Association for Behavior Analysis International

Dozens of studies support Precision Teachings claim that fluent behavior ensures retention of what is learned, shows resistance to fatigue, is unaffected by distraction, and can easily be developed into a more complex task. A critical feature of acquiring fluent behaviour is fast practice, however, methods used to understand the significance of fast practice have been hampered by experimental confounds as a result of using human participants (Dougherty et al., 2004). A repeated acquisition procedure, first described by Sidman and Rosenberger (1967) to study teaching methods, has shown some promise in resolving these confounds (Porritt et al., 2009). Using hens as an animal analogue of human responding, our goal is to utilize this procedure to remove these confounds in studying the effects of fast practice and clarify the relationship between practice speeds and learning outcomes.

Previous studies have identified potential sources of competing stimulus control in tests for stimulus equivalence. The current experiment employed the Nintendo Wii remote as a response device to investigate whether competition between comparison stimuli would affect suboperant action dynamics. Following one-to-many training on conditional discriminations sufficient to establish three three-member equivalence classes, participants were presented with a test for equivalence responding that included five trial types. These included, (1) traditional equivalence trial in which the incorrect stimulus had been previously presented as a correct comparison stimulus and (2), trials, in which a novel unrelated word was provided as the incorrect comparison. In the remaining three trialtypes, the incorrect stimulus was either orthographically similar (3) to the sample stimulus, phonologically similar (4) or both (5). Results suggested that the previous history of reinforcement for the incorrect stimulus in the traditional equivalence trial gave rise to greater competition than phonological or orthographic similarity between the sample and incorrect comparisons.

The present study aimed to evaluate whether differential levels of resurgence would be observed under a context of reinforcer-controlled variation and a context of extinction-induced variation. College students were required to emit sequences of 5 presses on keys F and J of a keyboard. Reinforcers were produced only by the target sequence 1 (JFJJF) in Phase 1 (Reinforcement), and by the target sequence 2 (JJFJJ) in Phase 2 (Elimination). Extinction was in effect for all other sequences in both phases. In Phase 3, the participants were assigned to 2 groups: Variation and Extinction. For the Variation Group, a variation requirement was in effect, such that infrequent and remote sequences had greater chance to be reinforced than frequent and recent ones (the target sequence 1 never produced reinforcers); and for the Extinction Group, none of the emitted sequences produced reinforcers. Resurgence of the target sequence 1 occurred mainly for the Extinction Group. Also, the level of sequence variation (U value) was higher in the last phase than in the previous ones for the Extinction Group, and it was unsystematic for the Variation Group. It was concluded that reinforcer-controlled variation may weaken resurgence due to the concurrent reinforcement of other behaviors.