Telling stories as preparation for learning: A Bayesian analysis of transfer performance and investigation of learning mechanisms

Abstract

BackgroundTextbooks are essential for natural science university education. However, recent evidence indicates that their design may not be ideal for learning, whereas narratives might overcome the associated limitations. AimThis study compares transfer performance and involved learning mechanisms upon learning scientific concepts either provided in an expository text alone, embedded in a historical narrative, or as expository text prefaced with the historical background. SampleParticipants were 163 undergraduate natural science students. MethodsWe randomly assigned students to one of the three conditions and used a Bayesian modeling approach to compare the prior knowledge-dependent transfer performance upon instruction. Additionally, we investigated the impact of the different conditions on affective and cognitive mechanisms. ResultsResults indicate that students with lower prior knowledge benefit most from narrative-embedded content. Students with higher prior knowledge profited if narratives were used as preparation for follow-up expository instruction. Self-efficacy and cognitive load measures were positively related to the narrative conditions and partly mediated learning from narrative instruction. ConclusionThe study conceptualized and offers support for using narratives as preparation for future learning for enhancing transfer performance in university natural science education, additionally highlighting when and why narratives might support learning.