Choice Among Multiple Alternatives

Abstract

Choice behavior has been widely researched, but few studies have examined the effect of multiple (more than two) alternatives on choice. Mathematical models of choice behavior predict that response allocation will not be affected when multiple alternatives are present, but recent research has not found support for this prediction. In Experiment 1, five pigeons were presented with three concurrently-scheduled alternatives. These alternatives were assigned to different VI schedules of reinforcement in ratios of 9:3:1. In some conditions all three keys were available, and in others only two keys were available. Choice was not affected by the total number of available alternatives; however, for two birds in the three-alternative conditions and three birds in the two-alternative conditions, preference was more extreme for the pair of alternatives with the lower overall pairwise reinforcer rate (3:1) than the pair with higher overall reinforcer rate (9:3). This result differed from past research and arose because the birds made significantly fewer responses per reinforcer to the “9” alternative than the other alternatives. Experiment 2 investigated whether changing the overall reinforcer rate affected preference among four alternatives. Four concurrently-scheduled keys were presented to the birds; each key was assigned to different VI schedules of reinforcement in ratios of 8:4:2:1. Reinforcement was arranged probabilistically using a reinforcer interval and this was varied between conditions. Following this, two of the alternatives, “2” and “4”, were put into extinction while the RI value was changed from 60-s to 5-s. The birds’ behavior was independent of RI length across all conditions. However, in the two-alternative conditions, the birds’ log response ratios for the 8:1 pair were less extreme compared to the four-alternative conditions, the birds were responding less to the “8” alternative when there were fewer alternatives. This occurred for two out of four birds in the RI 60 two-alternative conditions and three out of four birds in the RI 5 two-alternative condition. Experiment 3 investigated whether arranging reinforcement using a rapidly changing alternatives procedure changed how the birds made choices. Four pigeons were presented with three rapidly changing alternatives. Each bird was presented with the three different reinforcer ratios (9:3:1) in one of six possible arrangements, and after 10 reinforcers all of the keys were blacked out for 10-s. Following this blackout the alternatives would be in a new location. The obtained reinforcer ratios were similar to those arranged, but the birds did not show any differences in preference among the alternatives; that is, the birds were unable to learn the rapidly changing alternatives procedure. Together, the data from the present experiments do not support the established theories of choice behavior, and it is not obvious at this point how one model could accommodate such a range of results at a mathematical or conceptual level. The data presented in this thesis show that more research is needed in order to understand how choosing among multiple alternatives affects behavior.