Effect of aging on inferential reasoning about class inclusion

This study investigates the way inferences about class inclusion change fro m childhood to old age. Two experiments were carried out, each consisting o f presenting subjects with five class-inclusion tasks of increasing difficu lty: coordination of the whole and the parts, class intersection, quantific ation of inclusion, modification task, and screen task. In the first experi ment, the manipulated classes were concrete task that preexisted in long-te rm memory and therefore required a low activation level in working memory. Children's performance was compared to that of elderly people with a low le vel of education. The second experiment dealt with classes that required es tablishing new relations between them in working memory. Teenagers' scores were compared to those of old people with a high level of education. Both e xperiments revealed the existence of a cognitive regression phenomenon in o ld age, regardless of education level. However, the nature of the difficult ies subjects have building inferences seems to be different in children and in the elderly. In children, inference errors were caused by a misundersta nding of the dual membership relation of the included class, whereas elderl y people had trouble selecting relevant information to work out the inferen ces. Difficulty selecting relevant data can be interpreted as a decrease in attention capacity and working memory, both of which are characteristic fe atures of aging.