Recent Research in Science Teaching and Learning

Deborah Allen
2014 CBE - Life Sciences Education  
Feature Current Insights and included majors and nonmajors as participants. The frequency of use and types of active-learning methodologies described in the 225 eligible studies varied widely. Quantitative analysis of the eligible studies focused on comparison of two outcome variables: 1) scores on identical or formally equivalent examinations and 2) failure rates (receipt of a "D" or "F" grade or withdrawal from the course). Major findings were that student performance on exams and other
more » ... ments (such as concept inventories) was nearly half an SD higher in active-learning versus lecture courses, with an effect size (standardized mean weighted difference) of 0.47. Analyses also revealed that average failure rates were 55% higher for students in the lecture courses than in courses with active learning. Heterogeneity analyses indicated that 1) there were no statistically significant differences in outcomes with respect to disciplines; 2) effect sizes were lower when instructor-generated exams were used versus concept inventories with both types of courses (perhaps because concept inventories tend to require more higher-order thinking skills); 3) effect sizes were not significantly different in nonmajors versus majors courses or in lower versus upper-division courses; and 4) although active learning had the greatest positive effect in smaller-enrollment courses, effect sizes were higher with active learning at all enrollment sizes. Two types of analyses, calculation of fail-safe numbers and funnel plots, supported a lack of publication bias (tendency to not publish studies with low effect sizes). Finally, the authors demonstrated that there were no statistically significant differences in effect sizes despite variation in the quality of the controls on instructor and student equivalence, supporting the important conclusion that the differences in effectiveness between the two methods were not instructor dependent. In one of the more compelling sections of this meta-analysis, the authors translated the relatively dry numbers resulting from statistical comparisons to potential impacts on the lives of the students taking STEM courses. For example, for the 29,300 students reported for the lecture treatments across all students, the average difference in failure rates (21.8% in active learning vs. 33.8% with lecture) suggests that 3516 fewer students would have failed if enrolled in an active-learning course. This and other implications for the more beneficial impact of active learning on STEM students led the authors to state, "If the experiments analyzed here had been conducted as randomized controlled trials of medical interventions, they may have been stopped for benefit." That is, the control group condition would have been halted
doi:10.1187/cbe.14-09-0147 pmid:25452480 pmcid:PMC4255344 fatcat:s2qwgcynjzgzdkfvrxgng2t3ce