Aging and skilled problem solving.

Information-processing models of problem solving too often are based on restrictive age ranges. On the other hand, gerontologists have investigated few problem-solving tasks and have rarely generated explicit models. As this article demonstrates, both fields can benefit by closer collaboration. One major issue in gerontology is whether aging is associated with irreversible decrement or developmental plasticity. If both processes occur, then an appropriate strategy for investigating aging is to equate age groups for molar problem-solving performance and search for differences in the underlying components. This strategy was adopted to examine the relation of age and skill to problem solving in chess. Chess players were selected to vary widely in age and skill such that these variables were uncorrelated. Problem-solving and memory tasks were administered. Skill level was the only significant predictor for accuracy in both a choose-a-move task and a speeded end-game evaluation task. Age (negatively) and skill (positively) jointly determined performance in an unexpected recall task. Efficient chunking in recall was positively related to skill, though negatively related to age. Recognition confidence, though not accuracy, was negatively related to age. Thus despite age-related declines in encoding and retrieval of information, older players match the problem-solving performance of equivalently skilled younger players. Apparently, they can search the problem space more efficiently, as evidenced by taking less time to select an equally good move. Models of chess skill that stress that role of encoding efficiency, as indexed by chunking in recall, need to be modified to account for performance over the life span.