The Role of Personality Temperament and Student Learning in Principles of Economics:
Further Evidence
Author(s): Andrea L. Ziegert
Source: The Journal of Economic Education, Vol. 31, No. 4 (Autumn, 2000), pp. 307-322
Published by: Heldref Publications
Stable URL: http://www.jstor.org/stable/1183145
Accessed: 12/05/2009 11:55
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=held.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].
Heldref Publications is collaborating with JSTOR to digitize, preserve and extend access to The Journal of
Economic Education.
http://www.jstor.org
Research in Economic Education
In this section, the Journalof EconomicEducationpublishesoriginaltheoreticaland empiricalstudies of economiceducationdealing with the analysis and evaluationof teaching methods, learning, attitudesand interests,
materials,or processes.
PETER KENNEDY, Section Editor
The
Role
of
and
Student
Economics:
Personality
Learning
Further
in
Temperament
of
Principles
Evidence
AndreaL.Ziegert
Understandingeconomics is a process of gatheringinformation,makingsense
of information,building conceptualmodels, and using these models to evaluate
and analyze different situations and alternatives.Jungianpsychological theory
suggests that different personality temperamentsprefer to receive and process
informationdifferently.My hypothesis is that personalitytype plays an important role in a student's understandingof economics, as measuredby standard
measures of student performance.More precisely, given the more analytical
nature of economics, some personalitytypes may succeed in economics more
readily than do others.Furthermore,faculty personalitytype is also importantin
determiningthe way in which economic concepts are presented.Thus if students
and faculty have the same personalitytype, they are more likely to receive and
process informationin a similar fashion, which may lead to increased performance in economics courses. Understandingthe ways in which differentpersonality temperamentsgather and process informationcan lead to more effective
pedagogies that will benefit all economics students.
Borg and Shapiro (1996) found that studentpersonalitytype is an important
determinantof studentperformancein economics. In this article, I improve and
extend their work in two importantways. I replicateBorg and Shapiro'sanalysis
with a largersampleof studentsand a larger,more diversegroupof faculty.I also
Andrea L. Ziegert is an associate professor of economics at Denison University (e-mail:
[email protected]).The author wishes to thank membersof Miami University'sDepartmentof
Economicsfor theirparticipationin this project and Dennis H. Sullivanand anonymousrefereesfor
their helpful comments.
Fall 2000
307
investigate the role of personalitytype on both a student's knowledge of economics (stock) and on learningin an economics course (flow) using alternative
measuresof studentperformancein economics.
PERSONALITY TEMPERAMENTS AND LEARNING
Preference Scales
The role of personalitytype in learningwas developedby C. G. Jungandmade
operational by Isabel Briggs Myers and KatherineBriggs in the well-known
Myers-Briggs(personality)TypeIndicator(MBTI)(Myers 1975). Althoughonly
Borg and Shapiro(1996) have applied this informationto economics courses in
particular,considerableresearchhas used the MBTI to examine the role of personality in learning in general. In addition,the MBTI is frequentlystudied and
used in business settings to improveteamworkand workerproductivity.Because
many economists are unawareof this research,a brief discussion of personality
types and learningin an economics classroom is in order.
Myers and Briggs classified all mentalactivity accordingto four dichotomous
preference scales: Introversion(I)-Extraversion (E), Sensing (S)-Intuitive
(N), Thinking(T)-Feeling (F), andJudging(J)-Perceiving (P). The preference
scales identify and describe an individual'spreferencesfor relatingto the world
(I-E), processing information(S-N), makingdecisions (T-F), and lifestyle and
time orientation(J-P). Each scale is a continuum:An individualcan exhibit different degrees of these personalitytraits. In addition, the more a person reinforces a particularpreference,or that preferenceis reinforcedby teachers and
other influentialindividuals,the strongerthose characteristicswill be in the person's personality.Each individualcan be describedby his or herpersonalitytemperamentor type, which is composed of one preferencefrom each of the four
preferencescales.
Introversion (I)-Extraversion
(E)
The most misunderstoodof the preferencescales, introversion-extraversion
does not describean individualas eithera wallfloweror the life of the party.This
preferencescale refers to the focus of one's attentionor source of energy.Individuals with a preference for extraversion (E) are "outwardturning"and are
energizedby the world outside of themselves.They are sociable, action oriented,
risk takers,and have a wide varietyof interests.In the classroom,extravertstudents respondquickly to questions and often need to talk out loud to know what
they think. Extravertscan handle multiplestimuli withoutlosing focus and generally prefer talking ratherthan writing as a method of communication.They
have a wide varietyof interests.About 75 percentof the U.S. populationareestimated to be extraverts,and the remaining25 percentare introverts.
On the other hand, individualswho prefer introversion(I) are inwardturning
and find the source of their energy in the inner life of ideas and concepts. They
are private,reflective, and often exhibit a depthratherthana breadthof interests.
308
JOURNALOF ECONOMICEDUCATION
In the classroom, students with this personalitypreference need time to get to
know their instructorsand classmates;they need time for reflection before discussion and preferone-to-one ratherthan groupactivities. If introvertsare silent
in class, it does not necessarilymean they are uninterested,but ratherthey need
time to reflect upon the informationand place it in context before discussing it
with other students. Introvertswill often express an opinion only when asked.
Introvertspreferwritingratherthan talkingas a mode of communication.
At the end of the teachingday, an extravertedprofessoris likely to be wound
up andenergizedby teaching.He or she often has difficultyunwindingand wants
to talk about and process what happenedin the classroom. A professor with a
preferencefor introversion,althoughoften energeticin the classroom,is drained
at the end of the day. Bringing out the connections in the classroom is often
exhausting,and an introvertedprofessorneeds solitude to reenergizeand process
what happenedin class. Researchsuggests thatnationwide,introvertsoutnumber
extravertson college and universityfaculties (Centerfor Applicationof Psychological Type 1993). This may partiallyexplain the prevalenceof passive rather
than active pedagogies in college classrooms, as professors teach in a manner
more consistent with their own personalitypreferences or, perhaps, their own
experiencesas students.
Sensing (S)--Intuitive (N)
The sensing-intuitivepreferencescale refers to the way in which an individual takes in information.Approximately75 percentof the U.S. populationprefer
sensing (S) and the remainderare intuitive (N). A sensing individualtrusts the
data and informationthat comes from the five senses. These studentsfocus on
details and specifics, often work sequentially,preferexperience-rootedlearning,
and have a practicaland present orientation.In economics, these students may
focus on replicatinggraphsdrawnon the boardwithoutunderstandingthe underlying concepts (i.e., they know that MC crosses ATC at a minimum,but they do
not know why).
Intuitivespreferto take in informationthroughintuitionor hunches.They like
the "big picture,"focusing on patternsand concepts first,ratherthen detail.They
learn from insight and enjoy learning new skills. These students like solving
problems,enjoy change, and are often innovativein theirthinking,althoughthey
may frequentlymakeerrorsof fact. Studentswith an intuitivepreferencearelikely to get boredwith a professorwho reviews problemsin class thatillustrategeneral concepts using details because they find such activities obvious and repetitive.
In the classroom, a sensing professormay ask studentsfor specific examples
of economic phenomena(when you go to the store, how do you decide what to
buy?)and then develop a conceptor theoryin termsof a concreteexample or student experience.On the otherhand,an intuitiveprofessorwill often presenta theory or concept first and may be frustratedwhen some studentsfail to see its relevance or generalizationto specific facts or circumstances.
Fall2000
309
Thinking (T)-Feeling (F)
Thinking does not imply intelligence nor does feeling imply emotion. This
preferencescale describes ways in which individualsevaluate informationand
make decisions. Thinking-feeling preferences are evenly divided among the
overall population:50 percent are thinking (T) and 50 percent are feeling (F).
Thinkers tend to make decisions impersonally,making objective judgments,
which emphasizeanalysis and place a high value on fairness.Emotionsand feelings are importantto thinkersbut are valued as only one of many aspects to be
considered in the decision. In the classroom, thinkers love competition and
debate and try to sway others to their position. The feeling person makes decisions on the basis of personalvalues and places a high value on harmony.Their
decisionmakingstyle is subjective and empathic.They like to give and receive
affirmation.They do not like debatebutprefercollaborationto competitionin the
classroom.' Feeling studentslearn best once they have developed a relationship
with both their instructorand studentcolleagues.
Thinking versus feeling professors differ most in their approachto student
evaluation.Thinkersare likely to have the gradedistributionfor the course on the
syllabus on the first day of class. With few exceptions, applying this pre-determined distributioneasily sets grades.A feeling professor,on the other hand, is
more likely to give more weight to individualstudentgrowthand circumstances.
Potential differences in student evaluation by thinking and feeling professors
may raise a new issue of fairness.If when evaluatingstudentperformance,a feeling professorplaces more emphasison the amountof studentlearningduringthe
term (value added)ratherthanperformancerelativeto fixed learningobjectives,
then students must understandthe rules for grading or the assessment may be
both inaccurateand unfair.
The thinking-feeling preferencescale is uniqueamong the personalitypreferences. It is the only preferencescale thathas significantdifferencesin the distribution of preferencesamong men and women. Approximately60 to 65 percent
of women tested were classified as feeling, whereas60 to 65 percentof men tested were thinking(Centerfor Applicationof PsychologicalType 1993).2
Judging (J)-Perceiving (P)
The final preferencescale, judging-perceiving, indicatespreferencesin relating to the outer world: an individual's need for order, closure, openness and
adaptability.This preferencescale is evenly distributedamongthe population:50
percent of the populationare classified as perceiving (P) and the remainderare
judging (J). Judging individualsprefer an orderly and structuredenvironment.
They are goal oriented and often have detailed plans for achieving their goals.
When a decision is needed, they areoften anxious untila decision is made-they
seek closure. In the classroom, these studentsoften complete papersand assignments before their due dates and are very organizedin their studying.They are
often upset, however,if scheduledtests or assignmentdeadlines are changed.
Perceiving individuals prefer a spontaneousand flexible environment.They
310
JOURNAL
OFECONOMIC
EDUCATION
like to keep their options open and are often tentativein their decisionmaking:
they are uncomfortablemakingdecisions quickly and often will postponedecisions until the last possible momentto ensurethatall possible options have been
explored. These students often pull "all-nighters"to complete requiredassignments. Imposing structureon these studentsmay actually be counterproductive,
because they requirea degree of flexibility to accomplish their tasks. This preference has importantimplicationsfor active learning or group activities in the
classroom. Combining J and P students on group projects can spell disaster
unless the J students'need for structureand the P students'need for flexibility
are discussed and some compromisereached.
A professor'soffice often betrayshis or her identificationon this preference
scale. The judging individual'sneed for orderoften translatesinto orderlyfiles
and organized,clean desks: Everythinghas a place and everythingis in its place.
On the other hand, a non-P observer may consider a perceiving individual's
office a disaster area: books, papers, and files are stacked everywhere,and the
desk surfacehas not been seen since the beginningof the term. Generally,however, P individualsknow where things are and can operatesuccessfully in such
an environment.
When the four preferencescales are combinedin all possible combinations,16
personality types result. These possible personality types and their estimated
incidence in the generalpopulationare listed in the last column of Table 1.
TABLE 1
Comparison of the Sample Distribution of Personality Type
of Faculty and Students with the U.S. Population
Students
Number
Faculty
U.S. population
(in percentages)
Percentage
Number
Percentage
ENFJ
ENFP
ENTJ
ENTP
ESFJ
ESFP
ESTJ
ESTP
18
9.1
0.0
0.0
9.1
0.0
18.2
0.0
2
1
0
0
1
0
2
0
11.8
8.1
3.5
2.1
20.9
3.5
21.3
1.3
73
50
22
13
129
22
131
8
5
5
5
5
13
13
13
13
INFJ
INFP
INTJ
INTP
ISFJ
ISFP
ISTJ
ISTP
0.0
9.1
27.3
0.0
0.0
0.0
9.1
0.0
0
1
3
0
0
0
1
0
2.9
2.1
1.5
0.7
6.9
0.9
11.2
0.0
18
13
9
4
43
6
69
0
1
1
1
1
6
5
6
7
Personalitytype
Source: For U.S. populationwas Keirseyand Barnes(1984, appendix).
Note. See text p. 308 for a descriptionof personalitytype.
Fall2000
311
LEARNING THEORY, TEMPERAMENTS, AND ECONOMICS
Previous research highlights the importanceof both learning and teaching
styles and preferences. Research using the Grasha-ReichmannLearning Style
Questionnaire(Reichmannand Grasha 1974) suggests that similarity between
faculty teaching styles and studentlearning styles improvesa student'sattitude
towardand success in economics (Wetzel, Potter,and O'Toole 1982; Charkins,
O'Toole, and Wetzel 1985). Borg and Shapiro (1996) extend this literatureby
Theirresultssuggest that studentswhose
focusing on personalitytemperaments.3
personalitytemperamentsmatchthatof theirprofessorsearnhighergradesin economics. In addition,introvertedstudentswere found to do betterin principlesof
economics thandid extravertedstudents.Furthermore,as noted by these authors,
studentgrade, the dependentvariableused in Borg and Shapiro'smodels, is not
necessarilya consistentmeasureof performanceacrossdifferentprinciplesclasses because individualprofessorsmay have differentgradingscales. Indeed,professors with differentpersonalitytypes may approachgradingvery differently.
IsabelMyersconsideredthe S-N andT-F preferencescales the most important
factors in groupingthe preferences(Myers and Myers 1991). These preferences
relate directlyto the learningprocess:the S-N preferenceinvolves the gathering
of information,and the T-F preferenceinvolves decisionmakingand problem
solving. Given the possible combinationsof these preferences(SF, ST, NF, NT),
this researchhypothesizesthat,ceterisparibus,the moretheoreticaland technical
characteristicsof microeconomicprincipleswill lead NT (intuitive-thinking)students to performbetterand SF (sensing-feeling) studentsto performless well in
principles of economics. This hypothesis has underlyinggender implications:
becausefemale studentshave a greaterprobabilityof being an F (60 to 65 percent
versus 35 to 40 percentfor males) than a T; women on averageare expected to
performless well thantheirmale counterparts,all else equal.
DATAAND STUDY DESIGN
The Data
In this study,microeconomicsprincipleswere taughtby full-timefaculty,with
Ph.D. degrees in economics, in classes of 35 or fewer studentsduringa 15-week
semester.Participationin the study was voluntary:11 of 18 facultymembersrepresenting 27 of 36 classes of micro principlesparticipated.Studentparticipation
was also voluntary.In all, 617 studentsof a possible 1,200 studentsparticipated
in the study.4Of the students in this sample, 400 studentstook the microeconomics TUCE (Test of Understandingof College Economics) (Saunders 1991)
exam at the beginning of the semester (PreTUCE),and 300 students took this
exam at the end of the semester (PostTUCE).5These differentrelative sample
sizes suggest a possible sample selection bias discussed by Becker (1997). However, characteristicsof participatingand nonparticipatingstudents suggest that
studentswho participatedin this projectwere not appreciablydifferentfrom nonparticipants.On the other hand, participatingfaculty were racially more diverse
312
JOURNAL
OFECONOMIC
EDUCATION
and, on average,had fewer years teaching experience than nonparticipatingfaculty.6In addition,the Keirsey Sorter,a 70-question, Myers-Briggs-Typeinstrument, was administeredduringthe semester to both studentsand faculty.7Student GPA,ACT scores, cumulativecredit-hourspassed previousto this semester,
total hours attemptedduring the semester of the study, and grade in microeconomics were also obtained from universitydata sources. Personalitytemperament types for participatingfaculty and studentsare found in Table 1. Students
represented15 of the 16 temperamenttypes, whereas the 11 faculty members
represented7 differentpersonalitytypes. Recall that Myers consideredS-N and
T-F preferencescales crucial for the learningprocess. Distributionsof the S-N
and T-F preferencescales for the sample of participatingstudentsand faculty
suggestedthatover half of the faculty were T, whereas43 percentof all students
(and only 32 percent of female students) were representedby this preference.
Nearly two-thirdsof the faculty and only one-thirdof the studentstested as N.
Study Design
The theoreticalmodel used to estimate the effect of personalitytemperament
on learningin economics is the familiareducationalproductionfunction
Output= F(studentabilities,studentdemographics,personalitytemperament),
where outputin economics is measuredalternativelyas course grade(GRADE),
raw score on the PostTUCEexam (PostTUCE),or the differencebetween PostTUCE and PreTUCEscores (PostTUCEminus PreTUCE).Studentabilities are
measuredby gradepoint average(GPA) andACT scores. Studenttime available
for studyingmicroeconomicprinciplesis proxied by numberof credit hoursthe
studentwas enrolled in duringthe semesterof the study (HRSTAKEN).Student
demographicvariablesinclude total credit hours earned (TOTALHOURS),sex
(MALE), race (NONWHITE),and various personality temperamentvariables.
Course attributes,as proxied by a series of faculty fixed-effects variablesrepresenting 10 of 11 participatingfaculty,acted as shift parametersin much the same
way as technology does in the more standardproductionfunction.In addition,I
includedseveraldifferentvariablesto test the Borg and Shapiro(1996) hypothesis that studentswhose personalitytype matchesthat of their instructorperform
betterin microeconomicprinciples.
EMPIRICAL MODEL AND RESULTS
The empiricalmodel is of the following form:
Output= F(GPA,ACT,HRSTAKEN,TOTALHOURS,MALE,
NONWHITE,TEMPER,MATCH),
where the dependentvariableis alternativelymeasuredas
GRADE = the lettergrade (A, B, C, D, F) received in economic principles,
with A = 4 ..., F = 0;
Fall2000
313
PreTUCE= score earnedon the TUCE exam takenduringthe first week of
class as measuredby numberof questions answeredcorrectly;
PostTUCE= score earnedon the TUCE exam takenduringthe last week
of class as measuredby numberof questionsanswered
correctly;
LEARN = PostTUCEscore minus PreTUCEscore.
The independentvariablesare measuredas
GPA = studentgradepoint average(0.00 to 4.00);
ACT = score on the ACT exam or its SAT equivalent;
HRSTAKEN= total credit hourstakenin the semesterof the study;
TOTALHOURS= total credit hourspassed previousto the semesterof the
study;
MALE = 1 if studentwas male; 0 if female;
NONWHITE= 1 if studentwas nonwhite;0 if white;
TEMPER= dummy variablesindicatingthe 16 temperamenttypes; or,
alternatively,variouscombinationsof two of the four major
temperamentpreferencescales;
TEMPMATCH= 1 if studentpersonalitytemperamentmatchedthatof their
professor;0 otherwise (variousspecifications);
PROFI-PROF10= a series of faculty fixed-effects variablesrepresenting10
of 11 participatingfaculty members.
A priori, studentaptitudeand abilities were predictedto have a positive impact
on TUCE exam scores or course grades: students with higher GPAs and ACT
scores were expected to performbetter in economics. Similarly, students with
more cumulative hours (TOTALHOURS)were predictedto performbetter.If
hours taken (HRSTAKEN)is indeed a proxy for available study time, students
carryingmore credit hourswere predictedto earnlower gradesthandid students
carryingfewer credit hours.As previouslydiscussed, SF personalitytypes were
predictedto performless well thantheirNT counterparts;however,studentswho
matchedtheir professor'spersonalitytype were expectedto do betterregardless
of personalitytype.
Estimationof these models requiredseparateestimationtechniquesfor each of
the possible proxies of student learning in economics. Ordinaryleast squares
(OLS) was used to estimate the numberof correctanswerson the PreTUCEand
PostTUCEexams, and an orderedprobit maximumlikelihood estimationtechnique was used to estimate course grade. The sample of students used in this
study did not include any ISTP personalitytypes; thus, only 15 of the 16 temperamentsare representedin the sample.To avoid perfect multicollinearitywith
the remainingtypes, I omittedISTJfrom the model. This temperamenttype was
eliminated to facilitate comparison with Borg and Shapiro's empirical results.
Significantnegativecoefficients on the remainingpersonalitytemperamentssug314
JOURNAL
OFECONOMIC
EDUCATION
TABLE2
IndependentVariable,SampleMeans,and StandardDeviationsforVariables
Used withStudentGrade(N = 617)
Variable
GPA
ACT
HRSTAKEN
TOTALHOURS
MALE
ENFJ
ENFP
ENTJ
ENTP
ESFJ
ESFP
ESTJ
ESTP
M
S.D.
INFJ
INFP
INTJ
INTP
ISFJ
ISFP
0.030
0.021
0.016
0.008
0.072
0.009
0.17
0.14
0.12
0.08
0.26
0.09
SF
ST
NF
I
S
T
J
0.322
0.343
0.249
0.271
0.665
0.428
0.807
0.46
0.47
0.43
0.44
0.47
0.49
0.39
M
S.D.
Variable
2.95
25.8
15.14
37.83
0.54
0.113
0.63
2.89
2.17
14.26
0.49
0.31
0.084
0.042
0.019
0.203
0.035
0.215
0.014
0.27
0.20
0.13
0.40
0.18
0.41
0.11
gested thatthese temperamentsperformedless well in microeconomicprinciples
than did the omitted categories, whereas significant positive coefficients suggested the remainingtemperamentsdid better.Descriptive statistics for each of
the variablesused in the analysis are shown in Table 2.8
Estimation Strategy
The opportunityto replicate a pedagogical experimentin economics is rare
indeed. Thus, for comparativepurposes, this study closely replicatedBorg and
Shapiro's(1996) analysisusing a larger,morediversesampleof studentsand faculty (Tables 3-5). Noting the deficiencies of using course grade as a measureof
studentperformance,I estimatedothermeasuresof studentknowledgeand learning in economics and also tested alternativehypothesesof the role of personality temperamentand learningin economics (Tables6-8).9
The sample analyzedby Borg and Shapiroincluded 119 studentsin 3 sections
of macroeconomic principles compared with 617 students in 27 sections of
microeconomic principles in my study. The two empirical specifications estimated were nearly identical with only minor differences:the variablefor race
(Black = 1, zero otherwise,versusNonwhite = 1, zero otherwise);the absenceof
a junior college transfer(JCTRANS)variablein my study;and differentproxies
for studentaptitudes[SAT versus ACT, and high school GPA (HSGPA)versus
college GPA(GPA)].In addition,both models omittedthe ISTJpersonalitytemperament,but my sample containedno studentswith the ISTP personalitytype.
The role of the 16 personalitytemperamentsin determininga student'scourse
grade is shown in Table 3. The separate effect of each personalitypreference
scale on gradeearnedis in Table4. The role of particularlearningstyles (SJ, NF,
NT, and SP) on gradeearnedin principlesof economics is in Table5.
Re-estimating Borg and Shapiro's models yielded conclusions similar to
Fall 2000
315
TABLE3
and GradeEarnedin EconomicPrinciples:
PersonalityTemperaments
A ComparisonwithBorgandShapiro(ISTJomitted)
Dependentvariable:CourseGrade
Borg and Shapiro
Ziegert
Independent
variable
Estimate
CONSTANT
NONWHITE
-9.274***
0.0743
Independent
variable
Estimate
CONSTANT
BLACK
JCTRANS
SAT
HSGPA
TOTALHOURS
ESFP
ENTP
ESTP
ENFP
-3.966***
-0.543
-0.453
0.004***
0.230
0.009*
-1.536
-1.353***
-1.328**
-1.138***
ACT
GPA
TOTALHOURS
ESFP
ENTP
ESTP
ENFP
0.099***
0.861***
0.002
-0.423
-0.721**
0.294
-0.248
INFJ
INTP
ENFJ
INTJ
ESFJ
INFP
ISFJ
ESTJ
ISFP
ENTJ
ISTP
-1.019
-0.668
-0.625
-0.613
-0.571
-0.388
-0.202
-0.036
0.214
0.667
0.743
INFJ
INTP
ENFJ
INTJ
ESFJ
INFP
ISFJ
ESTJ
ISFP
ENTJ
ISTP
-0.409
0.375
-0.2679
-0.253
-0.325**
-0.807***
-0.011
0.030
-0.161
-0.551"*
TEMPMATCH
PROF 1
PROF2
X2
df
N
0.578**
1.036***
0.678***
-
TEMPMATCH
PROF 1-10
71.3
23
119
0.064
70.8
31
617
at the.10 level;**significant
atthe.05 level;***significant
atthe.01 level.
*Significant
theirs. Studentaptitude,as measuredby either SAT or ACT scores and college
GPA, had a positive impacton gradein economics. For both samplesof students
analyzed, studentswho were extraverts(Ziegert:ENTP,ESFJ, and ENTJ;Borg
and Shapiro:ENTP,ESTP and ENFP) performedless well than their ISTJ student colleagues (Table 3). Sensing (S) studentsoutperformedintuitive (N) students, and thinkers(T) outperformedfeelers (F) (Ziegert);and introvertedstudents (I) outperformedextraverts(E) (Borg and Shapiro) (Table 4). For both
samples analyzed,SJ learnersearnedsignificantlyhighergradesthandid NT and
NF learners(Table5).
There were some differences.The coefficient on TOTALHOURSwas positive
for both samples, but only Borg and Shapirofound this result to be statistically
significant. In my sample, the majorityof students who enrolled in economic
principles courses were sophomores, so this result was not surprising.More
316
JOURNAL
OFECONOMIC
EDUCATION
TABLE 4
Personality Preference Scales and Grade Earned in Economic Principles:
A Comparison with Borg and Shapiro
variable:
CourseGrade
Dependent
and
Borg Shapiro
Independent
variable
Estimate
CONSTANT
BLACK
JCTRANS
SAT
HSGPA
TOTALHOURS
-4.825***
-0.610
-0.554
0.004***
0.179
0.012**
0.357*
0.403
0.110
0.256
0.565**
1.067***
0.688**
I (vs. E)
S (vs. N)
T (vs. F)
J (vs. P)
TEMPMATCH
PROF 1
PROF2
Ziegert
Independent
variable
Estimate
CONSTANT
NONWHITE
-6.00**
-0.008
ACT
GPA
TOTALHOURS
0.099***
0.845***
0.002
1 (vs. E)
S (vs. N)
T (vs. F)
J (vs. P)
TEMPMATCH
PROF 1-10
0.062
0.187*
0.166*
0.027
0.054
58.871
12
119
X2
df
N
71.6
19
617
at the.01 level.
at the.10 level;**significant
at the.05 level;***significant
*Significant
TABLE 5
Combinations of S-N, T-F, and J-P Personality Preferences and Grade Earned
in Economic Principles: A Comparison with Borg and Shapiro (SJ omitted)
Dependentvariable:Course Grade
Borg and Shapiro
Independent
variable
Estimate
CONSTANT
BLACK
JCTRANS
SAT
HSGPA
TOTALHOURS
-4.018**
-0.544
-0.669*
0.004***
0.183
0.013***
NF
NT
SP
TEMPMATCH
PROF 1
PROF2
-0.654**
-0.633**
-0.135
0.528**
1.030***
0.575*
X2
54
11
df
Ziegert
Independent
variable
Estimate
CONSTANT
NONWHITE
-6.721***
-0.019
ACT
GPA
TOTALHOURS
NF
NT
SP
TEMPMATCH
PROF 1-10
0.103***
0.841***
0.001
0.215**
-0.347**
-0.115
0.023
71
19
at the.10 level;**significant
at the.05 level;***significant
at the.01 level.
*Significant
Fall 2000
317
TABLE 6
Personality Temperament and Scores Earned on the TUCE Exam:
(ISTJ, ISTP omitted)
Independent
variable
Dependentvariable:
PostTUCEscore
Estimated
coefficient
t
PostTUCEminus
preTUCEscore
Estimated
t
coefficient
CONSTANT
GPA
ACT
ESFP
ENTP
ESTP
ENFP
INFJ
-3.59**
1.27***
0.58***
-2.78**
-1.22
-.132
-1.89**
-2.05*
1.88
3.52
8.53
2.26
0.81
0.68
2.19
1.78
-6.79***
1.11***
0.33***
-2.66*
-0.33
-1.89
-3.00***
-0.81
3.12
2.68
4.30
1.79
0.21
0.94
3.00
0.64
INTP
ENFJ
INTJ
ESFJ
INFP
ISFJ
ESTJ
ISFP
ENTJ
-2.20
-2.10***
2.97**
-3.17***
-2.83*
-2.57***
-1.18*
-2.60
-0.18
0.82
2.67
2.11
4.89
1.77
3.10
1.84
1.35
0.17
-3.45
-2.92***
2.51*
-2.75***
-4.64**
-3.45***
-1.46**
-2.22
-0.31
1.24
3.34
1.16
3.53
2.76
3.68
2.00
1.10
0.26
PROF 1-PROF 10
0.37
400
adj. R2
N
0.26
329
at the.10 level;**significant
at the.05 level;***significant
at the.01 level.
*Significant
TABLE 7
Personality Preference Scales and Scores Earned on the TUCE Exam
Independent
variable
Dependentvariable:
PostTUCEscore
Estimated
coefficient
t
PostTUCEminus
preTUCEscore
Estimated
coefficient
CONSTANT
GPA
ACT
-6.33***
1.26***
0.59***
3.57
3.51
8.88
-9.76***
1.19***
0.32***
4.86
2.89
4.18
I (vs. E)
S (vs. N)
T (vs. F)
J (vs. P)
0.81"**
-1.03**
2.14***
0.20
1.98
2.30
5.48
0.35
0.72
-0.74
2.14***
0.62
1.55
1.46
4.83
0.99
PROF 1-PROF 10
adj. R'
N
0.37
400
0.26
329
at the.10 level;**significant
at the.05 level;***significant
at the.01 level.
*Significant
318
JOURNALOF ECONOMICEDUCATION
TABLE8
Combinationsof S-N andT-F PersonalityPreferencesand ScoresEarned
on the TUCEExam(NT omitted)
PostTUCEminus
preTUCEscore
Dependentvariable:
PostTUCEscore
Estimated
Independent
variable
Estimated
coefficient
t
CONSTANT
GPA
ACT
ST
SF
NF
-4.05***
1.33***
0.60***
-0.98**
-3.20**
-2.29***
2.18
3.77
8.88
1.39
4.42
3.05
coefficient
-7.14***
1.28***
0.33***
-1.12
-3.07***
-2.88***
3.39
3.14
4.30
1.42
3.80
3.49
PROF 1-PROF 10
adj. R2
n
0.36
400
0.28
329
at the.05 level;***significant
at the.01 level.
at the.10 level;**significant
*Significant
important,although Borg and Shapiro found supportfor the hypothesis that a
match between student and faculty personality type enhances performancein
principlesof economics, my study did not.1o
Estimations of alternativeperformancemeasures in principles of microeconomics are given in Tables 6-8. The first estimation, the numberof questions
answeredcorrectlyon the PostTUCEexam, is a measureof a student'sstock of
economic knowledge at the end of the term. The second, the numberof additional TUCE questions answered correctly at the end of the course compared
with those at the beginning (PostTUCE minus PreTUCE),is a measure of the
flow of studentlearningduringthe course.The TUCE score is a more preciseand
consistent measure of student performancethan course grade when analyzing
studentoutcomes across variousinstructorsbecause it avoids instructor-specific
differencesin grading.In addition,the models presentedin Tables6-8 are "tested down"versionsof the empiricalmodel previouslydeveloped."IVariablessuch
as MALE, NONWHITE, HRSTAKEN, TOTALHOURS,and TEMPMATCH
were excluded from these models because they were not statisticallysignificant
predictors(at the 0.10 level) of student performancein microeconomicprinciples. I investigatethe role of the 16 personalitytemperamentsin Table6,2 I1look
at the separateeffects of each personalitypreferencescale in Table7, and I investigate the role of the N-S and T-F personalitypreferencescales in Table 8.13
In each of the six models estimated,studentaptitudesas proxied by GPA and
ACT had a significantpositive effect on TUCE scores. The data in Table 6 suggest that eight personalitytypes, predominatelyextraverts(E) and feeling (F)
types (ESFP,ENFP, ENFJ, ESFJ, ESTJ, INFJ, INFP, and ISFJ) performedsignificantly less well on the TUCE exams than did ISTJ students, whereas INTJ
studentsperformedsignificantly better.All other personalitytemperamentsdid
not performsignificantlydifferenton the TUCE exam than did the omitted perFall 2000
319
sonality type, ISTJ.Conclusions based on the data in Table 7 supportthat introverts (I) performbetterthan do extraverts(E) and thinking(T) studentsoutperform feeling (F) students on the TUCE exams. Furthermore,intuitive students
(N) performbetterthando sensing students(S) on the PostTUCEexam. The data
in Table 8 provideevidence that NT studentsoutperformSF and NF studentson
the PostTUCEand PostTUCE-minus-PreTUCE
models and also outperformST
studentson the PostTUCEexam.
CONCLUSIONS
These results suggest that personalitytypes do affect studentperformancein
economics whethermeasuredby course grade or by performanceon the TUCE
exam. Unlike previousresearch,I found supportfor the hypothesesthatthinking
(T) students outperformfeeling (F) students on both course grades and TUCE
exams. Given that men are more likely to be classified as thinkers(T) than are
women, and women are more likely to be classified as feelers (F) than are men,
this result has implicationsfor the relative performanceof men and women in
principlesof economics. After adjustingfor personalitypreferences,the "gender
gap"in economics, which historicallyhas favoredmen, disappears:genderis not
a statistically significant predictorof performancein economics whether measured by course grade or performanceon the TUCE exam.
In addition,intuitive(N) studentsperformbetterthan do sensing (S) students
on the TUCE exams, and sensing (S) studentsearn highercourse grades thando
intuitive(N) students.This suggests thatdifferentpersonalitytypes may have different preferredmethods of testing or alternatively,course grade may not be a
sufficiently precise measure of student performanceto evaluate accuratelythe
role of personalitytemperamentsand learningin economics. To the extent that
standardizedtests such as the TUCE are writtenby economists who may have a
disposition for intuition(N) ratherthan a sensing (S) preference,14these instruments may be biased in favorof intuitive(N) students.Furthermore,the fact that
intuitive students (N) outperformsensing (S) students may suggest a potential
mismatch of learning and teaching styles. If the majorityof students have the
sensing preferenceand thus preferto learncontextually,but principlesof microeconomics tend to be presented"theoryfirst,"then sensing studentsmay be at a
disadvantagerelativeto intuitivestudents.
Finally,unlikepreviouswork,this researchdid not find supportfor the hypothesis that matching student and faculty personalitytemperamentsenhances student performancein economics. This difference may be partiallyexplained by
differencesin the two samplesof studentsstudied.The Universityof NorthFlorida (UNF), site of the Borg and Shapirostudy is a commutercampus where economics is taught in classes of 35 to 50 students.The average SAT score of the
UNF sample of studentsstudied was 987. On the other hand,Miami University
(MU) is a strictlyresidentialcampus,and economics is taughtin classes of 35 or
fewer students. The average SAT equivalentof the MU students sampled was
1,030-1,060. Thus, smaller class size and slightly higher aptitudesmay allow
students to adapt better to personality-baseddifferences in teaching styles that
320
JOURNAL
OFECONOMIC
EDUCATION
are different from their own personality preferences.Alternatively,the larger
numberof faculty membersin this study, 11 versus 3 in Borg and Shapiro,may
suggest these results are more reliable.More researchis neededto furtherexamine this relationship.
Takentogetherthese results suggest the importantrole that personalitytemperamentplays in the learningand teaching of microeconomicprinciples.This
result has severalimportantimplicationsfor pedagogy in introductoryclasses in
economics. If we want to increaselearningin economics, then we need to devise
ways of presentingcourse materialthat are accessible to all studentsregardless
of how they receive and process information.At a minimum, this suggests the
need for a varietyof classroompedagogies and a betterunderstandingof the role
of personalitytemperamentin learningand teaching.
NOTES
1. Some scholarshave suggestedthatthe contentof economics with its focus on competitionrather
than cooperationmay be difficult for some individuals(Nelson 1996).
2. Researchon preferencetype and occupationsuggests thatsome occupationsare dominatedby F
men and T women, that is, men who are counselors are predominantlyF, whereaswomen who
are engineersare predominantlyT.
3. Borg and Shapiro(1996) make use of the Keirsey-Bates(1984) analysis of learningstyles and
personality temperamentthat focuses on four learners: Dionysian (SP), Epimethean (SJ),
Promethean(NT), and Appollonian(NF).
4. By comparison,there were 3 professorsand 119 studentsin the Borg and Shapirostudy (1996).
5. None of the professorsin this sample includedthe PreTUCEexam resultsin the computationof
studentcourse grades in micro principles.Some faculty includedthe PostTUCEexam resultsin
the students'grades,but the weight given these scores were not available.The differentialimpact
of counting the PostTUCEresults as a componentof course grade was capturedin the series of
faculty fixed-effects variables(PROFI-PROFIO)included in each of the estimatedmodels. In
the models presentedin Tables3 and 4, four faculty fixed-effects variableswere statisticallysignificantrelativeto the omittedfaculty memberat a significancelevel of 10 percentor less; in the
model presentedin Table5, three faculty fixed-effects variableswere statisticallysignificantrelative to the omitted faculty memberat a significance level of 10 percentor less. In each model
estimated,the F test of joint significancefor faculty fixed-effects variablestakenas a groupsuggests the coefficients on these variablesare significantlydifferentfromzero at a confidencelevel
of 0.01.
6. Comparisonsof participatingand nonparticipatingfaculty and studentsare available from the
authorupon request.
7. Tuckerand Gillespie (1993) and Quinn,Lewis, and Fischer(1992) found significantconcurrent
validity between the MBTI and the Keirsey Sorter,which supportsthe use of either instrument
to determinean individual'spersonalitytype.
8. Descriptivestatisticsfor variablesused in this analysis are providedfor a sampleof 617 students.
Descriptivestatisticsfor smallersubsamplesof studentsused in this article(n = 329 and n = 400)
were not appreciablydifferentfrom the largersample from which they were drawnand are available from the authorupon request.
9. In additionto the variablesspecified in the model, faculty fixed-effects variableswere included
for 10 of the 11 faculty membersin an effort to control for class-specific differences.The coefficients and model summarystatistics presentedin Tables 3 through 8 reflect the inclusion of
these fixed-effects variables,although these variablesare not reported.In the models presented
in Tables 7 and 8, only one faculty fixed-effects variablewas statisticallysignificantrelativeto
the omittedfaculty memberat a significance level of 10 percentor less. In the models presented
in Table 6, no faculty fixed-effects coefficients were statistically significant.However,in each
model estimated, the F test of joint significance for faculty fixed-effects variablestaken as a
group suggested the coefficients on these variableswere significantlydifferent from zero at a
confidence level of 0.01.
10. I also used alternativemeasuresof matching in the analysis but found them to be statistically
insignificant.
Fall2000
321
11. See Kennedy(1997, ch. 5) for a discussionof testingprocedures.The complete model developed
in this study was estimated for both dependent variables, PostTUCE and PostTUCE minus
PreTUCE.However,coefficients on MALE, NONWHITE,HRSTAKEN,TOTALHOURS,and
variousspecificationsof TEMPMATCHwere not statisticallysignificantat the .10 level of significance.
12. The differentsample sizes (400 and 329 for the OLS models in Tables 6, 7, and 8 and 617 in
Tables2, 3, and 4 for the maximumlikelihoodmodels) reflect differencesin dataavailabilityfor
the dependentvariables,PostTUCEscore, PreTUCEscore, and course grade.The formerrelied
on the voluntaryparticipationof students,whereasthe latterwere obtainedfrom universitydata
sources.
13. Tested-down versions of the Borg and Shapiro (1996) model were also estimated using
PostTUCE and PostTUCEminus PreTUCEas dependentvariables. Model summarystatistics
suggest, that for the presentsample of studentsstudied,the models presentedin Tables 6-8 are
more robustthan the model developed by Borg and Shapiro(1996). These results are available
from the authorupon request.
14. Recall that in this sample two-thirdsof faculty were N whereas only one-thirdof students had
the N preference.
REFERENCES
Becker, W. 1997. Teaching economics to undergraduates.Journal of Economic Literature35 (3):
1347-73.
Borg, M. 0., and S. L. Shapiro. 1996, Personalitytype and studentperformancein principlesof economics. Journalof Economic Education27 (Winter):3-25.
Centerfor Applicationsof PsychologicalType. 1993.An assortmentoffacts about type in education.
Gainesville, Fla.
Charkins,R. J., D. M. O'Toole, and J. N. Wetzel. 1985. Linking teacherand studentlearningstyles
with studentachievementand attitudes.Journalof EconomicEducation 16 (Spring): 111-20.
Keirsey,D., and M. Bates. 1984. Please understandme: Characterand temperamenttypes. 5th ed.
Del Mar,Calif.: PrometheusNemesis.
Kennedy,P. 1997. A guide to econometrics.3rd ed. Cambridge,Mass.: MIT Press.
Myers, I. 1975. Manual: Myers-Briggstype indicator. Palo Alto, Calif.: Consulting Psychologists
Press.
Myers, P., and K. Myers. 1991. MBTI:Introductionto type in organizationalsettings. Palo Alto,
Calif.: ConsultingPsychologists Press.
Quinn, M. T., R. J. Lewis, and K. L. Fischer. 1992. A cross-correlationof the Myers-Briggs and
Keirsey instruments.Journalof College StudentDevelopment33 (3): 279-80.
Reichmann,S., andA. F Grasha.1974.A rationalapproachto developingand assessing the construct
validity of a studentlearningscale. Journalof Psychology 87: 213-23.
Saunders,P. 1991. The thirdedition of the Test of Understandingin College Economics. The Journal of Economic Education22 (Summer):255-72.
Tucker,I. E., and B. V. Gillespie. 1993. Correlationsamong three measuresof personalitytype. Perceptual & Motor Skills 77 (2): 650.
Wetzel, J. N., W. J. Potter,and D. M. O'Toole. 1982. The influence of learningstyles and teaching
styles on studentattitudesand achievementin the introductoryeconomics course: A case study.
Journal of Economic Education 13 (Winter):33-39.
322
JOURNALOF ECONOMICEDUCATION