How college instructors can enhance student achievement: testing a learning styles theory.

Terregrossa, Ralph A. ; Englander, Fred ; Wang, Zhaobo 等

INTRODUCTION

Students' academic achievement depends critically on the teaching effectiveness of the instructor. In many cases, college instructors are evaluated for purposes of retention, tenure and promotion based on their teaching effectiveness, as measured by the gains in academic achievement accomplished by their students. Two important factors that determine the effectiveness of an instructor are the instructor's breadth and depth of knowledge of his or her discipline and the instructor's ability to convey that knowledge. The Dunn and Dunn learning styles model (DDLSM) (Dunn, 2000), contends that the ability of an instructor to effectively convey knowledge to students depends on the extent to which the instructor's teaching method matches the pattern of learning style preferences of his/her students. The cornerstone of the DDLSM is the hypothesis that the optimal method of instruction in any discipline is that method that matches students' learning styles.

Students' learning style preferences are identified from their responses to the Building Excellence (BE) learning styles survey (Rundle & Dunn, 2009a). The BE survey is designed to reflect the revised version of the Dunn and Dunn (Rundle & Dunn, 2009b) learning styles model which includes environmental, emotional, sociological, perceptual, physiological and psychological learning style categories. The BE is an online survey containing 118 selfreflective questions that are answered on a five-point Likert scale. The BE survey identifies all twenty-six learning style elements contained in the six categories that comprise the DDLSM (Rundle & Dunn, 2009b). For example, the preference for noise is determined by answering 'strongly agree', 'agree', 'uncertain', 'disagree' or 'strongly disagree' to the following statements:

* I concentrate best in quiet surroundings with no sound or people talking.

* I concentrate best when there is sound, or when music is playing in the background.

* I concentrate best in a quiet place--especially when I am working on difficult tasks.

* I concentrate best with sound in the background, when I am working on difficult tasks.

A refined list of the learning style preferences identified by the BE survey are included as independent variables in regression model to explain variations in students' academic achievement. Qualitative variables reflecting the influences of gender and any potential differences in the rigor of the three semester exams on student achievement are included as additional explanatory variables. The analysis presented here examines whether the learning style variables have a statistically significant effect on student achievement. Additional analysis allows a ranking of the relative importance of the contribution of the alternative learning style variables and learning style categories to students' academic achievement. Moreover, students' learning styles in the accounting and economics courses are compared to determine if differences exist between the disciplines regarding the significance of the marginal contribution and relative importance of the learning style variables and categories to student achievement. This comparison provides insight regarding the merits of tailoring instructional methods to more closely match students' learning styles to enhance students' academic achievement. Examples are provided of potential teaching methods that may be used by instructors to match the most influential learning style preferences of the economics and accounting students in order to enhance both teaching effectiveness and student achievement.

LITERTATURE REVIEW

Coffield, Moseley, Hall and Ecclestone (2004) examined the validity, reliability and practical application of major learning styles models, including the DDLSM (Dunn, 2000) model. Although the authors find that the model is reliable in terms of its predictive ability, they conclude that the findings of the numerous studies that support the DDLSM (Dunn, 2000) model are suspect because of the studies' methodological limitations and because of a lack of independent research conducted on the DDLSM (Dunn, 2000) model.

Hawk and Shaw (2007) also examined the validity and reliability of major learning styles models as well as the respective instruments used to identify learning styles, including the Dunn, Dunn and Price (2006) Productivity Environmental Preference Survey (PEPS). The PEPS, designed to reflect the earlier version of the DDLSM (Dunn, 2000) model, includes the environmental, emotional, sociological, physiological and psychological categories, which are composed of twenty interrelated learning style preferences. Hawk and Shaw (2007) report that solid evidence exists that supports the reliability and validity of the PEPS instrument to identify students' learning styles. Consistent with Hawk and Shaw (2007), the study by Terregrossa, Englander and Wang (2009a) provides evidence supporting the construct validity of the PEPS instrument, the internal validity of the Dunn and Dunn (2000) model and the hypothesis that students' learning styles contributes significantly to student achievement.

The Building Excellence (BE) survey (Rundle and Dunn, 2009a), designed to reflect the revised version of the Dunn and Dunn (Rundle & Dunn, 2009b) learning styles model, is wider in scope than the PEPS instrument in terms of identifying students' learning styles. Unlike the PEPS instrument, the BE survey includes a separate learning styles category accounting for an expanded list of perceptual modality preferences, including visual-text, visual-graphic, auditory (learning by listening), verbal-kinesthetic (learning by verbalizing) and tactile/kinesthetic. The BE survey is a relatively new instrument and therefore has not been substantially scrutinized in statistical models designed to explain variations in college student achievement.

The methodology developed in this analysis represents the first known utilization of students' learning styles measured by the BE instrument in a regression analysis framework to examine the contribution of students' learning styles to their academic achievement. In this way, the methodology applied in this study represents an improvement over previous research that investigated the relationship between college students' academic achievement and their learning styles.

Learning Styles

The revised Dunn and Dunn learning styles model (Rundle, & Dunn., 2009b) theorizes that an individual's learning style is composed of a combination of interrelated perceptual, environmental, physiological, emotional, sociological, and psychological categories. The perceptual category includes preferences for alternative perceptual modalities, including auditory, or learning by listening; visual-picture, or learning by seeing images, illustrations or pictures; visual-word, or learning by reading; tactile and/or kinesthetic, or learning through hands-on experience and by doing; and verbal-kinesthetic, or learning by verbalizing.

The environmental category includes preferences for background sound versus silence, bright or soft light, cool or warm temperature and formal versus informal seating. The physiological category reflects the student's ability to remain energized, focused and alert. This category includes preferences for intake of snacks or drinks while learning, the time of day when the student does his or her best work, and whether the student needs to be moving while learning.

The emotional category includes preferences for internal versus external motivation, persistence, or starting and finishing one project at a time, conformity to societal norms, and structure, or a preference for internal or external direction. The sociological category reflects whether students prefer to work alone, with a partner or with a group of peers, and whether students prefer to learn with an authoritative versus collegial adult. This category also reflects whether students prefer to learn using a variety of methods or by using an established routine.

The psychological category includes the preference for either a reflective or compulsive approach to making decisions and solving problems. This category also identifies the students' thought processing method, hypothesized to include analytic, global and integrated processing methods. The integrated learners have both analytic and global characteristics and utilize the alternative styles depending on the nature of, and interest in, the material to be learned.

Analytic learners learn best in a quiet, brightly lighted and formal (e.g., sitting at a desk) environment. They like to work alone, tend to be persistent (e.g., prefer to start and finish one project at a time), and do not snack while learning. They also learn more easily when details are presented in a sequential, step-by-step manner that builds toward a conceptual understanding of the idea to be learned.

Global learners learn more easily when they understand the total concept first then subsequently focus on the underlying details. They learn best with background sound, soft light in a relaxed environment (e.g., sitting on a couch or in a coffee shop). They prefer to work with others, tend not to be persistent (e.g., work simultaneously on several projects), take frequent breaks, enjoy snacks when learning, and prefer to be taught with the use of illustrations and symbols. Global learners prefer new information to be presented anecdotally, especially in a humorous way that explains how the concept relates to them. The revised DDLSM (Rundle & Dunn, 2009b) hypothesizes that analytic, global and integrated learners can be identified by their preferences for noise, light, design, persistence and intake.

OBJECTIVES AND HYPOTHESES

The main objective of this study is to determine if instructors could effectively utilize information about their students' learning styles to enhance students' academic achievement. To achieve this objective, undergraduate students enrolled in principles of macroeconomics and accounting courses are analyzed to test the hypothesis that academic achievement is significantly correlated with student learning styles, ceteris paribus. A second objective of this study is to rank order the relative impact of the alternative learning style categories on student achievement. The hypothesis tested is whether a particular learning styles category (collection of related, individual learning style variables), when considered as a subset of the total number of explanatory variables, and confronted by the remaining categories, has a significant effect on student achievement. A third objective of this study is to determine the relative importance of the alternative learning style variables that compose the learning style category that has the greatest impact on student achievement. Standardized regression coefficients are used to accomplish this objective. A fourth objective of this study is to determine whether learning styles differ in a meaningful way between students in accounting courses and economics courses. If the learning style variables that have a significant impact on student achievement in the two disciplines differ, then alternative teaching strategies to reflect such differences are warranted.

METHODOLOGY

Test scores of students are regressed against a refined list of learning style variables, gender and other control variables to determine if learning styles have a significant impact on student achievement. Additional analysis is offered to determine the relative importance of the alternative learning style variables and learning style categories, respectively.

Data for the twenty-six learning style preferences were collected from the BE survey that was administered to over twenty-six hundred entering freshmen in the fall semester of 2004. The sample analyzed in this study included a total of sixty-one students, twenty-seven in economics and thirty-four in accounting. Students in both courses were taught in a college of business, accredited by the AACSB, at a university located in the New York City. The macroeconomics course was taught by an associate professor of economics, and the accounting course was taught by an associate professor of accounting.

The same OLS regression model was applied to both disciplines. Students' test scores, a proxy for academic achievement, served as the dependent variable. Three tests were administered in the macroeconomics course during the semester for a total of eighty-one observations, and four tests were administered in the accounting course for a total of 136 observations. The explanatory variables included a refined list (explained below) of eighteen learning style variables, a binary variable accounting for the influence of gender (females serving as the base category), and binary variables accounting for the influence of possible differences in the difficulty level of the alternative tests, with the last test serving as the base category. The OLS regression model is summarized in the following equation:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where E[X.sub.ij] is the ith student's score on the jth test, i goes from 1 to 34 and 1 to 27 for the accounting and economics courses, respectively, and j goes from 1 to 4 and 1 to 3, for accounting and economics courses, respectively; LS[V.sub.bi] is the bth learning style variable for ith student; GDummyi is the gender binary variable with female serving as the base category; [TDummy.sub.ti] are the binary variables representing the different characteristics of the tests, with the last test serving as the base category in both disciplines (In the economics regression model only the first and second test dummy variables are included since only three tests were administered during the semester); and [e.sub.ij] is a stochastic error term.

With regard to the refinement of the learning style variables, there naturally exists a high degree of correlation among many of the learning style variables. Consequently, the parameters associated with the learning style variables do not have a separate and independent influence on student achievement. It would therefore be difficult to separate the effect of one particular learning style variable from the effect of the other correlated learning styles variable(s) on student achievement. Hence, the ability of the learning style variables to adequately explain differences in student achievement would be compromised. Pindyck and Rubinfeld (1998) suggest that it may be reasonable in such cases to drop one of the collinear variables from the model and re-estimate it. The results of the re-estimated model then can be compared with the original model "to gauge the effect of the multicollinearity in the original model (Pindyck & Rubinfeld, 1998, p. 97)."

The method suggested by Pindyck and Rubinfeld (1998) was utilized to identify which learning style variables are highly collinear and candidates for exclusion from the student achievement regression model. The regression models, including all twenty-six learning style variables, were estimated and the variable inflation factor (VIF) scores were analyzed to determine which learning style variables had a score equal to or above ten, the generally accepted score that signals the presence of a troubling level of multicollinearity. In the economics course, all twenty-six learning style variables had VIF scores well above the acceptable limit and in the accounting course, ten of the learning style variables were above the acceptable limit. The covariance matrix of all twenty-six learning style variables also was analyzed to identify pairs of variables whose correlation coefficient exceeded 0.40, a significantly high value. In addition to the variables that obviously are highly collinear by virtue of the nature of the DDLSM (e.g., the different time-of-day preferences), variables within and among learning style categories with both exceedingly high VIF scores and pair-wise correlation values were excluded from the final version of the regression model.

The resultant regression model included eighteen of the twenty-six learning style variables. The VIF scores for all the resultant learning style variables were below the critical value of ten, the standard errors of the estimated coefficients were substantially reduced and, therefore, the explanatory power and reliability of the student achievement regression model were increased. The forward and backward stepwise regression methods were applied and the results were consistent with the inclusion of the resultant eighteen learning style variables, instead of the total twenty-six learning style variables, in the regression model. These results indicate that the methodology applied to refine the list of learning style variables provided a reasonable method of dealing with the inherent multicollinearity.

EMPIRICAL RESULTS

The results for the economics and accounting regressions are summarized in Tables 1 and 2, respectively. The explanatory variables are reported in descending order of relative importance, as determined by the standardized regression method. In the principles of economics course, the adjusted R-squared was .49, a relatively high number for student production function regression models of this sort. In the studies by Terregrossa, Englander and Wang (2009a) and Englander, Terregrossa and Wang (2009b), which utilized the PEPS instrument that reflected the earlier version of the DDLSM, the adjusted R-squared was .27 and .22, respectively. The increase in explanatory power may be attributed to the greater accuracy that Morten-Rias et al. (2008)

attribute to the BE survey as compared to the PEPS instrument. The value of the F-statistic, 4.69, is statistically significant (a = .0001) and indicates that the model explains a large proportion of the relative variance of economics students' test performance.

Several learning style variables contributed significantly to student achievement in the economics course. In descending order of importance, the tactile/ kinesthetic, intake, light and authority variables were positively and significantly ([alpha] =.01) correlated, the mobility and persistence variables were negatively and significantly ([alpha] =.01) correlated, and the visual-picture and auditory variables were negatively and significantly ([alpha] =.05) correlated with student achievement. The last finding is particularly revealing because, as Becker and Watts (2001) report, the teaching method predominantly used by instructors in the principles of economics courses is the lecture format supported by written notes and graphs drawn on the board. This method of instruction is discordant with this sample of students' learning styles and, according to the DDLSM, therefore may be an ineffective teaching method. The negative coefficient of the mobility variable indicates that movement by students in class may be a distraction that negatively influences student achievement and therefore should be limited. It is noteworthy that the visual-text variable was rated last in terms of relative importance, indicating that, for this group of students, reading the material in the textbook contributed the least to student achievement, ceteris paribus.

Contrary to Durden and Ellis (2003) and Krohn and O'Connor (2005), who observed greater academic achievement by males than females in economics courses, gender did not have a significant effect on student achievement, ceteris paribus. Although both test dummy variables were negative and significant ([alpha] =.05), indicating that students performed significantly better on the last exam relative to the first and second exams, they were relatively unimportant variables, ranked thirteenth and fifteen, respectively. In the accounting course, the adjusted R-squared value, .39, also is comparatively high for student production function regression models. As in the case of the economics regression model, the relatively strong explanatory power of the accounting regression model may be attributed to the increased accuracy of identifying students' learning styles associated with the BE survey as compared to the PEPS instrument. The value of the F-statistic, 4.96, is statistically significant (a = .0001) and indicates that the model explains a large proportion of the relative variance of accounting students' test performance.

Several learning style variables contributed significantly to the accounting students' academic achievement. Listed in descending order of importance, the positive coefficients for the persistence and auditory variables indicated that students performed better when they started and finished one project at a time, and from listening to the professor's lectures. The negative coefficient for tactile and/or kinesthetic variables indicated that students did not prefer to learn accounting by actually creating balance sheets or other financial statements. The negative coefficients for the late-morning and temperature variables indicate that students perform better learning either earlier or later in the day and in a cooler environment. The negative coefficient for the visual-text variable indicates that students' achievement is negatively affected by relying on written notes on the board or reading chapters in the textbook. The negative coefficient of the variety variable, ranked tenth, indicates that students prefer a routine teaching method.

With regard to the control variables, gender was significant ([alpha] =.05) and ranked eighth in order of importance, indicating that the male students performed significantly better than females in the accounting course. All of the coefficients of the three test dummy variables were positive and significant. Students performed statistically better on the first exam ([alpha] =.01), second exam ([alpha] =.05) and third exam ([alpha] =.01) relative to the last test. This result may possibly be attributed a more rigorous final exam or to the students' additional burden of having to prepare for and take several final exams in a short period of time at the end of the semester.

The restricted least squares regression method was used to determine which categories of learning styles contributed most significantly to the accounting and economics students' academic achievement. The joint F-test suggested by Pindyck and Rubinfeld (1998) was utilized to determine if each of the six learning style categories had a statistically significant influence on student achievement. The results of the joint F-tests are reported for the economics and accounting students in Tables 3 and 4, respectively.

The joint F-test consists of estimating the unrestricted regression model described by equation (1), retrieving the unrestricted sum of squared residuals (SSRUR), then partitioning the unrestricted model into the six learning style categories (see 'Learning Styles' sub-heading) and one category for the group of binary variables, excluding a category, estimating the resultant restricted model and retrieving the restricted sum of squared residuals (SS[R.sub.R]). If the excluded category does not contribute significantly to student achievement, then the difference between the SS[R.sub.R] and SS[R.sub.UR] is negligible. If the excluded category does significantly contribute to student achievement, then the difference between the SSRR and SSRUR is significant and the hypothesis that the coefficients of the independent variables that comprise the excluded category are equal to zero is rejected.

The results of the joint F-test for the economics students indicate that the environmental and physiological categories have the most important and significant ([alpha] = .01) influence on student achievement, followed by the perceptual category, significant at the five-percent level. The emotional and sociological categories were significant at ten percent level, and the psychological category did not have a significant impact on student achievement.

The results of the joint F-tests for the accounting students indicate that four of the six learning style categories, the psychological, physiological, environmental and sociological categories, were significant at the one-percent level. The perceptual category was significantly related to student achievement at the ten percent level, and the emotional category was not quite significant at that level.

Comparison of Business Students' Learning Styles

The comparison of students' learning styles in alternative business disciplines by Stout and Ruble (1991) indicated that accounting majors exhibited learning styles that did not differ from other business majors. The studies by Loo (2002) and Baker, Simon and Bazeli (1986) concluded that the learning styles of accounting majors differed from other business majors. The results of this analysis indicate that the learning style preferences and learning style categories that contribute significantly to academic achievement for the accounting students are similar in certain ways, but differ in meaningful ways, from the economics students.

Considering first the evidence regarding the learning style categories, both the environmental and physiological learning style categories significantly contributed to student achievement for both the accounting and economics students. This result suggests that surveying students prior to the beginning of the semester to determine their preference for the time of day that the course is scheduled in addition to their preferences regarding the classroom's environment (i.e., the sound level, temperature, lighting and seating arrangements) and then accommodating those student preferences may enhance student achievement in both disciplines.

The emotional category was significant for the economics students but not for the accounting students, and the psychological category was significant for the accounting students but not for the economics students. These results suggest that it may benefit accounting students, but not economics students, to accommodate their preferences for learning style preferences that distinguish the global, analytic or integrated processing styles, including background sound, light intensity, seating arrangement, persistence and snacking in class. Similarly, it may benefit economics students, but not necessarily the accounting students, to address students' preferences for motivation, conformity and internally versus externally imposed structure.

Comparison of students' learning style preferences reveals that, although the persistence, tactile/kinesthetic and auditory learning style preferences contributed significantly to student achievement in both the economics and accounting courses, the direction of the correlation was diametrically opposed for all three preferences. This result indicates that accommodating students' preference for persistence may enhance student achievement in both economics and accounting. However, the positive sign of the coefficient for the accounting students and the negative sign for the economic students suggest that the pedagogical methods utilized to accommodate students' preference in this case must differ.

Similarly, the tactile/kinesthetic variable was significant and highly ranked for students in both disciplines, but the opposite signs indicate that the accommodation for this preference must differ between economics and accounting. The positive sign of the coefficient for the economics students indicates that utilization of more hands-on learning teaching approach may enhance student achievement, whereas the negative sign of the coefficient for the accounting students indicates a less hands-on teaching approach may enhance achievement.

The auditory variable also significantly contributed to achievement for economics and accounting students, but in opposite ways. The negative sign of the coefficient for the economics students and the positive sign for the accounting students indicate that the accounting students' academic achievement may be enhanced with greater utilization of lectures as a pedagogical tool, and less use of lectures may enhance economics students' achievement.

PEDAGOGICAL ACCOMODATION OF STUDENTS' DIVERSE LEARNING STYLES

The negative coefficient of the persistence variable indicates that, for the introductory macroeconomics students, achievement may be enhanced if the information is presented so that each lesson focuses on an individual component of a multidimensional model, instead of analyzing the entire, integrated model. For example, rather than explaining in a single lesson how the price level in the economy is determined by the forces of aggregate demand and aggregate supply, it would be more effective to separately analyze the individual components of the aggregate demand and aggregate supply curves in a series of shorter class sessions.

The initial sessions would focus on the determinants of aggregate demand, including the levels of expenditures by consumers, businesses, government, exporters and importers. These sessions would be punctuated with "side-bar" material such as a review of major historical events, how the components of aggregate demand and aggregate supply respond to such events, the process by which such major events influence the nation's domestic policies and the specific difficulties that policy makers encounter in formulating policies to counteract such events. The following sessions would explain how the slope of the aggregate demand curve is determined by the wealth, interest rate, and the exchange rate effects. The focus would then shift to an explanation of how the aggregate demand curve shifts in response to changes in the various components. The same pedagogical approach would then be applied to analyze aggregate supply. Finally, by combining the insights gained by the aggregate demand and aggregate supply lessons, the students presumably would gain a better understanding of the behavior of the price level. Throughout the analysis, the time spent on one topic should be relatively short, and students should be allowed to take frequent breaks and to work on several topics simultaneously.

The positive coefficient of the tactual/kinesthetic variable indicates that student achievement in economics may be enhanced if students are given the opportunity to move while learning and to manipulate learning resources (e.g., to transfer what they are learning to another medium) and to participate in real-life activities. One potential teaching method that utilizes all three of these learning strategies is the team learning method. For example, the team learning method could be used to teach students about the Great Depression of 1929 and the Great Recession of 2007-2008. First, the class would be divided into teams of students and the teams would physically move their seats to form circles. This introduces movement into the lesson. The instructor would explain that the task for each team is to identify at least three important causes of the Great Depression, identify any parallels to the Great Recession of 2007-2008 and make policy recommendations to stimulate employment and economic growth. Students would be instructed to use their textbooks, the internet or any other available source of pertinent information. Each group must record the information and then present it to the class. The instructor should roam the classroom to keep students on task and to answer any questions. (This introduces externally imposed structure to the task and is consistent with the positive and significant preference for authority by the economics students). The instructor then divides the chalk board so that each team can summarize its findings and recommendations. Finally, students are instructed to grade each team's effort. Utilization of the team teaching method provides students an opportunity to manually manipulate learning resources, to be mobile within the classroom, to participate in a discussion regarding an important contemporary issue and allows for active participation by students in the assessment of what they learned.

The negative coefficient of the auditory variable indicates that student achievement in economics may be enhanced if the instructor utilizes an alternative to lectures in conveying information. Utilizing written material is ruled out by the insignificance of the visual-text coefficient; it was ranked last in terms of importance. Likewise, the negative and significant coefficient of the visual-picture variable indicates that students are averse to graphical analysis. However, the negative and significant coefficient associated with the tactual/kinesthetic variable indicates that the team learning approach described above is congruent with the students' preferred perceptual strengths and may be a viable, superior alternative to the lecture format.

With respect to strategies to adapt teaching methods for this sample of accounting students, a specific teaching method would utilize, for example, a Financial Analysis Project (Project). During the semester, students learn various financial ratios, which are applied to income statements and balance sheets. About halfway through the semester students would be told to select a company of interest to them and analyze the financial health of that firm by applying the financial ratios to that company's income statement and balance sheet for the current year and preceding year.

The Project is a long-term assignment that begins at the mid-point of the course and is due on the last day of class. The Project draws on the students' cumulative knowledge from the first half of the course and requires students to apply financial ratios to a comparative income statement and balance sheet. Students must determine the significance of improving or worsening financial trends. Students are also instructed to think of questions they would ask company management based on the results of the financial analysis. The Project, which requires the students to invest a considerable amount of time thinking and strategizing, is a teaching method that takes advantage of the students' positive preference for persistence.

In order to understand the purpose and usefulness of the income statement and balance sheet, students will be confronted with numerous technical materials and concepts. The instructor, recognizing that these accounting students have a preference for an auditory learning style, would rely on a lecture and discussion format as an effective approach to communicate key concepts to students. The instructor develops the lecture so it directly relates to the learning objective and generates class discussion. Students' responses to questions posed by the lecture would serve as a very important signal as to whether the instructor is getting through.

The instructor would also look for ways to reinforce the content of the lecture. An effective way of motivating the students to continue to think about these concepts is to assign homework exercises and problems. However, the negative coefficient for the tactile/kinesthetic variable indicates that student achievement may not be effectively enhanced by such homework exercises and problems. One possible approach in overcoming this difficulty would be to reserve lecture time to discuss the purpose of the homework in terms of what is to be learned, including detailed discussions of the approaches necessary to complete the exercises and solve the problems, and describe the problems or exercises as representative of test questions that will appear on the upcoming exam. In addition, the assignments could be described as a bridge leading to the content of the next lecture. The strategy of assigning homework, if used in connection with the lecture as explained here, capitalizes on the learning style preferences for persistence and auditory learning.

CONCLUSION

An instructor can assess whether a student has or has not learned the material covered in any discipline by administering a test. However, the student's test score, high or low, does not by itself reveal the factors that contributed to the student's performance. Test scores by

themselves do not indicate what factors were beneficial to student achievement and what factors were detrimental. Test scores alone give no indication of whether the teaching method utilized by the instructor or whether the unique characteristics of the student's learning styles contributed negatively or positively to student achievement.

Instructors may adjust their teaching method in response to students' test scores to enhance their students' academic achievement, but this trial and error method of designing an effective teaching method would likely be ephemeral, changing with every new class of students. A teaching method discovered by trial and error that is effective with one group of students may not be effective for another group.

A feasible alternative to the trial and error approach to designing an effective teaching strategy is to first identify the students' learning style characteristics and then determine how the alternative learning style characteristics affect students' academic achievement. The instructor subsequently could utilize teaching methods that correspond to the learning style characteristics that correlate positively with student achievement and avoid utilizing teaching methods that negatively correlate with students' academic achievement.

The results of this study support the hypothesis that college students' academic achievement correlates significantly with their learning style characteristics in both economics and accounting courses. Moreover, in a number of important respects, the learning styles that characterized the economics students were diametrically opposed to the learning styles that characterized the accounting students. The implication of these findings is that a one-size-fits-all approach to teaching is inefficient, and that alternative teaching methods must be designed to comport with the learning style characteristics of students in alternative disciplines to effectively promote students' academic achievement.

An important conclusion that may be drawn from this study is that, due to differences in learning style characteristics that exist between students in alternative disciplines, what may be a productive method of instruction for students in one discipline may be counter-productive if used in an alternative discipline. Therefore, for any field of study, teaching proficiency and student achievement may be increased by identifying students' learning style preferences prior to the beginning of the semester and then tailoring pedagogy to closely match students' preferences. Alternatively, once the students' learning styles are identified, the students then can be assigned to an instructor who utilizes a congruent teaching style. Matching learning styles with congruous teaching methods would improve the efficiency of pedagogical resources utilized by students, instructors and administrators alike. This paper illustrated how teaching strategies may need to be adapted to reflect the particular learning styles profile of students in a given class and provided different pedagogical prescriptions for the students in the introductory economics and introductory accounting classes.

LIMITATIONS OF THE STUDY

Several possible limitations of this study should be noted. The sample size of economics and accounting students was twenty-seven and thirty-four, respectively. As a rule, the generalizability of empirical results may be limited with such small sample sizes. However, the basic reason the samples were small is that the dependent variable, exam scores, was collected for two different classes of students. Many colleges aim to maintain such small class sizes. Although a class taught in a large lecture hall format with 400 students would afford a much larger sample size, such an environment would change the educational environment. Empirical results flowing from that environment would not necessarily apply to a traditional class of thirty students.

Another possible limitation to the present study is that the dataset which was made available to the authors did not include data for (a) student ability (e.g., SAT scores or student GPA), (b) student time management issues (such as hours per week a student works in the labor market or hours per week a student devotes to the course at hand) or (c) student behavioral choices (such as the pattern of alcohol consumption and hours per week of fitness activities). Previous research by the present authors has examined the influence of these types of variables on student performance and found them to be statistically significant. It is possible that the absence of such variables in the present study has contributed to omitted variable bias.

RECOMMENDATIONS FOR FUTURE RESEARCH

The research of this study (as well as much of the earlier learning style related research presented by the present authors and others listed below) has focused on the impact of learning style preferences on the grade performance of students. It may be hypothesized that a student's learning style profile also influences other academic outcomes such as a student's choice of major and the success that a student with a given learning style pattern has in a given major (as measured by grades, whether the student graduates and a student's longevity in his/her initial choice of major). The authors are currently pursuing the data for such a study.

REFERENCES

Baker, R. E., Simon, J. R. & Bazeli, F. P., (1986). An assessment of the learning Style preferences of accounting majors. Issues in Accounting Education, 1(1), 1-12.

Becker, W. E., & Watts, M. (2001). Teaching economics in the 21st century: still chalk-and-talk. American Economic Review, 91, 446-451.

Coffield, F., Moseley, D., Hall, E., & Ecclestone, K. (2004). Learning styles and pedagogy in post-16 learning: A systematic and critical review. London: The Learning and Skills Research Center, www.lsda.org.uk/files/ PDF/1543.pdf, (accessed July 12, 2007).

Dunn, R. (2000). Capitalizing on college students' learning styles: theory, practice, and research. In R. Dunn and S. A. Griggs (Eds.), Practical approaches to using learning styles in higher education (pp 3-18). Westport, CT: Bergin & Garvey, Inc.

Dunn, R., Dunn, K, & Price, G. E., (2006). Productivity environmental preference survey. Lawrence KS: Price Systems, Inc.

Durden, G., & L. Ellis, (2003). Is class attendance a proxy variable for student motivation in economics classes? An empirical analysis. International Social Science Review, 78, 4246.

Hawk, T. F., & Shaw, A. J. (2007). Using learning style instruments to enhance student learning. Decision Sciences Journal of Innovative Education, 5, 1-19.

Krohn, G. & C. O'Connor. (2005).Student effort and performance over the semester. Journal of Economic Education, 36, 3-28.

Loo, R, (2002). A meta-analytic examination of Kolb's learning style preferences among business majors. Journal of Education for Business Majors, 77, 252-256.

Pindyck, R. S., & Rubinfeld, D. L. (1998). Econometric models and economic forecasts (4th ed.). New York City: McGraw-Hill Irwin.

Rundle, S., & Dunn, R. (1996-2009a). Building Excellence[R] (BE) Survey. www.learningstyles.net.

Rundle, S., & Dunn, R. (1996-2009b). Building Excellence (BE) Survey 2000 Research Manual. http://www.asb.dk/fileexplorer/fetchfile.aspx?file=7783.

Stout, D., & Ruble, T. (1991). A reexamination of accounting student learning styles. Journal of Accounting Education, 9, 341-354.

Terregrossa, R., Englander, F., & Wang, Z. (2009a). Why learning styles matter for student achievement in college economics. Journal for Economic Educators, 9 (1), 16-31. www.mtsu.edu/~jee.

Englander, F., Terregrossa, R., & Wang, Z. (2009b). Student behavioral choices influencing performance in introductory microeconomics. International Journal of Education Research, 4 (2), 129-143.

About the Authors:

Ralph A. Terregrossa is an Associate Professor of Economics at St. John's University in New York City. He holds a Ph.D. in economics from Binghamton University. He has published articles in The Quarterly Review of Economics and Finance, International Advances in Economic Research and Educational Review.

Fred Englander is a Professor of Economics at Fairleigh Dickinson University in Madison, New Jersey. He received his Ph.D. in economics from Rutgers University. He has published articles in the Southern Economics Journal, the Business Ethics Quarterly, Science and Engineering Ethics and the Journal of Education for Business.

Zhaobo Wang is an Associate Professor of Production and Operations Management at Fairleigh Dickinson University, Madison, New Jersey. He received a Ph.D. in operations research from Rutgers University. He has published articles in the Journal of Educational and Behavioral Statistics and the Journal for Economic Educators.

Theodore Wielkopolski is an Adjunct Professor of Accounting at Fairleigh Dickinson University, Madison, New Jersey. He is a Certified Public Accountant with twelve years experience in a public accounting firm plus seventeen years experience as an accounting executive in the manufacturing, construction and real estate industries.

Ralph A. Terregrossa

St. John's University

Fred Englander

Zhaobo Wang

Theodore Wielkopolski

Fairleigh Dickinson University

Table 1

Economics Student Achievement Regression Results

                                                         Pr >
                      Parameter   Standard               [absolute
Variable              Estimate    Error      t-Value     value of t]

Intercept               84.69        4.56      18.58        <.0001
Tactile/Kinesthetic      0.72        0.20       3.61         0.001
Intake                   0.52        0.17       3.09         0.003
Mobility                -0.43        0.15      -2.91         0.005
Light                    0.41        0.12       3.40         0.001
Alone/Pairs             -0.56        0.29      -1.94         0.057
Persistence             -0.50        0.18      -2.75         0.008
Authority                0.48        0.18       2.73         0.008
Visual-Picture          -0.45        0.20      -2.26         0.028
Auditory                -0.34        0.13      -2.50         0.015
Motivation               0.33        0.24       1.36         0.180
Variety                 -0.34        0.27      -1.26         0.212
Informal Seating         0.22        0.09       2.42         0.019
Test Dummy1             -6.41        2.69      -2.38         0.021
Gender                   6.00        4.55       1.32         0.192
Test Dummy2             -5.44        2.69      -2.02         0.048
Late Morning             0.10        0.16       0.62         0.536
Structure                0.06        0.13       0.46         0.647
Sound                   -0.05        0.12      -0.46         0.646
Conforming               0.05        0.14       0.39         0.697
Temperature              0.03        0.16       0.18         0.856
Visual-Text              0.02        0.09       0.22         0.829

F-statistic = 4.69    P-Value = .0001        [R.sup.2] = 0.6252

Standard Error of Estimate = 9.893             Coefficient of
                                               Variation = 12.865

                      Standardized
Variable              Estimate       VIF Score

Intercept                    0            0
Tactile/Kinesthetic       0.69         5.79
Intake                    0.61         6.16
Mobility                 -0.53         5.20
Light                     0.48         3.17
Alone/Pairs              -0.48         9.72
Persistence              -0.47         4.61
Authority                 0.46         4.53
Visual-Picture           -0.39         4.81
Auditory                 -0.36         3.23
Motivation                0.32         8.73
Variety                  -0.25         6.23
Informal Seating          0.24         1.51
Test Dummy1              -0.22         1.33
Gender                    0.20         3.58
Test Dummy2              -0.19         1.33
Late Morning              0.08         2.87
Structure                 0.08         4.71
Sound                    -0.07         3.79
Conforming                0.05         3.05
Temperature               0.03         3.98
Visual-Text               0.03         2.22

F-statistic = 4.69    Adj. [R.sup.2] = 0.4918

Table 2

Accounting Student Achievement Regression Results

                                                          Pr >
                      Parameter   Standard                [absolute
Variable              Estimate    Error      t-Value      value of t]

Intercept               72.99       7.56        9.65        <.0001
Persistence              0.84       0.18        4.76        <.0001
Test Dummy1             20.71       3.59        5.77        <.0001
Tactile/Kinesthetic     -0.58       0.23       -2.53         0.013
Auditory                 0.33       0.17        1.97         0.052
Late Morning            -0.57       0.15       -3.72         0.000
Temperature             -0.34       0.16       -2.11         0.038
Visual-Text             -0.30       0.18       -1.67         0.097
Gender                   9.34       4.27        2.19         0.031
Test Dummy3             10.41       3.59        2.9          0.005
Variety                 -0.43       0.21       -2.09         0.039
Authority                0.28       0.21        1.34         0.184
Test Dummy2              7.88       3.59        2.2          0.030
Alone/Pairs              0.40       0.19        2.07         0.040
Visual-Picture           0.26       0.18        1.43         0.156
Intake                   0.14       0.09        1.55         0.125
Conforming               0.18       0.17        1.04         0.301
Structure               -0.10       0.12       -0.82         0.414
Motivation              -0.14       0.24       -0.57         0.570
Informal Seating         0.13       0.17        0.76         0.447
Mobility                -0.08       0.13       -0.61         0.542
Sound                   -0.06       0.11       -0.53         0.596
Light                   -0.06       0.10        -0.6         0.548

F-statistic = 4.96    P-Value = .0001        [R.sup.2] = 0.4914

Standard Error of Estimate = 14.798          Coefficient of
                                             Variation = 18.896

                      Standardized
Variable              Estimate       VIF Score

Intercept                    0            0
Persistence               0.63         3.93
Test Dummy1               0.47         1.50
Tactile/Kinesthetic      -0.36         4.61
Auditory                  0.35         7.09
Late Morning             -0.35         1.97
Temperature              -0.28         4.05
Visual-Text              -0.26         5.35
Gender                    0.24         2.68
Test Dummy3               0.24         1.50
Variety                  -0.22         2.57
Authority                 0.19         4.47
Test Dummy2               0.18         1.50
Alone/Pairs               0.17         1.52
Visual-Picture            0.15         2.59
Intake                    0.15         2.00
Conforming                0.12         2.84
Structure                -0.08         2.24
Motivation               -0.08         4.62
Informal Seating          0.07         1.87
Mobility                 -0.07         2.79
Sound                    -0.06         2.82
Light                    -0.06         1.97

F-statistic = 4.96    Adj. [R.sup.2] = 0.3923

Table 3

Joint F- Test of the Learning Styles Categories in Economics

                               Degrees of
Strand          Restrictions    Freedom     F-Statistic   P-Value

Environmental        4            4,60         5.66       0.0006
Physiological        6            6,60         4.71       0.0005
Dummies              3            3,60         4.15       0.0098
Perceptual           5            5,60         2.67       0.0305
Emotional            4            4,60         2.49       0.0525
Sociological         5            5,60         1.98       0.0948
Psychological        2            2,60         0.91       0.4088

Notes: 1. Joint [F.sub.q,n-k] = [(SS[R.sub.R] -SS[R.sub.UR])/q]/
SS[R.sub.R] /n-k. 2. Restrictions equal the number of number of
learning style variables in each category. 3. The degrees of freedom
(DF) in the numerator and denominator equal the number of
restrictions and the number of observations (81) less the number of
parameters (21), respectively.

Table 4

Joint F-Tests of the Learning Styles Categories in Accounting

                               Degrees of
Strand          Restrictions    Freedom     F-Statistic   P-Value

Dummies              4           4,114         11.01      1.38E-07
Psychological        2           2,114          6.19      1.21E-05
Physiological        6           6,114         12.05      1.8E-05
Environmental        4           4,114          6.05      0.00018
Sociological         5           5,114          5.20      0.00024
Perceptual           5           5,114          1.96      0.08961
Emotional            4           4,114          1.91      0.11278

Notes: 1. Joint [F.sub.q,n-k] =[(SS[R.sub.R] -SS[R.sub.UR])/q]/
SS[R.sub.R]/n-k. 2. The restrictions equal the number of number of
learning style variables in each category. 3. The Joint F-test degrees
of freedom (DF) in the numerator and denominator equal the number of
restrictions and the number of observations (136) less the number of
parameters (22), respectively.