607080
Language, Work, Expectation & Performance
Native EnglishworkerLOTE & bilingworker
Australian International Education Conference 2007 www.idp.com/aiec 9
Test Pace and BackgroundTest Pace and BackgroundTest pace and Performance
UniC S2 2006
01020304050607080
Test
Sco
re%
Test
Rel
ated
MC
Q
Oth
er M
CQ
Test
Rel
ated
Sho
rt A
nsw
ers
Oth
er S
hort
Answ
ers
Test
Sco
re%
Test
Rel
ated
MC
Q
Oth
er M
CQ
Test
Rel
ated
Sho
rt A
nsw
ers
Oth
er S
hort
Answ
ers
Faster Than Designed Slower Than Designed
Ave
rage
Sco
re (%
+/-
SE)
European NameNon-European Name
• Used family name as a proxy for language background, which could only be directly linked to outcomes for survey completers at smallest institution (“European” ≅ “Native English Speaker”)
• In Western Australia most likely error inclusion of native English speakers in “NonEuropean”
•Most non-Commonwealth European immigration 1948-1960s, minority of student exchanges•Most other migration post-1975, O/seas students
Test Pace and Background
0
10
20
30
40
50
60
70
Faster Than Designed Slower Than DesignedTest Pace
Perc
ent o
f Pac
e G
roup
European NameNon-Euro Name
n=148 n=74 n=19 n=29
χ2 = 12.27, 1df, p<.001
European Name 20min 47sec (41.6s/Q)NonEuropean 23min 48sec (47.6s/Q)F=9.8, 1&268, p=.002Those with Non-European names took longerMost “Slower than Design” Non-European names
Test Pace, Background and Exam Outcome
Test Pace>=1min/Q< 1min/Q
50
52
42
46
54
44
48
40
Average M
ark (percen
NonEuropean Name
Av. SE of Difference
European NameAcross Components Interaction term F=9.90, 1&1076df, p=.002
Students with NonStudents with Non--European Names Gain Less European Names Gain Less Advantage from FeedbackAdvantage from Feedback
Feedback Advantage and Population Background
Fail
30
Hig
hDis
t
40
Cre
dit
50
60
Dis
tinct
ion
Pas
s
70
Fee
dbac
k ar
ea M
CQ
mar
k fin
al e
xam
(%)
euro
noneuro
unknownSE of difference of means
Unbalanced ANOVA, Institutions as blocks Fnametype= 4.67, df1&1163, p=0.031
Feedback Area Advantage by Population Background
0
10
20
30
40
50
60
70
80
90
100
NONSTARTER INCOMPLETE COMPLETE
Feedback Test Completion
Feed
back
are
a fin
al M
CQ
/ Non
feed
back
A
rea
(%)
euro
noneuro
Advantage of feedback indicated by value of “completed” cf “nonstarter”
Significantly greater for those with European-derived surnames
Australian International Education Conference 2007 www.idp.com/aiec 10
SummarySummary
►► More likely to have lower marks than expected and to have difficMore likely to have lower marks than expected and to have difficulty ulty accepting lower marks as accurate reflections of their knowledgeaccepting lower marks as accurate reflections of their knowledge
NESB students more likely to receive marks lower than expected, NESB students more likely to receive marks lower than expected, to find to find workload too heavy and to fear failing. Less likely to have expeworkload too heavy and to fear failing. Less likely to have expectations ctations regarding university met (Krause et al., 2005)regarding university met (Krause et al., 2005)
►► Significant numbers of LOTE students are in paid work, which hasSignificant numbers of LOTE students are in paid work, which hasgreater impact upon their performancegreater impact upon their performance
International students more problems than domestic International students more problems than domestic LOTEsLOTEs (Krause et al., (Krause et al., 2005)2005)
►► Greater expectations of online resources and greater appreciatioGreater expectations of online resources and greater appreciationn-- disproportionate to what is delivered?disproportionate to what is delivered?
Greater reliance on online resources (Krause et al., 2005)Greater reliance on online resources (Krause et al., 2005)
ImplicationsImplications
►►LOTE students as a group have higher LOTE students as a group have higher expectations and are more likely to be expectations and are more likely to be disappointed in their marksdisappointed in their marks
Help with development realistic expectationsHelp with development realistic expectationsLOTE students in particular would benefit from LOTE students in particular would benefit from early assessments and feedbackearly assessments and feedbackMultiple assessment tasks (with feedback) Multiple assessment tasks (with feedback) important in first semesterimportant in first semesterChances for detection of need for intervention Chances for detection of need for intervention in pattern of use of online resourcesin pattern of use of online resources
Australian International Education Conference 2007 www.idp.com/aiec 11
AcknowledgementsAcknowledgements
►► Leith Sly, IT support Curtin UniversityLeith Sly, IT support Curtin University
►► Support for this study has been provided by The Support for this study has been provided by The Carrick Institute for Learning and Teaching in Carrick Institute for Learning and Teaching in Higher Education Ltd, an initiative of the Higher Education Ltd, an initiative of the Australian Government Department of Education, Australian Government Department of Education, Science and Training. Science and Training.
►► The views expressed in this presentation do not The views expressed in this presentation do not necessarily reflect the views of The Carrick necessarily reflect the views of The Carrick Institute for Learning and Teaching in Higher Institute for Learning and Teaching in Higher Education.Education.
The Project The Project -- to Produce Feedback for Online Tests & to Produce Feedback for Online Tests & Construct an Online Instrument of Reflective PracticeConstruct an Online Instrument of Reflective PracticeStage 1 – The students and feedback
Who were we writing feedback for?
How did they view feedback?
Experience of Feedback?
What did they do with feedback?
What feedback were they expecting?
15 Focus Groups
Glover (2004)-based Paper Questionnaire
Trialled on level 3 HB students
Reduced, Refined, Administered
Analysed
Personal, Authoritative[most dislike] “general feedback that doesn't directly co-incide with me it may not specifically apply to me and therefore may be useless” (CURTIN) “written feedback from your assessor” (CURTIN)
Positive[most dislike] “comments that make me feel stupid makes me don't wanna learn” (UWA)
Explanatory[most dislike] “just give the correct answer without any explanation or say nothing at all because I do not understand at all even though I have found the correct answer myself”(CURTIN)
Pertinent, Digestible“huge comments that don't directly relate to learning objective doesn't help learning”(CURTIN)
Only 30% thought they “always” or “mostly” received enough feedback
Rated feedback more useful with increased
•age •feedback experience•expected performance
(LOTE)(less employment)
Australian International Education Conference 2007 www.idp.com/aiec 12
The Project The Project -- to Produce Feedback for Online Tests & to Produce Feedback for Online Tests & Construct an Online Instrument of Reflective PracticeConstruct an Online Instrument of Reflective PracticeStage 2 Stage 2 –– The feedbackThe feedback
Who were we writing feedback for?
How did they view feedback?
Experience of Feedback?
What did they do with feedback?
What feedback were they expecting?
What sort of feedback was needed?
What sort of errors were being made?
What was the point of our questions?
Wrote 1605 feedback (4467 more)
comments for MCQ tests for one module at each institution
Focus Groups
Glover (2004)-based Paper Questionnaire
Trialled on level 3 students
Reduced, Refined, Administered
Analysed
Examined past MCQ responses with analyses of who made errors, why? discriminating power etc
Administered tests with feedback
Found
• the clutter of ages – cleaned up obvious duplicates & problems
• many more terminology-based questions than we thought we had
• fairly frequent ‘misreading’ errors
•most obviously correct option in a “choose the incorrect”question)
•“hypo”/ “hyper” confusions
Problem Area Problem Area --LanguageLanguage
hard to find the correct statement among all the jargon (UWA)
I would often scan through the options quickly without having an idea of the kind of answer I was looking for, thus leaving me confused about the options, especially if there's minor variations in wording. (UWA)
it used words i was unfamilar with and it confussed me and i lost cncentration all togethr(CURTIN)
10.5% of all students nominated “terminology”as a problem area highlighted by the test (commonest response)
We were told that the information we needed to know was in the lec notes, yet the way the questions were worded did not reflect this. (CURTIN)
Hard to remember some names and functions etc. (CURTIN)
just that some multiple choice answers were similar, it kinda confused me. therefore, i took a guess of the two (CURTIN)
Having different names for differrentparts is also confusing (ECU)
get put off reading long questions, lose sight of what their
trying to say (UWA)
I have trouble understanding the terminology. Need to study this more. (CURTIN)
Australian International Education Conference 2007 www.idp.com/aiec 13
Initial QuestionnaireInitial Questionnaire►► DemographicsDemographics
Age, sex etcAge, sex etcPaid Work (hrs)Paid Work (hrs)LanguageLanguageAnticipated markAnticipated mark
►► Satisfaction with Satisfaction with past feedbackpast feedback
AmountAmountTimingTiming
►► Use of FeedbackUse of Feedback►► Anticipated Anticipated
FeedbackFeedback►► Helpfulness of Helpfulness of
feedback feedback
12 types of feedback
“Helpfulness of Feedback” section
Whole Questionnaire took about 15 minutes to complete, 1099 responses
Tests with FeedbackTests with Feedback
►► UniAUniA50 questions50 questions
No variationNo variationBlackboardBlackboard
►► UniBUniB & & UniCUniC30 questions30 questionsEach chosen Each chosen randomly from randomly from topic sets of topic sets of 5 5 --15 questions15 questionsWebCTWebCT
Which of the following statements is CORRECT?
Designed to take about 1 minute per question
Australian International Education Conference 2007 www.idp.com/aiec 14
Student Evaluation of FeedbackStudent Evaluation of Feedback
Helped Understand Wrong Answers
0 10 20 30 40 50 60
Strongly Disagree
Disagree
neutral
Agree
Strongly Agree
Agre
emen
t
Percent of Responses
Apply to Other Human Biology Modules
0 10 20 30 40 50 60
Strongly Disagree
Disagree
neutral
Agree
Strongly Agree
Agr
eem
ent
Percent of Responses
Length of Feedback Messages
0 10 20 30 40 50 60
Too Long
Too Short
Full
Brief
Just Right
Agr
eem
ent
Percent of Responses
They are a great help! Thanks for the time & effort! Definitely it would be great to have access to more similar feedback in the future, it is a fantastic study tool (UniB)the feedback comments really help me alot. all other biology units should have this system to help students to understand better.(UniC)
Found them helpful for the questions I got wrong as I could find out right there and then how I went wrong. Before, I had to try to remember what the question was to then be able to look it up in the book. Please keep them for next year. (UniC)
they helped a lot ,as the next time I did the test I got 48/50!! (UniA)
workernonworker
Workornot!
60
40
20
0
Cou
nt
Bar Chart
innaccurateaccurate
AccurateYN!
Nonworker Worker
20
0
80
40
100
60
Count
AccurateInaccurate
Whole Sample
workernonworker
Paid Work Status
64
62
60
58
56
54
52
50
Test
Sco
re (p
erce
nt)
InnaccurateAccurate
Accurate Evaluation?
Expectation, Test Performance and Paid Work
ANOVAworkF=7.321, 1&127df, p=.008
ANOVAexpectF=7.321, 1&127df, p=.077
ANOVAintF=7.5.572, 1&127df, p=.020
Perceived Accuracy of Perceived Accuracy of Assessment of UnderstandingAssessment of Understanding
►► Workers more likely to feel that their test results do not accurWorkers more likely to feel that their test results do not accurately ately reflect their understandingreflect their understanding
►► Those with lower marks more likely to consider inaccurateThose with lower marks more likely to consider inaccurate►► Workers who accept accuracy have lower scores than Workers who accept accuracy have lower scores than nonworkersnonworkers
UniC - Only one under test conditions
χ2 =4.833, 1df, p=.028
Australian International Education Conference 2007 www.idp.com/aiec 15
Feedback does not compensate for Feedback does not compensate for disadvantage of disadvantage of ““languagelanguage”” backgroundbackground
Fail
Pas
sC
redi
tD
istin
ctio
nH
ighD
ist
none
uro
euro
FU
LL T
ES
T
none
uro
euro
PA
RT
TES
T
non
euro
euro
NO
TE
ST
0102030405060708090
100
Fina
l Exa
m M
CQ
%
Final Grade
Feedback Area Performance Final Exam
Unbalanced ANOVA (institutions as blocks) Fnames = 10.41, df=1&1139, p<0.001; Ftests = 5.511, df=2&1139, p=0.004; Fnames.grades.test =2.90, df8&1139, p= 0.003