The effectiveness of computer-assisted instruction in statistics education: A meta-analysis

Abstract

The purpose of this study was to investigate the effectiveness of computer-assisted instruction (CAI) in statistics education at the college level in the United States. This study employed meta-analysis to integrate the findings from 25 primary studies which met a specific set of criteria. The primary studies were selected from journal articles, ERIC documents, and dissertations. Results of the meta-analysis produced an overall effect size estimate of 0.43, indicating a small to medium positive effect of applying CAI in teaching college-level introductory statistics on students' achievement. Several study characteristics were examined for the association with the effect magnitude. These characteristics included the publication year, the publication source, the educational level of participants, the mode of the CAI program, the type of CAI program, the level of interactivity of the CAI program, the instructional role of the CAI program, and the sample size. The results of the analogous analysis of variance showed that different modes of CAI programs produced significantly different effects on students' achievement in learning statistics. Expert systems and drill-and-practice programs were the most effective modes and were followed by multimedia, tutorials, and simulations. Computational statistical packages and web-based programs were the least effective modes. The teacher-made CAI programs were significantly more effective than the commercially-developed CAI programs. The effectiveness of CAI program in teaching statistics did not differ significantly according to the study characteristics of the publication year, the publication source, the educational level of participants, the level of interactivity of CAI program, the instructional role of CAI program, and the sample size.