Abstract

As the most extensively researched type of formulaic sequence in SLA, collocation is regarded as essential for proficient L2 use (Sonbul & Siyanova-Chanturia, 2024). However, given their vast number and the limited time available for classroom instruction, teaching collocations through word-focused activities is not always feasible. Consequently, a number of studies have investigated the effectiveness of incidental collocation learning. While many studies reveal that exposure to meaning-focused input—such as reading stories, listening to music, and viewing television—can yield substantial learning gains, others have reported only limited gains, arguing that collocations often go unnoticed due to their high semantic transparency and low perceptual salience (e.g., Szukalski, 2012).

Given these mixed findings, the present meta-analysis, as the first of its kind, aims to examine the overall effectiveness of incidental collocation learning and identify moderators that may account for variability in learning gains across studies. A three-level meta-analytic model was fitted to a dataset of 115 effect sizes extracted from 28 eligible studies. The results revealed a statistically significant, large effect (g = 1.26, p < .0001). This finding is comparable to that reported in Webb et al.’s (2023) meta-analysis on the effectiveness of incidental single-word learning (g = 1.14), suggesting that meaning-focused treatments for collocation learning may yield greater effectiveness than previously assumed. Subsequent moderator analyses examined 14 variables and identified five significant sources of heterogeneity: L1 congruency, frequency of occurrence, test format, L2 proficiency, and treatment duration. Specifically, larger effect sizes were observed in studies targeting congruent collocations, providing more encounters with target items, employing form-recognition tests, involving novice learners, and implementing treatments over multiple days. These findings underscore the importance of considering learner-, treatment-, and methodology-related factors when evaluating the effectiveness of incidental collocation learning. The study concludes with evidence-based pedagogical implications for optimizing incidental collocation learning.