Componential Approaches to the Training of Intelligent Performance.

Abstract

This article proposes that 'componential' approaches to the training of intelligent performance are a useful means for producing and understanding improvements in such performance. The outcomes (some of them preliminary) from three experiments conducted at Yale, are cited. The results of these experiments show that at least some optimal strategies for intelligent performance can be trained; that certain strategies are better, on the average, than others; that the properties of strategies that make them 'better' or 'worse' can, at least in some cases, be isolated through componential means; that certain strategies are preferable for people with certain ability patterns, but that other strategies are preferable for people with other ability patterns; that certain strategies are preferable for use with certain stimulus types, but that other strategies are preferable for use with other stimulus types; and that people may be quite cognizant of the strategy they are trying to employ while at the same time being quite incognizant of the strategy they are actually employing, or of the difference between the two strategies.