Using Retrieval Practice, Variability, and Spacing to Facilitate Understanding of Complex Information

• In Event: Toward an Understanding of How Learning Strategies Can Be Effective for Learners Across Contexts

Abstract

Background
Many evidence-based learning strategies have been shown to facilitate learning in isolation: retrieval practice (Pan & Rickard, 2018; Rowland, 2014), variable practice (Raviv et al., 2022; Paas & Van Merriënboer, 1994), and spaced practice (Carpenter et al., 2012; Cepeda et al., 2006; 2008). However, less is known about the optimal implementation of these strategies when used in combination, even though students report using multiple learning strategies when preparing for an exam (Karpicke et al., 2009). Thus, we aim to understand the complexity behind how learning strategies can be combined during practice to facilitate meaningful learning that will transfer to new contexts (e.g., inference or application type questions). Combining learning strategies might produce greater transfer that using just one strategy, potentially producing additive effects. Alternatively, the mechanisms underlying the various strategies may be redundant or conflict with each other, thus producing subadditive effects.
Method
Two experiments (N1=75; N2=61) were conducted to investigate the effects of combining learning strategies. A 2 (Practice Activity: Retrieval Practice, Re-study) x 2 (Practice Type: Repeated, Variable) design was adopted for Experiment 1 (see Table 1). For Experiment 2, a 2 (Practice Structure: Spaced, Massed) x 2 (Practice Type: Repeated, Variable) design was adopted (see Table 2). Practice activity and practice structure were manipulated within-subjects, between concepts, and practice type was manipulated between-subjects. The materials consisted of five mini-lecture style videos on geological science that contained 12 key concepts (see Butler et al., 2017). After watching the videos, participants engaged in practice on the concepts: retrieval practice consisted of short-answer questions that required the application of knowledge from the video (i.e., transfer), whereas re-study practice involved reading a “study point“ that contained the same information as the question and answer but in paragraph form (Table 3 for examples of questions and study points for a concept). Practice was structured differently depending upon condition: repeated or variable (i.e., same question versus different question about the same concept); massed or spaced (i.e., all in one session or spaced across days).

Results
Tables 4 and 5 show performance on the practice trials and final test for Experiment 1 and Experiment 2, respectively. In Experiment 1, a 2 (Practice Type) x 2 (Practice Activity) mixed ANOVA confirmed a significant main effect of practice type, F(1, 73) = 5.95, p = .02, ηp2 = 0.08, as variable practice produced superior transfer relative to repeated practice. Neither the main effect of practice activity nor the interaction were significant. In Experiment 2, a 2 (Practice Type) x 2 (Practice Structure) mixed ANOVA revealed a significant main effect of practice structure, F(1, 59) = 12.87, p < .001, ηp2 = 0.18, as spaced practice produced superior transfer relative to massed practice. Neither the main effect for practice type nor the interaction were significant. Overall, the predicted advantage of combining learning strategies did not emerge, suggesting that using strategies in isolation may be more beneficial and more research is needed to understand how best to use strategies in combination.

Authors

• Rachel N. Smith-Peirce, Washington University in St. Louis

• Andrew C. Butler, Washington University in St. Louis