Corrigendum

The present meta-analysis is different from Phelps (2012) in at least four ways. First, Phelps (2012) conducted an extensive review of quantitative, survey, and qualitative studies on testing effects published up to 2010 and provided results separately for each, while our study focuses on quantitative studies published up to 2015. Indeed, of the 282 independent effect sizes that we extracted and included in our meta-analysis, 84 were published between 2011 and 2015. Second, Phelps (2012) examined both low-stakes and high-stakes testing effects while our meta-analysis only examined the effects of low-stakes practice tests. Third, our meta-analysis reported key methodological information such as study selection criteria, index of inter-rater reliability (Kappa) and presented weighted mean effect sizes with concomitant statistics such as standard errors, 95% lower and upper confidence intervals, and tests of heterogeneity (e.g., Q and I² statistics). Researchers have recommended that such information be reported as they help readers understand and interpret the precision and heterogeneity of the effect sizes so as to provide more reliable and valid inferences from meta-analyses (Ahn, Ames, & Myers, 2012; Borenstein, Hedges, Higgins, & Rothstein, 2009; Harwell & Maeda, 2008). Fourth, while we acknowledge the extensive review and report of several moderators by Phelps, our meta-analysis examined several moderators not reported by Phelps. These moderators include, but are not limited to, comparison treatment, practice and final test format, transfer appropriate processing, study settings, measure reliability, prior knowledge, treatment fidelity.