Optimal self-explanation prompt design in dynamic multi-representational learning environments Yu-Fang Yeh a , Mei-Chi Chen b , Pi-Hsia Hung c , Gwo-Jen Hwang d, * a Department of Knowledge Management, Aletheia University, Matou Campus 70-11, Pei-Shin-Liao, Matou, Tainan County 721, Taiwan b Department of Business Education, National Changhua University of Education, No. 2, Shi-Da Road, Changhua City 500, Taiwan c Graduate Institute of Measurement Statistics, National University of Tainan, No. 67, Rongyu St., East District, Tainan City 701, Taiwan d Department of Information and Learning Technology, National University of Tainan, 33, Sec. 2, Shulin St., Tainan City 70005, Taiwan article info Article history: Received 14 November 2008 Received in revised form 7 October 2009 Accepted 18 October 2009 Keywords: Self-explanation Expertise reversal effect Cognitive load theory Computer-based learning Data structures abstract Self-explanation prompts are considered to be an important form of scaffolding in the comprehension of complex multimedia materials. However, there is little theoretical understanding to date of self-explain- ing prompt formats tailored to different expertise levels of learners to help them fully exploit the advan- tages of dynamic multi-representational materials. To address this issue, this study designed two types of self-explaining prompts: the reasoning-based prompts asked the learners to reason the action run of the animation; the predicting-based prompts asked the learners to predict the upcoming action of the anima- tion, and then asked for reasoning if they made a wrong prediction. Furthermore, multiple indicators including learning outcome, cognitive load demand, learning time, and learning efficiency were used to interpret the prompts’ effects on different expertise levels of learners. A total of 244 undergraduate students were randomly assigned to one of the three conditions: a control and two different self-explain- ing prompt conditions. The results indicate that the learning effects of self-explaining prompts depend on levels of learner expertise. Based on the results, this study makes recommendations for adaptive self- explaining prompt design. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Computer-based learning environments widely use multiple representations and animations to convey and visualize complex mate- rials. To take full advantage of dynamic multi-representational materials, learners are required to actively organize and integrate asso- ciated elements from different and transient information sources. However, students seldom spontaneously apply meta-cognitive strategies to processing dynamic multimedia materials in a cognitively deep manner (Bodermer, Ploetzner, Feuerlein, & Spada, 2004). The absence of the necessary meta-cognitive strategies results in superficial perceptions and processing of the dynamic multi-represen- tational materials. Accordingly, external supports are necessitated when learning in this kind of environment (Hwang, 2003; Hwang, Tseng, & Hwang, 2008). Prompting students to explain the to-be-learned material to themselves is a promising external support for understanding complex materials (Chi, 2000). The process of self-explaining helps learners fill the information gaps between the learning material and the existing schema, to recognize conflicts of understanding, and to amend flaws in the schema (Chi, 2000). Many empirical studies have shown that students who are prompted to frequently self-explain the learning material have greater transfer performance than those who are not (Ainsworth & Loizou, 2003; Aleven & Koedinger, 2002; Mayer, Dow, & Mayer, 2003; Wong, Lawson, & Keeves, 2002). Although prompting to self-explain has been recognized as a useful form of instruction for the understanding of complex materials, its effects on different characteristics of learners remain unclear in instructional research. For adaptive instruction development, it is worth studying if different kinds of prompts for self-explanation have the same optimal effects on all types of learners, and if not, how they differ. 0360-1315/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.compedu.2009.10.013 * Corresponding author. Tel.: +886 915 396558; fax: +886 6 3017001. E-mail addresses: yufang.yeh@gmail.com (Y.-F. Yeh), meichen@cc.ncue.edu.tw (M.-C. Chen), hungps@mail.nutn.edu.tw (P.-H. Hung), gjhwang@mail.nutn.edu.tw (G.-J. Hwang). Computers & Education 54 (2010) 1089–1100 Contents lists available at ScienceDirect Computers & Education journal homepage: www.elsevier.com/locate/compedu