Abstract
The performance of 192 adults in a functional learning task containing four types of cue-criterion relations—direct, inverse, U-shaped, and inverse-U-shaped—was compared. The ability of these adults at extrapolating learning to situations containing more extreme values than the ones in which it took place was also examined. Older participants (65–90-year-olds) were able to learn the nonlinear functions practically as well as younger participants (18–25- and 40–50-year-olds); they were also able to extrapolate well beyond the range of learned responses. When extrapolating, however, they frequently used a more limited range of responses than the younger participants.
This work was supported by the laboratory Cognition et Décision (Ecole Pratique des Hautes Etudes), by the UPRES Vieillissement et Développement (Français-Rabelais University), and by the UMR Travail et Cognition (Mirail University). The authors are grateful to Sheila Rivière, Michel Isingrini, and Jean-Claude Marquié for their thoughtful comments on an earlier draft of this article.