This article explores learner-generated drawing, a strategy in which learners construct representative illustrations in support of learning goals. Both applied and empirical literature is reviewed with the purpose of stimulating research on this strategy. Clear from this review is the gap that exists between prescriptive readings on learner-generated drawing and research-based understandings. To make sense of inconsistent empirical evidence, the research review is organized around a series of hypotheses grounded in current underA hallmark of sophisticated, expert-like performance is the ability to think flexibly and to transfer knowledge across contexts. In part, this process is reliant on an underlying cognitive structure in which knowledge is integrated across varying representations (de Jong and Ferguson-Hessler, 1986;Silver, 1979)
286Van Meter and Garner (Van Someren et al., 1998). When a student can translate a data table to a linear function (Haverty et al., 2000) or use illustrated text to solve transfer problems (Mayer and Sims, 1994), that student is assessed as functioning at a higher level relative to a student who struggles with these activities. As critical as integration is, learners have a difficult time working with more than one format and integrating verbal and nonverbal representations of complex content (e.g., Scanlon, 1998;Tabachneck-Schiif and Simon, 1998). Given both the value and challenge of this process, strategies that facilitate the integration of different representations, particularly those that cross modalities, have great potential for improving student learning. Learner-generated drawing is one such strategy because drawing involves the construction of an internal, nonverbal representation that is mapped onto elements of the alternative, provided representation. The purpose of this article is to make the case that learner-generated drawing is a strategy that warrants thorough, systematic study.Although learner-generated drawing received some attention in the mid to late 1970s, research interest dried up by the mid 1980s. We believe the loss of interest is partially attributable to inconsistent findings and a body of research which, on balance, is rather disappointing. Along with the obvious value briefly stated above, there are two additional reasons for renewing interest in learner-generated drawing. First, an abundance of prescriptive publications available to classroom teachers tout learnergenerated drawing as a strategy that can meet a number of educational objectives. This is true despite a lack of evidence to support most applications. It is not that these prescriptions are necessarily wrong; rather, the research evidence addressing drawing as a learning process is inconsistent, silent, or qualifying. To highlight the distance between research and practice, the review section of this article begins with an overview of the applied literature before synthesizing the empirical research. We hope the juxtaposition of an array of implementations against scant research evidence ...