Associations of student characteristics to measures of introductory Pascal computer programming achievement for suburban community college students (abstract only)

The purpose of the study was the construction and testing of student characteristic models for predicting the performance of 162 students on several dependent measures of computer programming achievement under generic setting and treatment conditions. The dependent measures were tests for program reading and writing, lab problems, composite performance and relative course success categories. Student characteristics were measured within the categories of demography, academic background, ability and personality; ability measures included iconic pattern detection, vocabulary, analogies and algorithmic and quantitative problem solving; personality measures included self-assessments of abilities, motivations and computer attitudes, and Gittenger's cognitive style. The predictor models accounted for 59%, 56%, 27% and 64% of the variance for the reading, writing, lab and composite measures of programming achievement, respectively. The composite achievement model also placed students into course success categories with 87% accuracy. The analysis results led to the following conclusions:Different sets of student characteristics were associated with the dependent measures; the strongest linkages were quantitative ability and high school grades to reading, verbal ability, negative computer attitudes inverse and achievement motivation to writing, age, introvert/regulated cognitive style and female gender to lab, and a blend of ability and personality measures to composite performance. Ability and personality were more effective than previous academic performance in predicting course achievement. Both iconic pattern detection ability and self-assessed math ability enhanced performance. Students with basic introvert/regulated cognitive style performed significantly better than students with extrovert/flexible style. High school computer coursework, positive computer attitudes and self-assessed programming ability were all unrelated to performance. Both students who dropped and those who finished with low achievement had similar characteristics; low ability levels and negative computer attitudes were the primary predictors of poor performance.