A Meta-Analysis of the Effectiveness of Genre-Based Writing Instruction in English as Foreign Language

One of the recent developments in the teaching of writing to second language learners is the genre-based approach which follows the more traditional product- and process-based approaches. Experimental and quasi-experimental studies on the effects of genre-based instruction (GBI) on writing in English as a foreign language (EFL) have been frequently reported in the last few decades. Findings of these studies are inconclusive and inconsistent. Past research indicates that different populations of language learners may benefit differently from this type of instruction. The present meta-analytic study followed two a) to explore the average effect of GBI on EFL writing based on both experimental and quasi-experimental studies of the past 26 years and b) to investigate how the variables of writing task type, educational level, and first language possibly moderate the effect of GBI on EFL writing. Four research questions were addressed: a) what is the average effect size of studies on GBI in EFL writing? b) Do EFL writers at primary, secondary, and tertiary levels benefit differently from GBI? C) Does GBI affect different task differently?, and d) Do speakers of different mother tongues benefit differently from GBI? In this meta-analysis, a total number of 255 studies thoroughly searched and collected from various academic websites constituted the original population of studies intended for analysis. Based on careful exclusion and inclusion criteria, 26 (quasi)experimental studies providing 28 effect sizes were included as the final sample. The studies were coded for effect size data and moderator variable indices. The analyses were performed using CMA software. The results of the analyses showed that (1) GBI has a small average effect on EFL writing (Cohen’s d= 0.298); (2) GBI significantly more effective for primary level participants; (3) GBI produced different effects for different task types producing significantly higher effects for paragraph writing, (4) GBI was significantly more effective for learners whose first language was Japanese. The findings of this study can be considered as beneficial issue for both pedagogical and research purposes. The results imply that EFL writing instructors should consider possible moderating variables when choosing their method of writing instruction.