Schedule sensitivity of instructed human operant behavior: Effects of a warning, length of training and variability pre-training with elaborate and minimal instructions.

Previously published studies have shown that behavior under the control of instructions is likely to follow the instructions and not change when the schedule of reinforcement changes. A series of four experiments were designed to expand our knowledge of why rules produce this insensitivity. Once responding had been established on a multiple DRL 6 s/FR 18 schedule of reinforcement, an unannounced change in contingencies was introduced. All four experiments contained sufficient numbers of participants to permit the application of statistics in the data analysis. The experiments investigated (a) the effects of instructions on the speed and accuracy with which a new behavior was learned; (b) the "sensitivity" or rapidity of adjustment of the behavior to an unannounced change in contingencies as a function of whether it had been learned with the help of instructions. Experiment I attempted to replicate previous "insensitivity of rule-governed behavior" findings and examined schedule sensitivity as the multiple DRL 6s/FR 18 schedule changed to extinction/extinction. Experiment 2 aimed to increase this series of studies' ecological validity, by changing the schedule to something other (FR 18/FR 18) than extinction/extinction. In an attempt to counteract the insensitivity effect, Experiments 1 and 2 also examined the effect of a simple warning that "conditions may change at any time" on the behavior's rapidity of adjustment to changes in contingencies. In light of the results of the first two experiments, a selectionist framework was adopted in Experiments 3 and 4 to provide a procedural explanation for the insensitivity effect. The effects of the length of training, and of a variety of instructions were investigated in relation to the insensitivity of rule-governed behavior effect. Overall, the results have shown that instructions facilitate the acquisition of a new behavior. One of the studies has shown that providing a simple warning that conditions may change increased sensitivity in behavior following a change in contingencies. The results have suggested the elimination of length of training as a potential factor on the insensitivity of rule-governed behavior, in this specific series of studies. Providing a variety of instructions impacted significantly more on the Minimal Instructions groups' than the Elaborate Instructions groups' sensitivity to a change in contingencies. The results are explained in terms of the restriction in the behavioral repertoire during a change in contingencies