This study aims to analyze the effectiveness of various learning models in improving the ability to write short stories in students. Through a meta-analysis approach, this study collects and examines the results of relevant experimental and quasi-experimental research. The data collection technique was carried out by tracing scientific publications that met the inclusion criteria, namely focusing on the learning model, involving the ability to write short stories as a dependent variable, and presenting statistical data that allows the calculation of effect size. A total of 15 articles that met the criteria were studied using Cohen's formula d to determine the magnitude of the influence of each learning model. The results showed that the Self-Directed Learning learning model as a treatment had the largest effect size, which was 4.28 and was categorized as very large. Meanwhile, the study that had the smallest effect size was the one that used the Two Stay Two Stray model as a treatment to obtain an effect size of 0.26 with a medium category. Thus, it can be concluded that an interactive and student-centered learning model is more effective in developing short story writing skills.
                        
                        
                        
                        
                            
                                Copyrights © 2025