Comparing retrieval-based practice and peer instruction in physics learning

Even though it has been shown that retrieval practice could foster deeper learning and better long-term retention in other domains such as psychology, it is rarely studied in the context of physics learning where students need to solve more complicated problems. To an even lesser degree is comparing retrieval-based learning with other active learning style methods adopted in the physics classroom. In this study, we compared the effects of retrieval-based learning and peer instruction based restudying on physics problem solving and transfer. In both conditions (retrieval and peer instruction), participants were first presented video lectures explaining the definition of speed and energy conservation. In the training session that immediately followed the video lectures, the retrieval condition was asked to recognize, recall, and apply the relevant physics concepts to solving problems; while the peer instruction condition was asked to discuss the two video lectures with each other in a group of three or four members. After this training, an immediate and a delayed (one-week) final test were administered. Both tests contained an initial task, isomorphic to the training materials, near transfer tasks, and far transfer tasks for each topic. Subjective judgment of learning (JOL) was collected immediately after watching the video lectures, after the training, and right before the delayed final test. We found a retention advantage of the retrieval practice for the initial tasks, but not for the near and far transfer tasks. We also found an advantage of the retrieval practice on far transfer on the immediate final test. Peer instruction training inflated the participants' JOL compared to the retrieval practice. Both JOL scores were significantly lower than after the video lecture but did not differ from before the delayed final test.