Pryse H. Duerfeldt

Some relevant variables in observational learning of the rat

Several aspects of observational experience were examined to determine their relative contribution to observational learning of rats. Six groups of 10 Ss viewed demonstrator rats perform different aspects of a visual discrimination task for water reinforcement on an elevated Y maze. Ss were then tested for trials and errors to reach criterion on the same task. Modeling aspects of the demonstrator’s performance were most important, while vicarious reinforcement had little effect. Informational and following aspects also had impact.

STABILITY OF PATTERNS OF SELF-REINFORCEMENT

56 Ss were permitted to supplement a 25% reinforcement schedule on a time-estimation task (TET) with ad lib. self-rewards (SR+). on a separate word-association task (WAT), which Ss perceived as unrelated to the TET, Ss were instructed to self-administer either SR+ or self-criticism (SR—) after training on non-contingent 50% reinforcement. High self-re-warders on the TET also significantly exceeded low self-rewarders on the WAT in SR+ frequency, while low self-rewarders were significantly higher on SR —. However, rates of SR+ responses on the TET and SR— on the WAT were not correlated. The implications for personality constructs of the generality of self-rewarding patterns and their independence of self-criticism frequencies are discussed.

MOTIVATIONAL PROPERTIES OF SELF-REINFORCEMENT

To examine the motivational effects of Self-reinforcement (SR), Ss were asked to match to sample tachistoscopically presented stimuli from four response choices 100 trials with non-repeated samples were given. Three groups received 60% non-contingent reinforcement in Phase I. A fourth (NR) group received no reinforcement during the experiment. In Phase II, an External reinforcement (ER) group received continued non-contingent reinforcement, an SR group was given control over reinforcement delivery, and an Extinction (EX) group received no reinforcement. In Phase III no Ss were reinforced. Analysis of correct responses indicated no group differences in Phases I or II but significant differences under extinction in Phase III. The SR group gave significantly more accurate recognitions than all other groups. In Phase 11, SR administration closely matched the preceding ER schedule.

Resistance to extinction in children as a function of N-length and number of different N-lengths

The Capaldi stimulus modificatipn theory accounting for resistance to extinction as a function of sequential variables is tested with human Ss in a lever pulling task. The sequential variables investigated were N-length and variety of N-lengths. Response measures were (a) attention time, (b) reaction time, and (c) a verbal report as to whether and when S “gave up.” Neither attention nor reaction times differed in extinction. Verbal data, however, gave support to the Capaldi hypothesis that longer and multiple N-length groups yield greater resistance to extinction than shorter and less multiple N-length groups.

The effect of tension level and contingent reinforcement on fear reduction

The degree of tension experienced by the subject and the availability of an eating response were examined for their effects on desensitization of tarantula fear as measured by (a) changes in the subjects approach behavior toward the tarantula after treatment, (b) changes in the subjects approach behavior toward the tarantula during treatment, and (c) changes in the subjects reported fear of tarantulas after treatment. All 64 female subjects significantly reduced their tarantula fears on all measures consequent to treatment. Groups experiencing minimum tension showed markedly greater approach behavior after treatment. Groups provided with an eating response showed greater approach behavior only under conditions of maximum tension. Interpretation of the data favors an extinction paradigm as the most parsimonious explantion for the effectiveness of desensitization types of therapy.

Replacement function of intertrial reinforcement

Under conditions of extended partial-reinforcement training (PR), small reward magnitude, and intertrial reinforcement (ITR), a partial reinforcement effect (PRE) was obtained. Results indicated the degree of ITR occurring during N-length sequences had no apparent effect on PRE, and suggested that ITR-discrimination better accounted for PRE than a frustration explanation. ITR was also shown to affect extinction performance of consistently rewarded (CR) Ss.