Punishment as a discriminative stimulus and conditioned reinforcer with humans

Mental hospital patients were reinforced for responding in a two-response operant situation. When a noise was used to punish one of the responses, all subjects shifted to the unpunished one. When the noise was then paired with positive reinforcement, the subjects responded to produce the noise. Also, a novel response was reinforced by noise in the absence of other reinforcers. This study with humans extends the findings of previous studies with animals in revealing how a punishing stimulus can acquire discriminative or conditioned reinforcing properties.     

Conditioned reinforcement in second-order schedules

Pigeons responded under a schedule in which food was presented only after a fixed number of fixed-interval components were completed. Two such second-order schedules were studied: under one, 30 consecutive 2-min fixed-interval components were required; under the other, 15 consecutive 4-min fixed-interval components were required. Under both schedules, when a 0.7-sec stimulus light was presented at completion of each fixed interval, positively accelerated responding developed in each component. When no stimulus change occurred at completion of each fixed interval, relatively low and constant rates of responding prevailed in each component; a similar result was obtained when a 0.7-sec stimulus change occurred at completion of each fixed interval except the one which terminated with primary reinforcement. The 0.7-sec stimulus correlated with food delivery was an effective conditioned reinforcer in maintaining patterns of responding in fixed-interval components despite low average frequencies of food reinforcement.     

An individual differences model for multiple regression

A model for multiple regression was developed which allows individual differences to emerge empirically. The model encompasses as special cases several of the previous attempts to improve psychological prediction by deviating from the usual linear multiple regression model. The model is tested with both artificial and real data. The results indicate that the model effectively reduces the variance of the error of prediction, and that the weights obtained are stable over different samples, and, to some extent, over different sets of predictors.This article is based upon a thesis submitted in partial fulfillment of the requirements for the doctoral degree at the University of Illinois. The author thanks Professor Ledyard R Tucker who served as committee chairman and offered considerable support and assistance.