多媒体学习中线索效应的元分析

作为多媒体学习中一种重要的指导性教学设计形式, 线索能否引导学习者的注意并促进学习效果, 目前结论尚不一致。研究运用元分析技术, 分别以保持测验、迁移测验的成绩作为学习效果的结果变量, 以线索区的注视时间、注视次数作为注意引导的结果变量, 探讨多媒体环境中的线索对学习效果和注意引导的影响, 并通过调节效应检验考察线索效应的边界条件。采用严格标准进行文献检索和筛选后, 共获取43篇符合元分析的文献。在保持测验成绩和迁移测验成绩上, 分别生成了44个(3910人)和41个(3906人)独立效应量; 在对线索区的注视时间和注视次数上, 分别生成了14个(938人)和11个(816人)独立效应量。结果发现, 线索的加入显著提高了保持测验成绩(g = 0.53)和迁移测验成绩(g = 0.36), 同时也增加了学习者对线索区的注视时间(g = 0.50)和注视次数(g = 0.70)。调节效应分析发现：在保持测验成绩上, 静态学习材料、陈述性知识、理工科条件下的线索效应分别高于动态学习材料、程序性知识和文科; 在迁移测验成绩上, 时空线索、静态学习材料、陈述性知识条件下的线索效应分别高于物理线索、动态学习材料和程序性知识; 而在眼动指标上, 线索对注意的引导主要受线索类型的调节, 物理线索条件下学习者对线索区的注视时间和注视次数均高于时空线索。结论认为, 线索的加入确实促进了学习者对学习材料的识记和理解效果, 同时也引导了学习者的注意分配; 但线索效应一定程度上受线索类型、学习材料动态性、知识类型、学科属性等因素的调节。未来研究需进一步考察线索效应的其他边界条件。

Cueing effect in multimedia learning: A meta-analysis

Cueing is an important instructional design technique to improve learning performance in multimedia learning. Recently, researchers focus on whether cueing could guide learners’ attention or facilitate learning outcomes. However, there is still no consistent conclusion about cueing effect of guiding attention, organization and integration in multimedia learning. In this study, we conducted a meta-analysis to investigate the effects of cueing on learning performance (e.g. retention and transfer test), attention guiding (e.g. fixation time and fixation count of cued areas), and moderating factors (e.g. physical and spatial-temporal cues, declarative and procedural knowledge, dynamic and static presentation, and subjects of learning materials). We used key words (e.g. cue, cueing, signaling, attention guidance, multimedia learning, animation, eye tracking, eye movement) to search literatures in database and got 43 related empirical studies. Forty-four independent effect sizes (3910 participants) were included in retention tests-related meta-analysis while 41 independent effect sizes (3906 participants) were included in transfer tests-related meta-analysis. Fourteen independent effect sizes (938 participants) were included in fixation-time-related meta-analysis while eleven independent effect sizes (816 participants) were included in fixation-count-related meta-analysis to test the attention-guiding effect of cueing. According to the characteristics of studies about cueing effect, we reasonably chose random-effects model as meta-analysis model. CMA2.0 was used to test the independent effect sizes. Finally, the impact of publication bias was tested to guarantee a more convincing conclusion. The results showed that cueing could significantly improve learning outcomes of retention test (g = 0.53, 95% CI = 0.36, 0.69]) as well as transfer test (g = 0.36, 95% CI = 0.23, 0.49]). At the same time, learners’ fixation time (g = 0.50, 95% CI = 0.30, 0.71]) and fixation count (g = 0.70, 95% CI = 0.43, 0.97]) of cueing areas attracted much more attention than uncued areas. This result supported the attention guiding hypothesis of cueing in multimedia learning. Moderator analyses indicated that cueing effect in learning outcomes was mainly moderated by types of cueing, dynamism of materials, subjects of materials and types of knowledge. Specifically, cueing resulted in higher retention and transfer scores in conditions of static materials and declarative knowledge rather than dynamic materials and procedural knowledge. Compared with the materials in liberal arts, cueing effect in retention test was weaker than that in science or engineering. In addition, spatial-temporal cues performed much better than physical cues in transfer test. However, in regard to attention-guiding, cueing effect was mainly moderated by cueing types itself. Fixation time was longer and fixation count was more in the condition of physical cues than spatial-temporal cues. The observed strong cueing effect suggested that presenting cues rationally can indeed improve learners’ memorization and comprehension of knowledge as well as distribution of attention. Physical and spatial-temporal cues, declarative and procedural knowledge, dynamic and static presentation, and subjects of learning materials may act as important moderate variables in cueing effect in multimedia learning. The meta-analysis of moderator factors supported the boundary hypothesis of multimedia learning principles (Mayer, 2010). Limitations and future research were also discussed in this study.