Rudimentary determinants of attitudes: Classical conditioning is more effective when prior knowledge about the attitude stimulus is low than high☆

Petty and Cacioppo's (1981, 1986) elaboration likelihood model of persuasion predicts that the classical conditioning of human attitudes is similar to other peripheral attitude change mechanisms in that conditioning should be more powerful when preexposure to and prior knowledge about the conditioned stimulus is low rather than high. To test this hypothesis, neutral words (high prior knowledge) and pronounceable nonwords (low prior knowledge) were matched in terms of subjects' prior attitudes and served as the conditioned stimuli. Forty-three subjects participated in a 2 (Gender: male, female) × 3 (CS-US contingency: word followed by electric shock, nonword followed by electric shock, word and nonword paired randomly with electric shock) × 2 (Experimental Stimulus: word, nonword) mixed-model factorial design. As predicted, the classical conditioning of human attitudes was more effective when nonwords than when words were followed by electric shock. Experiment 2 was a simulation experiment in which 22 students from the same subject pool read descriptions of the experimental conditions used in Experiment 1 and responded to questions about (a) subjects' attitudes toward the stimuli following the various CS-US contingencies, and (b) how the experimenter wanted the subjects to rate the stimuli. Results revealed that subjects expected (and believed the experimenter anticipated) simply that whatever stimulus was associated with shock would also become disliked.