动画教学代理对多媒体学习的影响:学习者经验与偏好的调节作用

以往研究发现，动画教学代理对多媒体学习效果的影响不一致，可能受到学习者特征和偏好的调节。本研究以“空调的组成部分及工作原理”为实验材料，采用两个实验控制教学代理有无、经验高低和代理偏好，探讨动画教学代理对多媒体学习的影响。实验1发现与无代理组相比，代理组对教学视频的注视点个数更多，平均眼跳潜伏期更短，学习兴趣也更高；低经验者在代理条件下的迁移成绩更好。实验2发现学习者在偏好代理和非偏好代理条件下的迁移成绩好于无代理组；偏好代理组感知到更低的认知负荷，对总体视频的注视点个数和注视频率更多，平均眼跳潜伏期更短，对学习内容的注视点个数更多。结论认为：在多媒体中加入教学代理不会减少学习者对学习内容的注意，能提高多媒体学习效果，支持社会代理理论假设；教学代理能提高低经验者的学习效果，但对高知识经验者无明显作用；加入学习者偏好的代理形象会促进学习，但加入学习者非偏好的代理形象并没有阻碍学习。

Effect of Animation Pedagogical Agent in Multimedia Learning: The Role of Learner' s Experience and Agent Preference

Previous studies showed that the effect of Animated Pedagogical Agent (APA) in multimedia learning was debatable. In this study, animations related to the components and working principle of air conditioning were used as multimedia learning materials, and SMI RED 250 Eye tracker was used to record eye movement data. In experiment 1, A 2(Agent: APA vs. non- APA) × 2(Prior knowledge: high vs. low) between-subjects design was employed to investigate whether APA can improve learning and the effect of prior knowledge. In experiment 2, a single factor (preference APA, non-preference APA,non-APA) design was employed to investigate the influence of learners' preference for agent. Results of experiment 1 found that compared with no APA, learners in the APA condition had much more fixation count, shorter saccade latency average for animation, and also had higher learning interest. Learners with low prior knowledge in the APA condition outperformed non-APA on transfer tests, but not for learners with high prior knowledge. Results of experiment 2 found that learners in the preference group and non- preference group performed better on transfer tests than non-APA group. Learners in the preference group perceived lower cognitive load, and had more fixation count for air-condition. In conclusion, APA can promote multimedia learning performance for learners with low prior knowledge, but not for high prior knowledge. APA can facilitate learning when learners preferred the APA image, and can improve the learning performance without reducing the attention to the core learning material. So it supported the Social Agency Theory.