CALL-Based Versus Paper-Based Glosses

Abstract

Second language (L2) reading comprehension is generally facilitated by both native language and L2 glosses. However, CALL learning opportunities are still not as common as they should be, and, therefore, the present meta-analysis contributes further evidence supporting the inclusion of more CALL experiences in reading contexts. The present study, a quantitative meta-analysis of 32 studies, observes that overall effect sizes are larger (g = .92) for computer-assisted language learning (CALL) glossing studies than for non-CALL glossing studies (g = .43), and a test of homogeneity reveals a significant difference (p < .001). Converting the effect sizes to a percentage scale, it was found that 81% of learners provided with CALL glosses perform higher than those without such glosses as opposed to 64% of L2 readers with traditional glosses. Reasons explaining the effectiveness of CALL glossing include the fast access CALL glosses provide and the flexible nature of CALL glosses. It is argued that the effect sizes for CALL studies should generally be higher than they are and, consequently, further primary studies providing more flexible textual options are needed for the L2 reader such as those provided by linked programs (e.g., iFinger).