Abstract
Four groups of partially reinforced Ss were run in a T-maze. During a delay period at the choice point 1 of 2 signal stimuli was presented to all Ss. The information value of the signals was varied for each group and was a function of the reliability with which the stimuli predicted reward or location. The results indicated that the secondary reinforcement value of the signal stimuli was related to their reliability as predictors of both reward and location, rather than to their association or non-association with primary reinforcement.