Abstract
128 Ss performed on two concept learning problems within which the sequence of relevant cues was essentially random. For one problem a constant proportion (k) of irrelevant cues changed whenever the relevant cue remained the same on adjacent trials (S-transition) whereas the values of the irrelevant stimulus dimensions varied at random whenever different relevant cues appeared on adjacent trials (D-transition). Within the second problem, k irrelevant stimulus dimensions changed at D-transitions, while the values of the irrelevant dimensions varied at random at S-transitions. For independent groups, k ranged from .00 to 1.00. Trial of last error proved to be an increasing linear function of proportion of irrelevant dimensions changed at D-transitions whereas the proportion changed at S-transitions made no difference. These results are difficult to understand in terms of an all-or-none model of concept identification in which it is assumed that cues are sampled at random.
Get full access to this article
View all access options for this article.