| The Soloman-Felder Inventory of Learning Styles (ILS{rcub} is a frequently used instrument for the assessment of learning styles in science and engineering. However, introductory computer science education has rarely been a focus of learning styles research. This study used the ILS, in a test-retest format, in introductory computer science courses at the University of Minnesota Duluth (UMD) as a tool to aid in the understanding of student achievement and retention. Over 300 students in a variety of classes participated in the study during fall semester, 2004. There were important findings in three areas: instrument reliability, learning style characterization, and the relationship of learning style to outcome. Instrument reliability was acceptable along most dimensions of the scale but weak along the sequential-global scale. Specific sources of concern were identified which could lead to improvement of the instrument. Profiles of the learning styles of students in each class were constructed. The visual-verbal scale was skewed to the right in each instance. Other distributions were fairly normally distributed. A significant association with gender was identified (females being less visually oriented than males). This is important because computer science has historically been characterized by low female enrollment. Relationships with outcome identified the active-reflective scale as significantly related in performance in computer programming classes. Active learners were more likely to do poorly. This is important because it may be linked to retention issues. A predictive model of student outcome success identified the active-reflective scale and ACT Composite scores as the key indicators. The study has implications for the ways in which computer science students are selected for enrollment, instructed and assessed and may be linked to larger issues of retention and gender. |
