The Effects of Collaborative Output Tasks on Second Language Grammar Learning : A Meta-Analysis

Abstract

Collaborative output tasks have been used in teaching grammar,based on Long’s interaction hypothesis and Vygotsky’s sociocultural theory. Many researchers have developed collaborative output tasks designed to intend for learners to focus on language form. Accordingly, many studies have investigated the effects of collaborative output tasks on second language acquisition; however, the findings have been inconsistent, calling for a meta-analysis of these studies. To this end, I conducted a meta-analysis to investigate the effects of collaborative output tasks, which are dictogloss, reconstruction, and text-editing tasks, on L2 grammar learning, particularly for adult learners, with the additional aim of comparing the effects of different types of output tasks. For the effect size calculation, only one effect size was used for each study, because several effect sizes could violate the assumptions of statistical analysis. JAMOVI was used to calibrate the effect sizes across studies. To determine the relative effects of the different types of collaborative output tasks, each effect size was compared by calculating the average effect size for each type of collaborative output task. Based on 16 effect sizes gathered from 12 studies representing 16 independent samples, I found that collaborative output tasks are proved to be effective in L2 grammar learning (p< .001). Further, it is revealed that the effect size for dictogloss is medium (ES = 1.38), and the effect sizes for reconstruction cloze task and editing task are large (ES = 1.62, 2.44).