Impact of Narrative versus Expository Instruction in Science Education on Recall, Understanding, and Transfer: A Meta-Analysis

Instructional materials in science education from primary school to graduate level are typically written as expository texts. Instead, narratives are only rarely chosen to teach scientific concepts, likely due to the inconclusive evidence regarding the effectiveness of narrative interventions. This meta-analysis synthesizes the results of 30 empirical studies with 72 independent effect sizes representing the academic performance of over 5300 students when learning with either narrative or expository materials. The narrative learning materials had significant and positive effect on performance (g = 0.16; 95% CI: 0.03-0.30, p < .05) but indicated high between-study heterogeneity (I^2 = 79%; 95% CI: 75-83%). Follow-up moderator analyses revealed that learning from personal or scientist-centered stories was more favorable than learning from fictional stories. The findings were robust across different education levels. Furthermore, the results indicated that narratives were especially favorable in supporting understanding and more impactful in formal than informal educational settings.