Investigating the Potential Moderating Role of Expertise on the Testing Effect with the Fictional Worlds of Harry Potter and the Marvel Cinematic Universe

Testing of information, compared to restudying, results in better long-term retention. Given the educational relevance, there is interest in the conditions and learner differences that may moderate the usefulness of testing. One possible moderator is pre-existing knowledge, which often influences learning. It is possible that the testing effect is stronger for experts because it guides new learning, stronger for novices because it helps novices catch up to experts, or so robust that testing works equally well for everyone. I explored if the testing effect was moderated by pre-existing knowledge about topics where expertise largely differs. I recruited participants who reported expertise in one of two fictional domains: either Harry Potter but not Marvel Superheroes, or vice versa. In four experiments, participants read and studied novel texts about both domains and took a multiple-choice test one week later. In Experiments 1, 2A, and 2B, study strategy (testing versus restudying) was orthogonally manipulated within-participants; in Experiment 2B only, participants also received feedback when studying via retrieval practice. Experiment 3 employed a mixed design in which each participant used only one strategy or another to address potential carryover effects from one strategy to the next. In all experiments, I found that participants with existing expertise on a given topic also better learned new information on that topic, regardless of study strategy. Retrieval practice significantly benefited retention only when participants received feedback; otherwise, I did not observe a significant testing effect. Critically, in Experiments 1, 2A, and 2B, the effects of retrieval practice (or lack thereof) were never moderated by prior knowledge, indicating that any learning benefits of testing do not depend on having high levels of pre-existing knowledge. In contrast, in Experiment 3, I did find a significant interaction: retrieval practice boosted performance for novice topics, while restudying did so for expert topics. Although the interaction is statistically significant and may be meaningful, this result is not definitive and should be interpreted with caution.