-
Title
-
BEHAVIORAL VARIABILITY UNDER SCHEDULES OF INTERMITTENT REINFORCEMENT.
-
Identifier
-
AAI8203334
-
identifier
-
8203334
-
Creator
-
TREMONT, PAUL JOSEPH.
-
Contributor
-
Brett K. Cole
-
Date
-
1981
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Psychology, Experimental
-
Abstract
-
Variability of interresponse time (IRT), response force time integral (RFTI), and response duration, was examined under Random Interval (RI) and Random Ratio (RR) reinforcement schedules, using 2 groups of 6 rats. For the RI group, the probability of the first response in a repeating time cycle (T) being reinforced was held constant at .100, while the length of that cycle was varied as follows: T = 1.5, 3.0, 6.0, 12.0, and 24.0 sec. The minimum mean interreinforcement intervals of the RI schedules thus generated was 15, 30, 60, 120, and 240 secs. For the RR group the probability (p) of reinforcement for each response assumed the following values: p = .100, .050, .025, .012, and .006, and the schedules thus generated were RR 10, 20, 40, 83.3, and 166.6. Sessions for the RI group were terminated after 600 T cycles; for the RR group they were terminated after the allotted number of reinforcements were obtained or after 45 min., whichever occurred first. Sixty reinforcements were allotted under the RR 10, 20 and 40 schedules, and 45 and 23 reinforcements under the RR 83.3 and 166.6 schedules respectively. Twelve daily sessions were run at each schedule value in the orders listed, and the initial schedule values were recovered for each group. Reinforcement consisted of 2.5 sec. access to .04 cc water and a response was defined as bar contact with a minimum 5 gram force for at least 20 msec. and subsequent reduction of contact force below 5 grams for at least 20 msec.;Response rate and post reinforcement pause functions, obtained from the final 4 sessions at each schedule value, were generally consistent with those found by other workers under similar conditions for both groups. Redetermination resulted in recovery of original rate and pause levels in most cases. The IRT distributions, which may be regarded as reflecting variability in behavior outside of the response class (not-R), were obtained by pooling data over the last 4 days at each condition. They revealed increasing variability, in not-R, with increases in interreinforcement time for the RI group and little systematic change with increases in average response requirement for the RR group. Variability of behavior within the response class (R), when estimated either from distributions of RFTIs or durations (again pooled over the last 4 days at each condition) was not systematically related to changes in intermittency of reinforcement on either RI or RR schedules, but overall was greater on the interval schedules. Both not-R and R variability measures showed, from their recovery data, that extended training had no irreversible effects on variability. Additionally, RFTI and response duration were shown to be highly correlated measures across all experimental conditions; interresponse times were not highly correlated with the RFTIs of the responses terminating them although long IRTs infrequently preceded high RFTIs.;These findings show that the increasing variability of behavior which accompanies decreases in the rate of response occurs, to a large extent, in behavior exclusive of the response class. Moreover, they were suggestive of the possibility that different experiments on variability may incorporate differing amounts of behavior external to the response class (not-R), into their measures of R variability. Experiments using response locus were considered as more conducive to incorporating not-R variability into the R measure, and the possible influence of not-R on behavioral measures is suggested as a matter of importance for behavior theory.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
-
Program
-
Psychology
