knowledge in the learner that are broadly relevant (such as mental discipline or general principles). On the general transfer side of the controversy was the doctrine of formal discipline, which held that learning certain school subjects such as Latin and geometry would improve the mind in general (i.e., teach proper habits of mind) and thereby improve learning and performance in other unrelated subjects. On the specific transfer side of the controversy was E.L. Thorndike, largely recognized as the founder of educational psychology, who sought to put the issue to an empirical test. In a famous set of early studies, Thorndike and Woodworth (1901) found that students who were taught a cognitive skill showed a large improvement on the taught tasks but not on other tasks. Thorndike was able to claim strong support for specific rather than general transfer: “Improvement in any single mental function rarely brings about equal improvement in any other function, no matter how similar” (Thorndike, 1903, p. 91).

This was not a good outcome for those dedicated to helping students develop the ability to exhibit general transfer—that is, to apply what they have learned in one situation to a novel situation. Subsequent work by Judd (1908) offered some hope by showing that transfer to new situations depended on the instructional method used during initial learning, with some instructional methods supporting transfer to new situations and others not. An important aspect of Judd’s finding is that transfer was restricted to new situations that required the same general principles as required in the original task, although it could be applied to situations requiring different behaviors.

Judd’s finding has been replicated in many contexts. For example, Singley and Anderson (1989) report on an experiment designed to study the acquisition and transfer of skills in text editing. A group of 24 young women (aged 18-30) from a secretarial school were first taught to use either one or two line editors (text editing software used to change individual lines of text) and then a screen editor (text editing software used to scroll throughout a page of text), while control groups spent similar amounts of time either learning and using one of the screen editors or simply typing a manuscript. The authors observed positive transfer, both from one line editor to the next and from the line editors to the screen editor, as indicated by reductions in total learning time, keystrokes, residual errors, and other measures in comparison to the control groups. They proposed that the very high level of transfer from one line editor to the next line editor was due to the fact that, although the surface features of the commands used in the two editors were different, the underlying principles were nearly identical. In addition, they proposed that the moderate level of transfer from the line editors to the screen editor reflected the fact that the procedures used in the two line editors are largely different from those used by the screen editor. Nevertheless, the two line editors and the screen editor do share several

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement