Now You Don't: Visual Attention for Illustrations During Mathematical Lessons

• In Event: 66.026 - Scaling Up Cognitive Learning Principles to Redesign a Mathematics Curriculum for Improved Learning

Abstract

Integrating visual materials along with text in mathematics curricula is often justified for several general reasons: illustrations may be mathematically necessary (e.g., graphs), contextualize the activity, make the activity more personally relevant and engaging, and foster links between representations. Our coding of items across the entire 3-year curriculum showed that CMP2 is rich with contextual and decorative illustrations in addition to mathematically necessary visuals.

Authors of the IES Study Guide “recommend that teachers combine graphical presentations … that illustrate key processes and concepts with verbal descriptions of those processes and concepts” (Pashler et al., 2007, p. 13). Following these recommendations, the evaluation and subsequent CMP2-R redesign was guided by three general cognitive principles for organizing curriculum information. Based on the multimedia principle, learning is enhanced when texts are accompanied by relevant illustrations (Mayer, 2005) because they foster deeper cognitive processing of the text. Based on the signaling principle, integrating text and visual information is facilitated through the use of color coding and adding missing information. Based on the coherence principle, information such as decorative images that may be interesting to the reader, but not relevant to the activity, can distract learners and diminish comprehension (i.e., seductive details effect; Lehman, Schraw, McCrudden, & Hartley, 2007; Levin, Anglin, & Carney, 1987).

We have examined these principles with eyetracking experiments. In the first, undergraduates (N = 41) read lessons about mathematical functions that had relevant text and a mathematically necessary graph of the function. In addition to the text and graph, the items included either: no additional illustration, a decorative illustration, a contextual illustration, or both types of illustrations (Figure 1). Eye tracking data showed that participants read the text but that little visual attention was directed towards either math-necessary or math-unnecessary images (Figure 2). Performance did not differ by condition.

In the second, middle-school students viewed a lesson from the original CMP-2 or from CMP2-R materials, which had added signals for corresponding information between the mathematically necessary visuals and the text and removed mathematically unnecessary images (Figure 3). Data analyses (N= 57) indicated that the text in the revised materials was easier for students with low levels of prior knowledge to process, based on the average fixation duration. As with the undergraduates, there was little visual attention towards the contextual and decorative illustrations in the original materials (Figure 4). However, middle school students attended to the mathematically necessary tables and graphs. Again, overall performance did not differ by condition.

Prior research using science materials showed benefits of applying these principles on the transfer of scientific concepts (e.g., Sweller & Chandler, 1991; Mayer, 2005). Education researchers have used such findings to derive far-reaching principles for curriculum design and instruction. However, our process-level data shows that students may engage with visuals in math materials differently, leading to different effects on cognitive and affective processing. The premise that effective design principles in one content area apply broadly deserves greater scrutiny as we consider scaling-up general cognitive principles to foster deep processing and learning.

Authors

• Mitchell J. Nathan, University of Wisconsin - Madison

• Virginia E. Clinton, University of North Dakota

• Jennifer L. Cooper, University of Wisconsin - Madison

• Martha W. Alibali, University of Wisconsin - Madison