Abstract

It is widely agreed that similarity is an important factor in transfer of training. However, the precise role of similarity in transfer is not well understood. We believe that the account can be clarified by a finer grained analysis of similarity. In this research we consider two factors that should affect transfer between two systems: (1) surface similarity, which we operationalize as similarity between individual device components; and (2) shared systematicity whether the learner possesses a coherent mental model of the original device that can apply to the second device. Subjects learned an operating procedure for a simulated device and then were asked to transfer that procedure to a new device. Two factors were varied: (1) the systematicity of the original device model - whether the subjects were given a systematic causal model or simply a set of operating procedures; and 92) the degree of surface similarity between corresponding device components (called transparency). The dependent measure was the number of trials to criterion in the original and transfer devices, and in the same target device with an additional load task. The results show effects of both variables. Having a systematic mental model greatly facilitated learning of the initial device, and may have also promoted transfer to the target device. Transparency had strong effects on transfer: subjects learned the new device fastest when the corresponding components were highly similar, and slowest when there were spurious similarities between target components and noncorresponding base components (the crossmapped condition). These results suggest that there are at least two separable factors that promote transfer: (1) the systematicity of the domain model; and (2) surface similarity of corresponding components.