Psychology ETDs


David Perkins

Publication Date



In order to test the hypothesis that resistance to extinction is directly related to the amount of response sub-class variability reinforced during acquisition, the number of different response subclasses reinforced during acquisition was directly controlled and subsequently related to resistance to extinction. A response subclass was defined as a response to one of eight key locations on a horizontally oriented response strip which was presumed to represent a topographical response continuum.

Seven groups of 3 pigeons each were first trained to peck the transilluminated response strip for 3-sec. access to grain reinforcement. Each group received a total of 752 reinforcements distributed over 14 consecutive sessions of acquisition followed by six 12 minute extinction sessions. For each group, reinforcement was made available at any given time for only one of the locations designated for reinforcement. Responses at all other locations went unreinforced. The groups differed with respect to number of keys reinforced, location of reinforced keys and programmed probability of reinforcement.

The finding of primary interest was the relative resistance to extinction of the various groups. An over-all analysis of variance showed that the groups differed significantly (p< .025) in response persistance under extinction. Further analysis indicated, in general, that resistance to extinction was directly related to the number of sub-classes reinforced and that persistance of responding during extinction was greater when non-reinforced locations separated reinforced locations than when reinforced locations were adjacent. A comparison relevant to the variable of key location, indicated that the actual location of pairs of adjacent reinforced keys did not affect resistance to extinction. Of interest also was the finding that probability of reinforcement was not related to resistance to extinction when number of keys was held constant. In general, therefore, resistance to extinction was found to be directly related to the amount of response sub-class variability produced during acquisition, thus supporting the general response variability theory of resistance to extinction.

Level of Degree


Department Name


First Committee Member (Chair)

Douglass P. Ferraro

Second Committee Member

Frank A. Logan

Third Committee Member

Louis E. Price

Fourth Committee Member

David W. Bessemer



Document Type


Included in

Psychology Commons