The effects of solution type and context on the transfer of solution to conditional probability problems for introductory undergraduate statistics students

Presenting difficult concepts to undergraduate students has been a challenge for instructors and instructional designers for some time. Conditional probabilities are a complex application for students, who tend to become confused when presented with the traditional Bayesian probability format as an initial introduction. This study compared performance on conditional probability problems after students studied either a problem solved using the traditional Bayesian probability format or when using a natural frequency format. A second variable investigated whether the example problem and the target problem presented for solution had matched or mismatched context. The results of this study support the predictions that the use of frequency formatted solution would lead to better performance in the presentation of conditional probabilities, and that having matched context in both the solved example problems and the target problems presented for solution would improve performance in solving the target problem.;While the results of the participant performance on the solved problems replicates and supports findings of other studies presented in the literature, participants struggled to transfer their learning to the subsequent multiple choice target problems. This lack of transfer suggests that the level of learning that took place was at a procedural level that did not provide a deep enough level of conceptual understanding to adequately transfer the learning to the different problem form of the multiple choice problems.