The paper discusses an experiment in which regression analysis was used to examine predictors of information retention following a college economics course. It was hypothesized that students taught in small classes with self-instructional materials would place a higher value on basic economic concepts and retain these concepts longer than students in large, traditional classes. Variables included class size, method of instruction, pre- and posttest scores, grade in the course, class, sex, SAT scores, and student attitudes towards economics. All students were tested over four years. New models were developed to solve problems of allocating correct explanatory power to each variable to correct shortcomings of existing retention models. Findings indicated a significant positive attitude change toward economics in smaller classes. It was concluded that, in spite of positive attitudes toward economics in small classes, cognitive performance remains comparable in small and large classes. Additional research and analysis is suggested. (Author/DB)