Retrieval enhanced learning: Does it matter what topic is to be learned?

Testing (i.e., the retrieval of information from memory) has been shown to improve retention of learning material in lab-based studies, classrooms, and in an array of medical education contexts. However, the magnitude of its effects are variable without full clarity regarding when or why stronger effects are found. This study explores the under-examined relationship between topic of study and the pedagogical benefits of testing across three research projects. Knowledge retention over a one-week interval was compared for those who undertook a test during learning vs. those who engaged in extended study in 3 separate experiments: (1) Renal and gastrointestinal materials, studied by N = 32 first year medical students; (2) Gastrointestinal materials presented with more or less effort towards integrating content, studied by N = 39 first year medical students; (3) Rheumatological materials that presented pathophysiology and clinical material in either integrated or separated fashion, studied by N = 109 undergraduate biology students. Across 8 distinct examinations of the mnemonic benefits of retrieval (with vs. without integration, with medical students or biology students, using a recall or a diagnostic accuracy outcome measure) test-enhanced learning effects were observed except for when gastrointestinal learning materials were studied. Various exploratory analyses were unable to account for why differences in topic were associated with differences in effect magnitude. While the literature exploring the effects of testing is justified in concluding it to be pedagogically beneficial, puzzles remain regarding the magnitude of effect that can be expected when learners study different topics. At worst it appears testing is innocuous, so its general advantage for learning justifies its continued use even while further research is needed to explore the extent of factors that influence retrieval-enhanced learning before best practices can be fully specified.