Impact of methodological differences in task sequencing on second language learning: a multi-level meta-analysis

Drawing on the SSARC model, TBLT research has investigated the role of task sequencing in L2 production, yielding mixed results. This study conducts a multi-level meta-analysis of methodological practices in task sequencing to provide more transparency. Sixteen studies published up to May 2024 were selected based on inclusion criteria, and each study was coded for analyzed features and contextual variables. Methodological aspects, including study designs, sampling procedures, analyses, and reporting practices, were also coded. Significant methodological variations were identified across studies, including differences in L2 proficiency, learning contexts, task modalities, types, complexities, delivery methods, and constructs. Task complexity emerged as a significant factor influencing L2 performance, with more complex tasks yielding larger effect sizes. Additionally, simple and complex sequencing tasks exhibited a medium effect size in enhancing accuracy, particularly in between-group studies. These findings highlight the importance of balanced learner representation and standardized research practices to improve L2 learning outcomes.