How Effective are Writing Interventions on Writing Outcomes, Qualities, and Attitudes? A Large-Scale Meta-Analysis for Postgraduate Theses in Türkiye

Abstract

This meta-analysis examined 137 thesis studies on writing interventions in Turkish schools. After removing outliers, overall effect sizes were calculated from 129 studies (627 test results) for writing outcomes, 103 studies for quality, and 32 studies for attitude. Robust variance estimation was used to address dependency in studies reporting multiple effect sizes. Results showed large effects on writing outcomes (g = 1.17) and quality (g = 1.43), with moderate effects on attitudes toward writing (g = 0.59). Among interventions, Self-Regulated Strategy Development, graphic organizers, pre-writing strategies, process approach, strategy teaching, text-based teaching, and various text structure instructions had larger effects. Meta-regression models revealed genre instruction studies had larger effects on quality, while studies with two different teacher implementers showed larger effects on attitude. Grade level+genre-based and pilot study+genre-based+grade level models explained 16% of variance in writing quality. These findings provide an evidence-based framework for effective writing instruction.

This meta-analysis examined 137 thesis studies on writing interventions in Turkish schools. After removing outliers, overall effect sizes were calculated from 129 studies (627 test results) for writing outcomes, 103 studies for quality, and 32 studies for attitude. Robust variance estimation was used to address dependency in studies reporting multiple effect sizes. Results showed large effects on writing outcomes (g = 1.17) and quality (g = 1.43), with moderate effects on attitudes toward writing (g = 0.59). Among interventions, Self-Regulated Strategy Development, graphic organizers, pre-writing strategies, process approach, strategy teaching, text-based teaching, and various text structure instructions had larger effects. Meta-regression models revealed genre instruction studies had larger effects on quality, while studies with two different teacher implementers showed larger effects on attitude. Grade level+genre-based and pilot study+genre-based+grade level models explained 16% of variance in writing quality. These findings provide an evidence-based framework for effective writing instruction.