Proceedings of the Annual Meeting of the Cognitive Science Society

Analogical transfer between source and target problems ought to be a major contributor to problem-solving and learning. Yet, data from laboratory studies show that successful spontaneous analogical transfer does not reliably occur in the absence of explicit hints to analogize, in the presence of a delay between source and target, or when there are extensive filler tasks, a finding attributed to the complexity of analogy retrieval and mapping. Here, we show that participants solving variants of the Cards problem often failed to show transfer between source and target problems that shared both conceptual and superficial similarities. Frequency of re-inspecting the task instructions was a significant predictor of transfer, with participants successful at T2 requiring fewer re-inspections. The results suggest that analogical transfer may be limited, not just by the difficulty of mapping between source and target, but by a lack of conceptual understanding of the source and its solution, even when the source is solved.