Effect size guidelines for individual differences in L2 learning studies: A second-order synthesis

Previous research has determined that effect sizes differ across disciplines, with discipline-specific effect size benchmarks needed to accurately describe effect sizes. Research into individual differences in second language (L2) learning is flourishing, however effect sizes guidelines are lacking. Using the established methodologies of previous studies in psychology, this second-order synthesis presents effect size benchmarks for correlation coefficients of individual differences in L2 learning. The analysis of 1719 effect sizes from 23 meta-analyses resulted in an overall 25th percentile of r = ±0.20, a 50th percentile of r = ±0.35, and a 75th percentile of r = ±0.50. In addition, using a taxonomy of individual differences categories, we distinguished between cognitive-, conative-, affective-, and personality differences. A one-way ANOVA showed significant differences between categories. Based on the quartiles, effect size guidelines were developed for each category as well as for L2 individual differences research in general.

Educational relevance statement

When researchers examine the effects that certain individual differences have on language learning, for example if certain personality traits may help or hinder someone in learning a second language (L2), they may use quantitative methods to explore these effects. When using quantitative methods, such as correlations or regressions, researchers need to report and interpret the effect sizes they find for each analysis that they do. However, interpreting effect sizes can be complex, as previous research has established that effect sizes tend to fluctuate depending on the research field. In this study, we explore effect sizes commonly found in L2 individual differences research and we develop practical guidelines that researchers may use to interpret their effect sizes in future. We also give practical advice as to why, when, and how researchers should interpret effect sizes in the field.