Moderators of the Relation Between Self-Regulated Learning and Academic Performance: A Meta-Analysis

• In Event: Integrating Different Approaches to Investigating Self-Regulated Learning

Abstract

Over the past two decades, educational and developmental psychologists have become increasingly interested in the antecedents, processes, and outcomes of self-regulated learning (SRL). The breadth and diversity of this literature makes it ideal for a synthesis of research, yet the relation between SRL and academic performance across secondary school has not been explored using meta-analysis. Incorporating research across theoretical perspectives, the current meta-analysis assessed how the processes of self-regulated learning (e.g., planning, monitoring, control) correlate with academic performance for children and adolescents.

Published and unpublished research on this topic was collected through electronic database searches, listserv requests, and email correspondence with authors. A coding scheme for collecting data from research reports was developed by the first author, and two trained members of our lab independently coded each study. Coding discrepancies were resolved through discussion between the first and third authors.

Based on the data gathered from this literature search and coding procedure, 62 independent samples (N=11,867) yielded a weighted average correlation of .184 between SRL processes and academic performance. In a separate meta-analysis, this correlation was found to be significantly stronger than the relation between cognitive strategy use (e.g., elaboration, rehearsal) and academic performance (r=.137, k = 72, N =16,538). However, there was more variation around both overall correlations than chance alone could explain.

To account for this variation, four moderator analyses were conducted. The meta-analysis on SRL processes revealed that the strength of average correlations differed significantly as a function of students' grade level, the academic subject, the specific self-regulatory process, and how these constructs were measured. The developmental trajectory of self-regulated learning suggests that its relation with academic performance should become stronger across primary and secondary school as students begin to master their own learning and do so more strategically (e.g., Winne, 2001). Our moderator analysis by grade level supports this hypothesis, with significant correlations emerging early in middle school and becoming stronger during high school. Collapsing across secondary school, there was a significantly stronger correlation for science (r=.316) than for math (r=.166) or English (r=.153). Among SRL processes, the correlation with academic performance was strongest for self-evaluation (r=.206), which was significantly different from metacognitive monitoring (r=.164) and planning (r=.142). How SRL processes were assessed also moderated their relation with academic performance. The strongest correlation was for "online" measures, such as observable traces of strategy use (r=.507; Schraw, 2010). Although questionnaires comprised over two-thirds of the correlations, their relation with academic performance was significantly weaker (r=.169). Measurement error and less variation in scale scores may have attenuated the correlations for self-report measures. However, weaker correlations may instead reflect a real shortcoming in the ability of self-report measures to capture students' self-regulation of learning (Aleven et al., 2010). Implications for measuring self-regulated learning, future research exploring its relation with academic performance, and educational practice will also be discussed.

Authors

• Amy L. Dent, Duke University

• Harris M. Cooper, Duke University

• Alison C. Koenka, Virginia Commonwealth University