Harp and Mayer (1998) delineated three classes of explanations for how the presence of images may undermine learning from text: diversion, disruption, and distraction.  This keynote will consider new research consistent with a proposed fourth alternative:  deception. Several studies have now demonstrated that the presence of images, or different types of images, can impact judgments of knowing, beliefs about learning, perceptions of understanding, and readers’ ability to monitor their own learning. The deception hypothesis prompts the question whether the presence of even relevant images (that depict information relevant to understanding a process, system or phenomenon) may sometimes serve as seductive details to readers.

In recent years, there have been lively discussions about the relevance of research findings on learning and instruction for practice. In many cases, even researchers have rather skeptic views whether their findings allows for such codified recommendations as they can be found in the medical domain (i.e., evidence-based medical guidelines). In the present talk, I will take a more optimistic view on the relevance of research on learning and instruction. I will illustrate my arguments for the case of learning from multiple representations. More specifically, I will argue that we should not follow the medical model of evidence-based practice that focusses on systematic reviews, ideally in the form of meta-analysis, for providing specific guideline on action. Instead, I take up the motto of the present conference ("Forest before trees: Towards identifying general patterns in text and graphics research") and argue that we should use general patterns of findings on learning from multiple representations for a variety of different types of "recommendations" (e.g., "warnings" or "showing up possibilities").

Principles on how to best learn from multimedia (i.e., learning with text and pictures) have been thoroughly studied and received much attention for decades. However, visualizations such as representational or decorative pictures also play an important role in materials constructed for testing students’ knowledge and competencies (e.g., in educational assessment programs). Can we simply transfer multimedia principles from learning to testing? How do students use multimedia material in their response process when answering test items, and is there an interaction between the effects of learning and testing with multimedia material? In an effort to address these questions, I will present several empirical studies that shed some light on the cognitive and the affective-motivational effects of different picture types in testing, with the help of process measures such as eye tracking parameters, response times and repeated student ratings. The findings reveal noticeable similarities between effects found for multimedia learning and multimedia testing. However, they also indicate how research in both areas may be mutually beneficial and stimulating.