The effects of affective pedagogical agent in multimedia learning environments: A meta-analysis

Affective pedagogical agent (PA) is an image of a character embedded in multimedia lessons with the ability to influence learners' affective experiences and learning performance. Prior studies on the effects of affective PA have shown inconsistent findings. In this study, we conducted four separate meta-analyses to address whether adding an affective PA to multimedia lessons can increase learners' retention performance, transfer performance, positive emotions, and intrinsic motivation, and to explore several moderators that may have contributed to the inconsistencies of previous studies. The research framework mainly includes introducing the concept of affective PA, reviewing research on the impact of affective PA on learning performance, emotions, and motivation, analyzing the moderators that may affect the effects of affective PA, performing a meta-analysis, and discussing the results based on the findings of the meta-analysis. We found 36 articles met the inclusion criteria. The results of the meta-analysis indicated that affective PA could increase learners' positive emotions (k = 25, g = 0.26), improve intrinsic motivation (k = 26, g = 0.26), and facilitate learning performance (retention: k = 35, g = 0.26; transfer: k = 45, g = 0.34). Furthermore, moderator analysis found that affective PA characteristics (i.e., appearance, the number of emotional cues, and body movement) and learning materials characteristics (i.e., subject domain, pacing of presentation) moderated the effects of affective PA. We discussed these findings from different theoretical perspectives. In general, affective PA could help students be happier and more motivated to learn in multimedia learning environments.