A critical review is made of Kagan's conceptualization of the cognitive style dimension of reflectivity-impulsivity as being indicative of stable differences in children's problem solving abilities. Kagan's approach is criticized on both methodological and conceptual grounds. The Matching Familiar Figures test (MFF) and the Kansas Refleetion-Impulsivity Scale for Preschoolers (KRISP) allow individuals to trade speed for accuracy. Both tests indicate different speed-accuracy tradeoff strategies which result from different interpretations of explicit and implicit instructions.

The stability of differences in problem solving as measured by the KRISP was investigated in the present experiment. Forty-one pre-school children were categorized according to cognitive style by means of the KRISP. Subsequently, each child did a 'forced-choice' line discrimination task. The results indicate (1) that the reflective-impulsive dimension does not represent stable differences in problem solving abilities, and (2) that reflective children do not engage in more efficient and detailed processing than impulsive children. It was found that the percentage of correct responses on the line discrimination task was equivalent for the two groups, and that whereas impulsive children adopted different strategies for easy and difficult discriminations, reflective children did not.

Concerning the total population of pre-school children, it was discovered that they use both the differences and the ratios of the pairs of line lengths to make their decisions. This, as well as the finding of a "distance effect" (the more different the choices the faster the decision time) suggests that the ways in which children store information and make decisions are analogous to those of adults.