Search

Considerable attention has been given recently to the perplexing problem of the relationship between attitudes and behavior (Ehrlich, 1969; Deutscher, 1969; Warner and DeFleur, 1969; Ajzen et al., 1970; Lastrucci, 1970; Tarter, 1970). Methodological issues have been debated, the lack of congruence between attitudes and behavior has been examined, and a variety of solutions have been offered. We are told, for example, that the continuing poor correlation between measured attitudes and overt behavior requires us to search for the intervening variable or variables which apparently obtain (Ehrlich, 1969); or to employ the "direct observation" of the phenomenon under investigation rather than try to extrapolate from paper-and-pencil tests to behavior (Deutscher, 1969); or to develop theories that, in turn, enable us to find indicators that are more valid than the paper-and-pencil type (Lastrueci, 1970); or to simply admit that attitudes as "presently conceptualized play no real role in behavior" (Tarter, 1970). A number of important aspects of attitude research, however, have been obscured or omitted in these discussions. For the surprising aspect of the situation is not, as has often been implied, the lack of congruence between attitudes and behavior, but the persistent use of research designs that are inappropriate for the complexity of the subject under investigation. Further, it is surprising that researchers have failed to draw out other important implications of their research. That is, the fact that an attitude does not lead directly to a behavior does not justify the assertion that attitudes play no role in behavior, or that attitude research lacks significant implications for social life. The basis for this latter statement will be shown below in a discussion of the values of attitude research. It may be that inadequate research designs, valued for their simplicity rather than their appropriateness, are a manifestation of the "publish or perish" syndrome. In any ease, this paper attempts to outline the problems and the values of attitude research, and to demonstrate thereby that such research is of great significance for the understanding of social phenomena. If that significance seems minimized by those studies that have found poor correlation between attitudes and behavior, the fault lies both in the failure to create research designs that reflect the complexity of the problem and in the tendency to exalt the importance of the proximate causes of overt behavior. [ABSTRACT FROM AUTHOR]

In this paper we review recent efforts to examine attitude-behavior inconsistency as a multivariate research problem. Rather than either assuming a one to one relationship or documenting the contrary; recent researchers have reconceptualized the problem as identifying the conditions which affect the relationship. Although numerous variables have been examined in the recent literature, most of the multivariate research seems to be addressed to three underlying issues: one, the extent to which measurement invalidity may depress the observed attitude-behavior relationship; two, the extent to which other competing but unmeasured attitudes may depress the observed attitude-behavior relationship; and three, the extent to which attitude incongruent social norms depress the observed attitude-behavior relationship. [ABSTRACT FROM AUTHOR]