TESTS FOR THE EXTENT OF GENERALIZATION OF CLASSICAL CONDITIONING UNDER IDENTICAL TRAINING SCHEDULES.
INDIANA UNIV BLOOMINGTON
Pagination or Media Count:
The possibility that equal rates of reinforcement may produce generalization between two dissimilar conditioned stimuli is important both for understanding generalization itself and for interpreting null outcomes resulting from within-S comparisons of two conditioning methods. Conventional procedures for testing the extent of generalization are not applicable, but suitable trial arrangements do allow tests of this question, the number and nature of these tests depending on a whether the reinforcement schedule is continuous or partial, and b whether conditioning can be measured on training trials. An application of one of the tests suggested that generalization of the conditioned GSR was total between a tone and a light. The statistical power of the tests under optimal trial arrangements is considered, and the implications of the result of the GSR study are discussed. Author