Overlearning and frustration effects during extinction of operant responding

The two experiments in this study investigated overtraining effects and frustration effects during extinction of combined runway-barpress responding by albino rats The chain of runway-barpress responding provided discrete starting and runway time scores as well as massed barpross frequency scores. Magnitude of food rewards was held constant. During extinction, trained subjects were allowed free access to a 'recovery box' and allowed to return repeatedly to the bar press box. Findings showed (1) a monotonic relationship for resistance to extinction as indexed by number of bar presses in extinction and (2) a nonmonotonic relationship for one of the runway indices (starting time) of resistance to extinction. The monotone curve supported the theoretical position (Hull; Pavlov) that resistance to extinction increases as a function of increase in training trials. The nonmonotone curve, from data from the same subjects, did not support the Hullian line of reasoning but was supportive of other experimental work (North and Stimmel, and others) which has yielded U-shaped curves for runway scores indexing resistance; to extinction. Results were explained as being congruent with previous overtraining bar press studies that have consistently supported the Hullian and Pavlovian positions that increasing training produces increased resistance to extinction. Apparatus differences, spaced versus massed training trials, time versus frequency data, and the extinction of rg-Sg combined with the development of were proposed as explanatory mechanisms which could have influenced the finding of the nonmonotone relationship for starting time. Indexing frustration by amount of time spent in the recovery box did net differentiate among the trained groups in the first experiment but did differentiate between highly trained Experiment A subjects and highly overtrained Experiment B subjects. The most highly overtrained subjects spent the greater amount of time in the recovery box. Highly overtrained subjects entered the recovery box no more often, however, than other trained subjects did. Further experimental work was recommended to clarify the influences of intertrial interval, magnitude of reward variation, and frustrative effects upon extinction processes. Attention was drawn to the often arbitrary choosing of extinction criteria in learning experimentation.'Psychology, Department o