Doctoral Dissertations

Date of Award

8-1981

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Psychology

Major Professor

John C Malone

Committee Members

William Calhoun, Jim Lawler, Mike Moshell

Abstract

During the past 20 years several multiple schedule interactions have been shown to accompany the development of a discrimination. Four of these are: 1) behavioral contrast, 2) local contrast, 3) dimensional contrast, and 4) local dimensional effects. These interactions were the subject of the present experiments.

In Experiment I, strong behavioral and dimensional contrast effects appeared as a discrimination developed during discrimination training. However, only marginal local effects (local contrast and local dimensional effects) appeared. Since behavioral and dimensional contrast occurred in the absence of parallel local effects, it seems reasonable to conclude that they are not simply the sums of local effects.

In the discrimination phase of Experiment I, only the two extreme stimuli were associated with reinforcement (i.e., were S+'s). Under such conditions, dimensional contrast appeared in the form of a W-shaped gradient. This W-shaped gradient has also been found by Hinson and Malone (1980), using a similar procedure. Since both demonstrations of this particular effect employed line-orientation stimuli, it is possible that it is not general, but occurs only with line-orientation continua. Experiment II examined this possibility by employing a column of lights instead of line-orientations. In this experiment, a gradient resembling a W shape was obtained. Thus, the W-shaped gradient does not seem to be specific to line-orientation continua.

Two features that may influence the occurrence and magnitude of dimensional contrast in S- components are average response rates and average reinforcement rates in S+ components. Since the two factors are usually positively correlated, it is difficult to determine their respective contributions to dimensional contrast. In Experiment III, however, response rates were manipulated, while reinforcement rates were held constant. This allowed an assessment of the independent contribution of average response rates to dimensional contrast. It was found that whenever response rates in an S+ were relatively high, response rates in the adjacent S- were relatively low. This demonstrated that response rates do influence the magnitude of dimensional contrast.

Dimensional contrast is a recently described phenomenon, and much remains to be known about it. Therefore, parametric studies are of importance. Experiment IV was a parametric study which showed that 1) dimensional contrast was more easily obtained when S-'s were associated with extinction than when they were associated with a relatively lean VI schedule of reinforcement, 2) dimensional contrast reappeared when the discrimination was retrained, and 3) omitting a stimulus did not affect the degree of dimensional contrast influencing the remaining stimuli.

In addition to the above, two models of operant discrimination, Blough's 1975 model and the "neural unit" model, were applied to the results of Experiments I, III, and IV. The "neural unit" model was shown to better approximate the actual data.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS