The Effects of Learning Styles on Course Performance: A Quantile Regression Analysis

Abstract

ABSTRACT In this study, we investigated the relationship between learning styles and student performance for students enrolled in a basic business statistics course. We performed this study in order to assess our success in creating a course that facilitated learning for students regardless of learning style. To accomplish this analysis, we utilized quantile regression, a statistical technique from the economics literature developed by Koenker and Bassett (1978) that supplements the classical ordinary least squares regression. Quantile regression allowed investigation of a more complete picture of student performance over the entire population as opposed to the mean effect from least squares regression. Learning style was not significant in determining a student's overall course score for the entire group of students. For small cohorts of students, learning styles were significant for exam average. That is, some students experienced either a disadvantage or advantage by their learning style for the course's exam component. However, for the overall course performance, a student with a particular learning style was neither advantaged nor disadvantaged. INTRODUCTION As faculty, one of our goals is to provide the best possible learning environment for our students. In order to create an ideal learning environment, it is important to understand our students' different learning styles (Coffield, Moseley, Hall, & Ecclestone, 2004; Dunn, Griggs, Gorman, & Beasley, 1995). preferentially focus on different types of information, tend to operate on perceived information in different ways, and achieve understanding at different rates (Felder, 1993, p. 286). Acknowledging that students have different learning styles then behooves the instructor to utilize a variety of teaching strategies in order to engage students of all learning styles (Buxeda & Moore, 1999). AU students will then have opportunities to use their preferred learning style and opportunities to improve their less-preferred learning style (Hawk & Shah, 2007). Having incorporated a variety of learner-centered and integrative teaching strategies into a business statistics course, we wanted to know if a student's learning-style had an effect on their course performance. It is important to design a course that allows students of all learning styles to succeed. As noted by Felder, Students whose learning styles are compatible with the teaching style of a course instructor tend to retain information longer, apply it more effectively, and have more positive post-course attitudes toward the subject than do their counterparts who experience learning/teaching style mismatches, (Felder, 1993, p. 286). If the results of our study had shown a relationship between the overall course score and a student's learning style, then additional components could have been designed into the course or students could have been appropriately advised about how best to adapt to the teaching style that does not match their preferred learning style (Campbell, 1991; Coffield, et al. 2004). To accomplish this analysis, we collected student performance data on the various components of the course (quizzes, exams and projects), attributes of student learning styles, achievement on pre- and post-assessment, and attendance in the course. We analyzed the data using ordinary least squares regression analysis and quantite regression (Koenker & Bassett, 1978). Quantile regression allowed investigation of a more complete picture of student performance over the entire student population distribution. For example, a least squares regression analysis for course score with learning styles as the independent variables estimated the mean effect of learning styles on course performance. Quantile regression, however, provided information about the performance of, for example, the lower performing 25% of the class. The significant factors that affected performance for the lower 25% could have been different from the significant factors that affected the performance for the top performing 25% and this difference could only be discovered using quantile regression. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 43 citations exist upstream.