This paper reviews evidence pertaining to the operation of acquiescent response style in psychological tests and to the possible interpretation of acquiescence as a personality variable. Acquiescence was initially conceptualized in terms of consistent individual differences in the probability of agreeing with items that subjects are. at a loss to answer, either because they lack knowledge or because they are, uncertain about the meaning of the item or because they cannot evaluate the desirability of the response. In addition to this agreement tendency, an acceptance tendency was empirically differentiated and conceptualized in terms of consistent individual differences in the tendency to accept many heterogeneous characteristics as descriptive of the self. Lindauer, 1963;Holtzman, 1965;Loevinger, 1959; McGee, 1962c; Messick, 2961a;Moscovici, 1963;Rorer, 1965 In attempt,ing to assimilate these findings into a coherent theory of acquiescence, should we account for some of the differences in results in terms.of contaminating influences in the acquiescence measures that might have distorted the correlations? Should we account f'or some of the differences on the basis of subject variations, not only in terms of selection effects but also on the grounds that acquiescence may function in different ways in different types of' people? Should we look for consistencies in the patterns of results for particular kinds of acquiescence measures and consider the possibility that more than one dimension of acquiescence may be operating?In keeping with the spirit of the variable under consideration, the answer is "yes" to all of these questions.-3-
Conceptions of Acquiescence Acquiescence and Test PerformanceTo Cronbach (19~,6) it seemed obvious that if a subject knew the answers to true-false achievement items, he would respond in terms of his knowledge -4-and that if a response set were to influence test performance, it must operate on items the subject did not know. 'Ihus, acquiescence should operate primarily on items that the subject is at a loss to answer, either because he is uninformed on the topic in question or because he is uncertain about the meaning of the items. This type of reasoning has led to the use of extremely difficult information items to measure acquiescence (Gage, Leavitt, & Stone, 1957).On items where the answer is not known, the subject is thought to have a characteristic probability of responding "true"; Le., he is assumed to toss a loaded coin. However, in most simple ways of scoring acquiescence on true-false tests, the tendency to say "tzrue ' and the number of items not known or thought ambiguous are confounded. A sUbject may receive a particular score either because he displayed a high probability of responding "true" to a relatively small number of ambiguous items or a smaller probability of responding IItrue" to a greater number of ambiguous items. 'Ihus, reliable individual differences in typical acquiescence scores do not necessarily reflect individual differences in the tendency ...