Psychology ETDs

Publication Date

9-4-1975

Abstract

The temporal flux of reinforcement probability is a prominent feature of many operant situations. Presumably, it is also a pervasive determinant of performance. The momentary probability of reinforcement or some derivative variable has been emphasized in many theoretical accounts of the performance engendered by operant schedules of reinforcement. The duration of a barpress specifies a more restricted class of behavioral and environmental events than doe the length of an interresponse time. Accordingly, it was conjectured that the quantitative details of the regulation of performance by momentary reinforcement likelihood would be revealed more readily if response durations were used as predictors of reinforcement probability. Discrete trials were scheduled at variable intervals. The expected value of the time between the initiation of successive trials was 22 sec. Trials were signalled by turning on a pair of cue lights in each of four chambers. Water-deprived rats were trained to press and release a response lever in the presence of the cue lights. The first experiment included three groups of eight rats for which the release of the lever occasioned the delivery of reinforcement with a probability which was an increasing linear function of the time since the lever had been pressed. The slope of this function was 1.0, .50, and .25, respectively, for these three groups. Reinforcement consisted of 2 sec access to .04 cc of tap water, delivered by a Lehigh Valley liquid dipper. For two additional groups of eight rats reinforcement probability was unrelated to response duration. This probability was .40 and .10, respectively, for Groups 4 and 5. In the second experiment each of 16 rats was exposed successively to each of the three correlated conditions which defined the first three groups of the first experiment. The dependent variable in both experiments was the duration of the barpress responses emitted by the subjects. For the conditions in which reinforcement probability was correlated with response duration, mean response duration was a decreasing function of the slope of the function relating response duration to reinforcement probability. There was no substantial difference between the two groups for which reinforcement probability differed, but was unrelated to response duration. Response durations were consistently much shorter for these two uncorrelated group than for any of the correlated conditions. The distributions of response duration were well described by theoretical predictions derived from the assumption that the momentary probability of a lever release was linearly related to the experimentally programmed reinforcement probability.

Degree Name

Psychology

Level of Degree

Masters

Department Name

Psychology

First Committee Member (Chair)

Douglas Peter Ferraro

Second Committee Member

Joseph Anthony Parsons

Third Committee Member

John Paul Gluck Jr.

Language

English

Document Type

Thesis

Download

Included in

Psychology Commons

Share

COinS