Engaging in a test over previously studied information can serve as a potent learning event, a phenomenon referred to as the testing effect. Despite a surge of research in the past decade, existing theories have not yet provided a cohesive account of testing phenomena. The present study uses meta-analysis to examine the effects of testing versus restudy on retention. Key results indicate support for the role of effortful processing as a contributor to the testing effect, with initial recall tests yielding larger testing benefits than recognition tests. Limited support was found for existing theoretical accounts attributing the testing effect to enhanced semantic elaboration, indicating that consideration of alternative mechanisms is warranted in explaining testing effects. Future theoretical accounts of the testing effect may benefit from consideration of episodic and contextually derived contributions to retention resulting from memory retrieval. Additionally, the bifurcation model of the testing effect is considered as a viable framework from which to characterize the patterns of results present across the literature. (PsycINFO Database Record (c) 2014 APA, all rights reserved).