You are on page 1of 11

"SEE HOW THEY LEARN": THE IMPACT OF FACULTY AND STUDENT LEARNING STYLES

ON STUDENT PERFORMANCE IN INTRODUCTORY ECONOMICS


Author(s): Kara Boatman, Richard Courtney and William Lee
Source: The American Economist , Spring 2008, Vol. 52, No. 1 (Spring 2008), pp. 39-48
Published by: Sage Publications, Inc.

Stable URL: http://www.jstor.com/stable/40657694

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Sage Publications, Inc. is collaborating with JSTOR to digitize, preserve and extend access to
The American Economist

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
"SEE HOW THEY LEARN": THE IMPACT OF FACULTY AND
STUDENT LEARNING STYLES ON STUDENT PERFORMANCE IN
INTRODUCTORY ECONOMICS1

by Kara Boatman,* Richard Courtney, and William Lee

Abstract

"See How They Learn": The Impact of Faculty and Student Learning Styles on Student Performance
in Introductory Economics
This paper reports the results of a recent study completed at Saint Mary's College of California, in
which the TUCE was administered to introductory economics students. Students and instructors also
completed the VARK questionnaire to identify which of the sensory modalities they prefer to use to
learn information. Results suggest that a strong visual learning preference positively influences stu-
dent performance. Our finding that neither ethnicity nor gender influence student performance con-
firms results of prior research, and suggests that ethnicity- and gender-based differences in student
performance may be at least partially caused by differences in learning style preferences.

I. Introduction prefer to learn economics. Second, while our study


confirms some reported results from similar
Much of the literature on economics education inresearch, it contradicts prior research results in
important areas. We hope that discussion of both
the past decade has questioned the effectiveness of
introductory economics courses (Becker 1997).the similarities and differences will contribute to
Economics instructors, these authors argue, mustthe academic community's understanding of criti-
move beyond the traditional lecture approach tocal success factors in the learning of economics.
more "active" teaching strategies such as coopera- The paper is organized as follows. Section II
tive learning if they are to successfully engage summarizes the relevant literature and describes
today's students. Without such changes, they warn, how our research contributes to this literature. In
students will be less likely to take economicsSection III, we describe the VARK learning styles
courses, and/or to continue in their economics stud-questionnaire and our rationale for using it in our
ies beyond required courses. The need for a variety study. Section IV presents our model and analysis.
of teaching methods is supported, in many cases, byWe discuss our conclusions and suggestions for
further research in Section V.
differences in student learning styles. If students
absorb and process information differently, then
shouldn't instructors present the material in differ-
ent ways in order to maximize learning outcomes? II. Literature Review
Several recent papers have focused on the
impact of student learning styles and personality Much of the literature on performance in intro-
types on the ability to learn economic concepts.ductory economics classes identifies standardized
This paper contributes to the literature in two test scores, prior academic performance and previ-
ways. First, unlike much of the existing researchous experience in the subject as significant predic-
that relies on personality type as a proxy for learn-tors of student success. In addition, researchers in
ing style, we use the VARK (Visual, Aural, Read/other fields have studied the impact of learning
write and Kinesthetic) inventory as our learningstyles on student performance in introductory
style indicator2. We believe the learning style courses (see, for example Bartlett, Hallock, Kellogg
modalities identified in the VARK inventory lend et al 1996, and Thomas et al 2002). Their results are
themselves well to an analysis of how students mixed; some conclude that learning style can impact

* Corresponding author, Saint Mary's College of California, kboatman@stmarys-ca.edu

Vol. 52, No. 1 (Spring 2008) 39

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
student success, while others find no significant rela- gender is not a statistically significant predictor
tionship. We focus here on the literature that intro- of success in economics. Finally, contrary to
duces student learning style as an additional "input" Borg and Shapiro and Charkins et al, Ziegert's
into models of economics learning "output." research does not suggest that commonality
Charkins et al (1985) search for a link between between instructor and student personality type
teaching styles and learning styles, and try to iden- improves student performance.
tify any impact of such a link on student learning. Our study extends this literature in several
These authors use a questionnaire that classifies ways. First, we rely on the VARK inventory to
teacher and student learning as dependent, inde- assess learning styles. In our view, this inventory
pendent or collaborative. They include the SAT provides a valuable measure of learning style
score as a measure of aptitude, the TUCE (Test of because it directly assesses how students prefer to
Understanding of College Economics) pre-test as a learn, rather than indirectly predicting their learn-
measure of the stock of knowledge, and the per- ing strengths through a personality assessment.
centage change in the pre- versus the post-test Second, we pursue the still unresolved issue of
TUCE score as a measure of achievement in intro- commonality between professor and student
ductory economics courses. They find that the learning preferences, and whether such common-
greater the divergence between teaching style ality
and significantly improves student performance.
Finally, comparison of our results with those in
student-preferred learning styles, the lower is stu-
dent achievement. previous studies will, we hope, shed light on how
the teaching of introductory economics can best
Borg and Shapiro (1996) use the Myers-Briggs
be tailored to different student populations to
Type Indicator (MTBI) to evaluate macroeco-
improve learning outcomes and attract more stu-
nomic principles students and professors to see if
dents to the field.
personality type affects student grades and if stu-
dents with personality types similar to those of
their professor perform better. They conclude that
III. The VARK Questionnaire
students whose personality type suggests a prefer-
ence for a structured learning environment per-
Our analysis relied on student responses to the
form significantly better in macroeconomic
VARK questionnaire developed by Fleming and
principles courses than those whose personality
Mills (1992). The questionnaire is comprised of
type suggests a preference for either independent
thirteen questions that generate a profile of modal
learning or extensive interaction during class. In
preferences regarding the presentation and pro-
addition, consistent with Charkins et al, these
cessing of information. The modalities are: visual,
authors find that students who share a personality
aural/auditory, read/write, and kinesthetic. Each of
type with their professor perform significantly these modalities is described in turn.
better.
Individuals with a visual modality prefer to learn
Ziegert (2000) tests the hypothesis that faculty information that is presented in charts, graphs, and
and student personality types affect student abili- symbols, rather than words.3 An auditory/aural pref-
ties to understand economics, as measured by the erence indicates that the individual prefers to receive
TUCE and course grade. Like Borg and Shapiro, information that is communicated orally. Students
she uses the Myers-Briggs personality type indi- with this preference would seem to learn best from
cator and finds that "thinking" students (those lecture and group discussion, among other tech-
that make objective judgments) outperform feel- niques. Individuals with a read/write modality indi-
ing "students" (those that make decisions based cate a preference for information that is presented
on personal values). Her research also suggests in text. Fleming (1995) emphasizes that the "read/
that "intuitive" students (those that focus on the write" preference is distinct from the visual prefer-
larger picture prior to details and learn from ence. While both diagrams and written words are
insight) outperform "sensing" students (those that absorbed visually, students with the former modal-
prefer experience-based learning). Ziegert finds ity prefer to work with textual material, while those
no evidence of a gender gap in her study; once with a visual preference prefer to work with sym-
personality differences have been accounted for, bols and/or graphs.

40 THE AMERICAN ECONOMIST

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
A kinesthetic preference suggests that the stu- introductory economics. For example, one would
dent learns best when the information is deliveredexpect a student with a strong visual modality to
through active learning activities, including con-enjoy the extensive use of graphs in introduc-
crete examples, practice, simulation or personal tory economics courses. A kinesthetically oriented
experience. These students learn theoretical con- student may appreciate an abundance of "real-
cepts when those concepts are applied to specific world" applications of economic principles and
situations. Students with a multi-modal preferencein-class experiments. Economics is distinct from
have two possible profiles. Some are "balanced many other college-level subjects in that it can be
learners" that have equally strong preferences fortaught in a variety of ways. Economics instruc-
two or more modalities and can use different tors can deliver information visually using graphs
and equations, aurally by relying on a lecture
strategies in different courses. Others must process
information in a number of ways before they format,
fully textually through reading assignments, or
understand it. by employing a combination of these and other
Each VARK question presents a situationtechniques.
likely Using the VARK questionnaire, eco-
to be within the respondent's experience and asksinstructors can consider student learning
nomics
him to select from among alternative actions.preferences
Each in selecting from among these teach-
ing techniques.
answer represents a modal preference. Respondents
may select multiple answers and all answersIn areour view, the VARK scoring system is
counted. The questionnaire was developed to iden-
preferable to those of other learning style invento-
tify which sensory modalities subjects prefer to because it distinguishes between mild,
ries. First,
strong and very strong preferences, we can assess
employ when learning or presenting information.
The questionnaire is scored by totaling the the
importance not only of learning style prefer-
number of answers that indicate a preference ence,for
but that of preference strength. Second, the
each mode. Because multiple and omitted answers
scoring system identifies respondents with multi-
are permitted, score totals may vary across respon-
modal preferences, rather than "forcing" them into
dents. A modal preference is then determined by
a particular category. This scoring feature adds a
establishing the numerical dominance of one mode
subtle dimension to the inventory that is lacking in
over the others. For example, if a student selects
other8learning style assessment methods.
read/write answers, 2 visual answers, 4 aural Finally, because the questionnaire identifies both
answers and 2 kinesthetic answers, the student preferences for presenting and processing informa-
would be viewed as having a read/write learningtion, we can assess teaching as well as learning
preference. For those respondents whose answers modal preferences by administering the question-
indicate that a single (rather than multi-modal) naire to both professors and students.
preference dominates, that preference can then be There has been little or no independent research
categorized as mild, strong or very strong, depend- on the quality, validity or reliability of the VARK
ing on the difference between the dominatingquestionnaire. However, in their review of learning
mode score and the next highest preference. style instruments, Hawk and Shah (2007) note that
We selected the VARK questionnaire for severalthe VARK model is the only one of five studied
reasons. First, the VARK system focuses on one that contains the read/write and kinesthetic dimen-
dimension of learning style: sensory preferences forsions. According to the VARK website 58 percent
how to absorb and deliver information. It is pre- of individuals report a match between the ques-
cisely this dimension that we focus on in our study. tionnaire results and their own perceptions of
However, unlike other sensory inventories, the learning preferences.4
VARK questionnaire distinguishes between two
types of visual learners: those who like to work
with text and those who prefer to work with graphs, IV. Model and Analysis
symbols and diagrams. Clearly, this distinction is
critical for economics students, since it implies the Our study was conducted during the fall and
need for very different teaching techniques. spring semesters of the 2005/2006 academic year.
The VARK learning modes identified lend Saint Mary's College of California is a private,
themselves well to the instruction and learning ofCatholic, liberal arts college located in Moraga,

Vol. 52, No. 1 (Spring 2008) 41

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
California. The College's traditional undergraduate enrolled in introductory micro- and macroeco-
program had an enrollment of 2,525 students nomics are business or economics majors. A few
during the fall 2005 semester, 39 percent of whom students take one of these courses to fulfill a
were men and 61 percent of whom were women. general education requirement or as an elective
Of the students enrolled in the College's tradi- The students were told about the study and its
tional undergraduate program, 55 percent were purpose, and only completed the questionnaire i
Caucasian; 20 percent Latino; 10 percent Asian they authorized access to and use of institution
American; 6 percent African American; with 1 data including gender, high school GPA, ethnic
percent American Indian/Alaska Native. ity and other personal information. Only 49 per
The mean verbal SAT score for freshmen enter- cent of the students surveyed were women,
indicating that women are underrepresented in
ing the College in the fall 2005 semester was 548,
while the mean math SAT score was also 548. The economics classes relative to total undergrad
ateinenrollment. In addition, of the students sur-
average high school GPA of freshmen entering
the fall 2005 semester was 3.11. veyed, 57 percent were Caucasian, 20 percen
were Latino, 13 percent were Asian- American,
Our analysis was designed to answer several ques-
tions:
7 percent were African-American, and 2 percen
were Native American. These percentages ar
similar to the statistics for the undergradua
• Does learning style preference significantly
program as a whole.
impact student performance in introductory
economics courses and, if so, how?
• Do gender and ethnicity impact student per-
formance? 3. TUCE
• Does commonality between professor and stu-
dent learning preferences improve student per- The Test of Understanding in College Economi
formance? has been in existence for 40 years and used exte
• How do the results of our research compare to sively by researchers in economics. The test is now
similar studies conducted at other institutions? in its fourth edition. According to Walstad (2007
"As with previous editions, the TUCE-4 has tw
main objectives: (1) to offer a reliable and valid
L Theoretical Model assessment instrument for students in principles o
economics courses; and (2) to provide normin
In keeping with the economic education litera-
data for a large, national sample of students in pri
ciples classes, allowing instructors to compare pe
ture, we base our empirical model on the theoreti-
formance
cal education production function, where output of in their classes on both pretests an
economic knowledge is a function of productiv-
posttests to the performance of the national sampl
of students and instructors" (Walstad, p. 1). Th
ity inputs, including measures of student ability
fourth edition of the test consists of a 30-item test
such as high school GPA and SAT scores. Student
characteristics, such as gender and ethnicity, are
for microeconomics and a separate 30-item test for
included since they may systematically affectmacroeconomics.
input
The TUCE has been criticized for several rea-
productivity. Finally, we hypothesize that certain
learning style modalities affect the student's sons. First, there are concerns that it may be biased
ability
to learn economics. against students who either don't perform well on
multiple choice tests or have little experience with
the multiple choice format. Second, because the
2. Data exam is standardized, it may not perfectly reflect
the material presented in individual classes. Finally,
During the first week of each semester, performance
four on the test may not effectively dis-
criminate among students with different levels of
instructors administered the VARK questionnaire
to 211 students. Of the four, one is a full profes-
cognitive achievement (Swartz et al, 1980).
sor, one is an associate professor, and two In arethis study, the fourth edition of the TUCE
(TUCE-4) was administered as a pre-test to 95
adjunct instructors. The majority of students

42 THE AMERICAN ECONOMIST

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
students in four microeconomics principles impose any requirements
courses in the fall 2005 semester and to 58 stu-
ual instructors should in
dents in two microeconomic principles courses into
in their final course
penalty
the spring 2006 semester. The pre-test was admin- associated with
either the pre- or post-te
istered during the first week of the semester. The
most likely did not stu
same test was administered as a post-test to stu-
TUCE and did not necessa
dents in these courses during the last week of the
semester. During the spring 2006 semester, the identifying the correct a
same procedure was followed for 68 students This
in explanation is consis
average increase in scor
three macroeconomic principles courses. However,
the third edition of the TUCE (TUCE III) was post-test was only 3 poi
used for the macroeconomics courses. because of the weak correlation between TUCE
Female students of all ethnic backgrounds and final grade measures of student performance,
averaged a 9 percent improvement in their TUCEwe should estimate different versions of the
score from the beginning to the end of the model using both the TUCE post-test score and
final grades as the dependent variable. We
semester, while their male counterparts averaged
a 10 percent improvement. Caucasian students acknowledge that exams and grading policies dif-
averaged an 8 percent improvement, compared fered
to across instructors, possibly biasing the
a non-Caucasian student average improvement results
of of the model using final grade as the
6 percent. dependent variable.
Several other important results emerged from pre-
liminary regressions and correlations. First, verbal
4. Preliminary Analyses SAT score had no impact on student perform-
ance in introductory economics courses, measured
Prior studies have used final course grade, TUCEeither by TUCE scores or final course grades. This
post-test scores, and pre- and post-TUCE score dif-result was somewhat surprising, since we do not
consider our introductory economics courses to be
ferences as measures of student learning or perform-
ance in introductory economics. A priori, we had noparticularly quantitative. In response to this out-
come and in contrast to earlier studies that use
strong view as to which is the best measure of stu-
total SAT score as an input, our model includes
dent performance, so we performed a correlation
only the student's math SAT score. Second, prelim-
analysis between the TUCE post-test score and final
grade. The results are presented in Table 1. inary regression results suggested that commonal-
Results of this analysis suggest that there is between professor teaching style preference and
ity
only a weak correlation between the TUCE
student learning style preference did not signifi-
cantly influence performance, measured either by
scores and the final grade in the course. There are
a number of possible explanations for this, but TUCE data or final grade data. While this result
we believe it may be because of how we adminis- confirms Ziegert's findings, it is somewhat surpris-
tered the TUCE. In an effort to attract as many ing, since we believe the VARK questionnaire
instructors to the study as possible, we did not represents a better measure of teaching style prefer-
ence than the Myers-Briggs indicator used by
Ziegert. One possible explanation for this result is
that, because of small class sizes at Saint Mary's
TABLE 1.
(typically 30 students or less), professors have time
Correlation Analysis
during class to employ a variety of teaching methods
Correlation
to appeal to different student learning preferences,
Coefficient
regardless of their own personal preference for pre-
with Final
senting information. For example, one of the profes-
Grade
sors in the study has a strong preference for the
TUCE Post-Test Score 0.41
reading modality, yet she encourages class and
Difference Between Pre- and
group discussions (aural), and uses numerous con-
Post-Test TUCE Scores crete examples (kinesthetic) in class. Another

43
Vol. 52, No. 1 (Spring 2008)

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
possibility is that, unlike many subjects, teaching where PERFORMANCE = either the final grade
introductory economics requires visual presenta- in the course, measured on a 4-point scale,6 or the
tion of information. That is, notwithstanding a pro- TUCE post-test score;
fessor's preference for other modes of presenting
SATMATH = the student's math SAT score;
information, he or she must use at least some
HSGPA = the student's high school grade point
graphs in order to effectively present the material.
average;
Thus, it may be the case that in economics,
GENDER = a dummy equal to "1" for female
instructors do not have the luxury of eliminating
students;
less preferred modes entirely. Given these prelimi-
ETHNIC =a dummy equal to "1" for non-
nary results, we eliminated common teaching and
Caucasian students;
learning style preference from our final model
specification. VISUAL, AURAL, READ/WRITE, KINES-
Finally, preliminary regressions indicated that THETIC = dummy variables indicating a strong
only math SAT scores and high school GPA had or very strong student preference for visual, aural,
any predictive power over student learning as reading or kinesthetic approaches to learning,
measured by the pre- and post-test difference in respectively.
TUCE scores. We believe this result may be attrib- We were concerned about potential multi-
utable to the relatively minor improvement (3 collinearity between high school GPA and Math SAT
points, on average) in TUCE scores from the score, which could make the impact of each of these
beginning to the end of the semester, as well as variables on student performance more difficult to
the incentive issues described above. Conse- isolate. We relied on the Variance Inflation measure
quently, we will use post-test TUCE scores andthe impact of collinearity on the precision
to measure
of the
final grades as our two measures of student estimates. Calculation of the Variance Inflation
learn-
ing in introductory economics. Factor for these two variables generated values of
less than 2 in each case, suggesting that the variance
5. Final Model of the estimates of these coefficients is not highly
inflated by multicollinearity.7
We estimated the following final model: We employed ordinary least squares (OLS)
regression analysis and report robust standard
PERFORMANCE = a + $l * SATMATH errors.8 Regression results using final grade as the
dependent variable are presented in Table 2.
+ ß2 * HSGPA + ß3 * GENDER + ß4 * ETHNIC
Note that the number of observations in the
+ ß5 * VISUAL + ß6 * AURAL model and in later regressions is significantly
lower than the initial number of students tested at
+ ß7 * READWRITE + ß8 * KINESTHETIC

TABLE 2.
Regression Results: Dependent Variable is Final Grade
Coefficient Robust Standard Error t-statistic

SATMATH 0.0047643 0.0007962 5.98***


HSGPA 0.6510529 0.1559663 4.17***
GENDER 0.0235933 0.1370492 0.17
ETHNIC -0.1546154 0.1345821 -1.15
VISUAL 0.6141443 0.1664058 3.69***
AURAL 0.0078785 0.2842594 0.03
READ/WRITE 0.1515377 0.2074554 0.73
KINESTHETIC -0.070325 0.2203403 -0.32
R2 .45
# Observations 119

***, **, and * represent significance at the .01, 05 and .10 leve

44 THE AMERICAN ECONOMIST

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
the beginning of each semester. The smaller final basic economic concepts. As we stated earlier, even
sample size reflects students that dropped the those introductory economics instructors with non-
course, as well as students that were absent on the visual preferences present a significant amount of
day that either the questionnaire or TUCE test was information visually. Students who prefer to
administered. process information in this way would reasonably
Before we discuss the learning style variables, it be expected to perform better. The lack of signifi-
is interesting to note how our results compare with cance of other preferences is consistent with our
those of earlier studies. Consistent with earlier stud- preliminary result that commonality between teach-
ies, high school GPA and math SAT score coeffi-ing and learning style preferences does not influ-
cients are both highly significant and positive. The ence student achievement; because professors
results indicate that a 10-point increase in the math employ a variety of teaching methods in addition
SAT score improves the final course grade by .04 to visual presentation, no one of these other learn-
points, while a one-point increase in high school ing style preferences improve a student's perform-
GPA increases the final course grade by 0.65 points. ance. However, the lack of significance of the
However, our results indicate that neither genderkinesthetic preference should be interpreted with
nor ethnicity exert a significant influence on student caution. This result may be due to the fact that
performance, as measured by course grade. These strategies such as in-class debates, experiments, or
results are consistent with Ziegert's explanation that field research were not extensively employed in
once learning preferences or personality types arethese classes and therefore kinesthetic learners did
taken into account, presumed gender and ethnicity not exhibit any advantage over other students.
effects on final course grade may well disappear. We also tested an alternative form of the model,
Our primary interest, of course, is in the impact where student performance was measured by the
of learning styles on student performance. The TUCE post-test score. Results are presented in
model indicates that students with a strong prefer- Table 3.
ence for visual learning did significantly better in These results were consistent with the previous
introductory economics courses than their counter- model, with one important exception. Non-
parts, while other learning style preferences did not Caucasian students performed significantly worse
appear to have a significant effect. Specifically, stu- on this test than Caucasion students. This is con-
dents with a strong visual preference could expectsistent with TUCE results at other institutions (see
a 0.6 point higher final grade than their counter- Simkins and Allen, 2001). Some researchers have
parts without this learning style preference. The pointed out that the TUCE is biased towards the
visual preference result is not surprising, since personality types or teaching styles of the econo-
introductory economics classes have a large graphi- mists who designed it (Borg and Shapiro, 1996). If
cal component, and students are expected to be the designers were predominantly Caucasian, then
able to interpret and create graphs that illustratethe multiple choice questions may be constructed

TABLE 3.
Regression Analysis: TUCE Post-Test Score is Dependent Variable
Coefficient Robust Standard Error t-statistic

SATMATH 0.0256575 0.0051705 4.96***


HSGPA 1.91783 0.7579155 2.53**
GENDER -0.0553041 0.7770452 -0.07
ETHNIC -1.690185 0.7068015 -2.39**
VISUAL 5.48591 0.8943118 6.13***
AURAL 2.222059 2.346619 0.95
READ/WRITE 0.4605724 1.378746 0.33
KINESTHETIC 0.4612324 1.105828 0.42
R2 .42
# Observations 112

***, **, and * represent significance at the .01, 05 and .10 l

Vol. 52, No. 1 (Spring 2008) 45

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
TABLE 4.
Final Grade Regression Excluding "Fs"
Coefficient Robust Standard Error t-statistic
SATMATH 0.004322 0.0007179 6.02***
HSGPA 0.5744007 0.1464242 3.92***
GENDER 0.0991141 0.1314311 0.75
ETHNIC -0.1135582 0.1256467 -0.9
VISUAL 0.6770246 0.1605702 4.22***
AURAL -0.0143067 0.2711653 -0.05
READ/WRITE 0.3018865 0.1346478 2.24**
KINESTHETIC 0.0051086 0.1935647 0.03
R2 .44
# Observations 116

***, **, and * represent significance at the .01, 05 and .10 level

with a strong visual learning


in a way that is more comprehensible topreference,
students students
with common ethnicity. withIn aaddition,
strong read/write thelearningTUCE
preference also
reliance on a multiple choice
performformat may
significantly better than put stu-
their counterparts.
dents with certain modalityOnepreferences
explanation for this result
at isathat students with
disad-
this preference
vantage, particularly students who may do notnothave exhibit
been attentive in
a class,
"read/write" preference. orWithmay not respect
even have attended
to class,
TUCE planning
instead
post-test scores, the results to read the textbook.
indicate that Instudents
some cases, this
with a strong visual learning preference
strategy failed and students failed can be
the course. Once
expected to receive scores results
overfor these
5 students
points were removed
higher from the
data set, the regression captured the impact of a
than those without this preference.
We then used the same two versions of the final strong read/write preference on those students who
model to analyze a data set that excluded students were more actively engaged in the class. For this set
who failed introductory economics. We excluded of students, the preference for learning from textual
these data because, in some cases, the students input improves their economics grade by 0.3 points.
failed because they never attended class. Their In order to maintain consistency with earlier
grades, therefore, could not be affected by learning results, we also ran the TUCE regression with a
style preferences. The results of this analysis are data set that excluded students who received an "F'
presented in Table 4. for the course. Results are presented in Table 5.
The results are consistent with the earlier analy- Here again the ethnic dummy coefficient is signifi-
sis, with one exception. Now, in addition to students cant and negative. In contrast to the corresponding

TABLE 5.
TUCE Post-Test Regression Excluding "Fs"
Coefficient Robust Standard Error t-statistic
SATMATH 0.0250895 0.0053909 4.65***
HSGPA 1.759701 0.7877765 2.23**
GENDER 0.0700276 0.7982181 0.09
ETHNIC -1.749293 0.7450964 -2.35**
VISUAL 5.611735 0.9525867 5.89***
AURAL 2.280651 2.418713 0.94
READ/WRITE 0.3249676 1.4958 0.22
KINESTHETIC 0.8538297 1.130787 0.76
R2 .38
# Observations 104

***, **, and * represent significance at the .01, 05 and .10 le

46 THE AMERICAN ECONOMIST

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
final grade model, however, a strong read/write ics courses. This implication is consistent with our
preference does not significantly improve student anecdotal observations that introductory econom-
performance in introductory economics, while a ics students tend to be intimidated by graphs and
strong visual preference continues to exert a signifi- hesitant to use them unless they are required to
cant positive effect. One explanation for this result do so. Consequently, instructors should consider
is that because professors use a variety of assess- spending more time teaching students how to con-
ment methods in determining final course grades, struct and interpret graphs, rather than assuming
the read/write preference is a more significant that students enter the course with graphing skills
factor with respect to grade than it is with respect to already in place.
the post-test TUCE score. The strong influence of the visual learning pref-
erence also leads us to a suggestion for further
research. Specifically, we wonder if introductory
V. Conclusions, Implications and courses in economics may themselves increase
Suggestions for Further Research student preferences for information presented in a
visual mode and more generally, if learning style
Our research suggests that a strong preference preferences remain consistent before and after
for receiving information in a visual mode posi- studying economics.
tively influences student performance in introduc- Our study also indicates that, for those students
tory economics courses. The significance of the who pass the course, a preference for processing
visual modal preference is robust to alternative information that is presented textually gives them an
specifications of the dependent variable. This advantage relative to their peers with respect to final
result has interesting implications for the teaching course grade. This result suggests that textbook
of introductory economics. If the goal of these selection is extremely important for these students,
courses is to attract students to the field who will since they may rely on it heavily in mastering course
be successful in upper division courses and per- material. Moreover, assessment methods such as
haps in graduate studies as well, then the empha-papers and essay examinations may provide these
sis on visual presentation may be appropriate.students with a greater opportunity to demonstrate
Certainly the mathematics that is introduced intheir understanding of the material.
upper division and graduate school courses requires Our finding that neither ethnicity nor gender
that students be comfortable with symbolic presen- influence student performance, as measured by final
tation of information. Yet, if the goal of introduc- course grade, confirms Ziegert's results and suggests
tory courses is to educate the greatest number that ethnicity- and gender-based differences in stu-
of students about the basic premises of economics,dent performance may be at least partially caused by
emphasis on graphs in these courses may bedifferences in learning style preferences.
counter-productive. Less than two percent of the Interestingly, our study suggests that ethnicity
students we studied exhibited a strong visualis associated with lower scores on the TUCE, but
learning preference. If our courses are taught innot with lower grades. This result warrants further
such a way that this very small number of students research, since it may suggest that the TUCE does
is likely to outperform their counterparts, we may not allow non-Caucasian students to demonstrate
fail in our mission to "spread the word" withtheir understanding of the material as well as other
respect to basic economic concepts. If in additionassessment methods.
to the traditional reliance on graphs we can pres- Our preliminary research results indicated that
ent economic concepts in ways that attract stu- verbal SAT scores do not influence student per-
dents with non-visual learning preferences, we formance in introductory economics, while math
may widen the appeal of introductory economics. SAT scores are highly significant determinants of
Moreover, the fact that only a small minority of student achievement in these courses. In addition
students in our classes exhibit a visual learning to being consistent with the importance of the
preference suggests that most introductory eco- visual learning preference, the importance of math
nomics students will not see graphs as a useful skills suggests that instructors can more easily
learning tool, but as an obstacle that must be over- identify at-risk students in introductory economics
come in order to succeed in introductory econom-by accessing their scores on the math SAT. The

Vol. 52, No. 1 (Spring 2008) 47

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms
result also suggests the importance of presenting to better understand the impact of this preference on
conceptual explanations in addition to graphs and learning in introductory economics.
equations in these courses.
The fact that students with strong aural and
kinesthetic learning preferences did not outper- Notes
form their peers suggests that instructors might
make greater use of concrete examples and class 1. The authors would like to thank an anony-
discussions in presenting introductory economics mous referee for helpful comments and sug-
material. The results of our research are broadly gestions.
consistent with other studies. However, because 2. Drago and Wagner (2004) apply the VARK
the VARK approach affords a narrow focus on learning styles to online education.
preferences with respect to the receipt and presen- 3. Note that a visual preference specifically
tation of information, we believe our results point excludes movies, videos or PowerPoint presen-
even more clearly to the need for adopting a vari- tations.
ety of teaching strategies and assessment methods 4. Only 4 percent of those individuals respond-
in introductory economics courses. ing to the website questionnaire reported that
While we believe the results of this study are the results did not match their own percep-
instructive, we hope to continue this research and tions of learning style preferences.
improve our study design. Specifically, we will ask 5. In fact, one of the professors administered the
instructors participating in the study to incorporate post-test without notifying students ahead of
the TUCE post-test into their final exam, so that it time.
more accurately and consistently reflects student 6. The scale is calibrated as follows: a "B" is a
achievement. Second, we will explore other assess- 3.0, a B + is a 3.3, and a B- is a 2.7.
ment methods in addition to the TUCE that might 7. Typically, a VIF of 10 or more suggests the
better reflect the impact of learning style prefer- presence of harmful multicollinearity.
ences on academic performance. Finally, we will 8. Robust standard errors were employed in
encourage faculty participants to include teaching order to improve the accuracy of the t-statistic
strategies that appeal to kinesthetic learners in order if heteroskedasticity is present.

48 THE AMERICAN ECONOMIST

This content downloaded from


115.242.193.90 on Mon, 22 Apr 2024 09:05:04 +00:00
All use subject to https://about.jstor.org/terms

You might also like