Experimental analysis of blocking in human operant behaviour

Abstract

The blocking effect is the reduced stimulus control achieved by an unfamiliar (target) stimulus, “X”, when it is paired during conditioning with a stimulus, “A”, that already has acquired control. Blocking has been relatively difficult to accomplish with nonverbal behavior in human participants, compared with either nonverbal behaviour in animals or verbal behaviour in humans. Thus my primary research purpose was to produce blocking in the nonverbal, operant behaviour of human participants. My secondary research purpose was to achieve blocking of verbal behaviour. My third research purpose was to evaluate the relationship between verbal and nonverbal results. Four experiments were conducted with introductory psychology students pressing keys in response to coloured stimuli on a computer monitor. In Experiments 1A and 1B, the blocking procedure was applied only to the negative discriminative stimulus (S−) which was paired with a response-cost contingency. Stimulus colours employed as the preconditioned and target stimuli were counterbalanced between Experiments 1A and 1B. Statistically significant results were obtained in the first case, but not in the second. Overshadowing likely occurred in both experiments, adding to blocking effects in Experiment 1A and subtracting from blocking effects in Experiment 1B. In Experiments 2A and 2B, overshadowing was better controlled and the blocking procedure was applied to both positive (S+) and negative (S−) stimuli. The key finding was blocking in Experiment 2B, with a large, statistically significant difference in response rates during the positive target stimulus. The success of blocking in Experiment 2B likely was a consequence of the powerful point-loss contingency employed with the negative stimulus, which greatly reduced the probability of generalized high-rate responding. It appears to be the first demonstration of blocking of a positive stimulus in an operant procedure employing human participants; it supports the continuity of learning principles across species. Verbal response measures in both Experiments 2A and 2B also suggested that blocking occurred. In Experiment 2A however, nonverbal measures failed to show blocking, suggesting that in humans verbal measures may be more sensitive to blocking manipulations than are nonverbal measures.