A meta-analysis of the reliability of second language reading comprehension assessment tools

Abstract

The present study aims to meta-analyze the reliability of second language (L2) reading assessments and identify the potential moderators of reliability in L2 reading comprehension tests. We examined 3,247 individual studies for possible inclusion and assessed 353 studies as eligible for the inclusion criteria. Of these, we extracted 150 Cronbach’s alpha estimates from 113 eligible studies (years 1998–2024) that reported Cronbach’s alpha coefficients properly and coded 27 potential predictors comprising of the characteristics of the study, the test, and test takers. We subsequently conducted a reliability generalization (RG) meta-analysis to compute the average reliability coefficient of L2 reading comprehension tests and identify potential moderators from 27 coded predictor variables. The RG meta-analysis found an average reliability of 0.79 (95% CI [0.78, 0.81]). The number of test items, test piloting, test takers’ educational institution, study design, and testing mode were found to respectively explain 16.76%, 5.92%, 4.91%, 2.58%, and 1.36% of variance in reliability coefficients. The implications of this study and future directions are further discussed.