The “Gambler Effect”: Choice and the Probability of Reinforcement of Small Response Outputs

Abstract

Several types of concurrent-chains studies in which preference was found disproportionate to the rate of primary reinforcement were discussed. The disproportionality was attributed to the unusual effect on preference of the smallest interreinforcement interval components of schedules of reinforcement. The present study investigated an aspect of this factor in the control of preference. Concurrently available fixed-ratio 10 schedules were chained to a fixed-ratio 30 schedule on one key and a multiple fixed-ratio 5 fixed-ratio 80 on the alternative key. In the course of the study the relative frequency of the smallest interreinforcement interval (FR 5) in the multiple schedule option was reduced from 0.50 to 0. Pigeons could maximize reinforcement by always choosing the FR 30 schedule or choosing the multiple schedule and “gambling” that the FR 5 would be produced. Three pigeons chose the multiple schedule exclusively with one exception until the FR 5 was removed entirely from the multiple schedule. At this point, all birds chose the fixed schedule exclusively. Preference defined as rate of initial links responding was, in most instances, inversely related to rate of primary reinforcement.