Reinforcer Quality and Self-Control
Self-control has been extensively studied using procedures in which subjects chose between two reinforcer alternatives. Traditionally, one of those alternatives delivers a small reinforcer after a short delay (SI), the other, a larger reinforcer after a long delay (LD). Choosing the SI is defined as impulsivity as it requires forfeit of the larger reinforcer; and choosing the LD is termed self-control. Four experiments were conducted to examine behaviour using non-human animal analogues of self-control situations. The subjects used for all four experiments were Norway-hooded rats. Experiment 1 used an SI - LD self-control paradigm to examine the effect of manipulating reinforcer quality on response distribution. Findings were that behaviour became more impulsive as the delay ratio became more extreme and this tendency was more systematic when different quality reinforcers were used for the SI and LD alternatives. Experiments 2 and 3 introduced a novel self-control paradigm designed as an analogue of choice situations in which individuals choose between two competing immediately available reinforcers each associated with a different delayed reinforcer. The procedure used was a concurrent-chains schedule that delivered primary reinforcement in the initial and the terminal links. The initial reinforcers were of equal amount and unequal quality; the terminal reinforcers were of unequal amount and equal quality. An impulsive choice was defined as choosing the alternative that delivered the most-valuable reinforcer in the initial link and the least-valued reinforcer in the terminal link. A self-controlled choice was defined as choosing the alternative that delivered the least-valuable reinforcer in the initial link and the most-valuable reinforcer in the terminal link. The results indicated that behaviour was more self-controlled when the terminal reinforcer quality was ethanol solution and increasing the delay between the initial and terminal links increased subjects' responding on the impulsive choice. Behaviour allocation in Experiment 3 was well described by the Contextual Choice Model (Grace, 1994) when the temporal context scaling parameter (k) was allowed to vary. Subjects that were relatively more impulsive had lower derived k values. The final experiment presented the subjects from Experiment 3 with concurrent variable interval (VI) VI schedules in which one alternative delivered plain-sucrose solution and the other ethanol-sucrose solution. Preference measures obtained from Experiment 4 were negatively correlated with the values obtained for the scaling parameter in Experiment 3, indicating that subjects which were more impulsive in the MN - ML paradigm had a stronger preference for ethanol. In summary, findings indicate that reinforcer quality may change the discriminability of reinforcer alternatives; and the influence of reinforcer quality on response allocation is well described by quantitative models based on the Matching Law.