Not All Transfer Is Created Equal: Making the Case for Different Types of Transfer

• In Event: The Form and Function of Transfer in Diverse Contexts

Abstract

Understanding when and how students use knowledge learned in one situation and apply it to another (i.e., transfer) is essential to more efficient and effective learning (e.g., Alexander & Murphy, 1999; Schoenfeld, 1999). Although research on transfer is rather mature, there appears to be little commonality among researchers in how transfer is conceptualized and operationalized (Lobato, 2006; Loughlin et al., 2011). This can often create problems in that evidence for transfer in studies may not be equivalent.

Recently, Loughlin et al. (2011) proposed a two-by-two matrix based on the literature that was found useful in categorizing studies of transfer. In this matrix, variations in the concept of transfer found within the literature were positioned based on what was being transferred (i.e., perceptual versus conceptual knowledge representations) and where transfer was occurring (i.e., within or between domains). For the categorized studies, transfer entailed carrying forward knowledge or processes previously learned (base) to the understanding or performance of a novel concept or task (target).

Thus, it would appear transfer would be influenced by what people have already learned (i.e., prior knowledge) and what they are able to perceive as potential relations between the known and the novel (i.e., pattern recognition). The current investigation empirically tests the validity of such a two-by-two matrix via two questions. First, does manipulating what is expected to transfer and where transfer is expected to occur influence individuals’ positive and negative transfer? Second, are prior knowledge and pattern recognition predictive of individuals’ positive or negative transfer?

Participants for this study were 121 undergraduates from the United States and New Zealand who first completed knowledge and pattern recognition measures followed by four problem-solving tasks. For the first, participants learned to solve a multiplicative probability problem (base). Participants who successfully completed the base problem then completed a distractor task on statistical variance. Next, participants were randomly assigned to one of four conditions in which the target, transfer problem varied according to the elements in the two-by-two transfer matrix (i.e., near-perceptual, far-perceptual, near-conceptual, and far-conceptual). Last, participants completed a problem on additive probability for which prior training would be counterproductive (negative transfer).

Preliminary findings indicated that positive transfer differed by condition (b=-.55, p<.01), but no difference was found for negative transfer (b=.03, p=.88). Prior knowledge was not a significant predictor of positive or negative transfer, likely due to the low levels of prior knowledge for this sample. However, there was a modest positive relation between pattern recognition and positive transfer. These data support the contention that what is being transferred and where transfer is occurring influences transfer performance. Examining these dimensions should contribute to more valid judgments about the nature of transfer. Moreover, the consideration of pattern recognition was a novel dimension to existing studies of transfer and warrants further exploration.

Authors

• Daniel Dinsmore, University of North Florida

• Stephanie Doyle, Victoria University of Wellington

• Peter Baggetta, University of Maryland

• Sandra Michelle Loughlin, University of Maryland