Relative Influences of Affect and Cognition on Behavior: Are Feelings or Beliefs More Related to Blood Donation Intentions? Sally D. Farley 1 and Mark F. Stasson 2 1 Virginia Commonwealth University and Albright College, Harrisburg, USA and 2 Metropolitan State University, St. Paul, USA Abstract. This study tested the relative predictive power of affect and cognition on global attitude and behavioral intention within the tripartite model of attitude structure. Participants (N = 264) completed questionnaires that included an item regarding blood donation experience, five semantic differential items, four behavioral intention items, and one global attitude item. Participants were randomly assigned to either an affective or cognitive instruction set for the semantic dif- ferential items. As predicted, semantic differentials were more highly correlated with both global attitude and behavioral intention when completed under the affective instructions than under the cognitive instructions. In addition, donors’ and non-donors’ attitudes on the semantic differential scales were distinguished from one another only when they were elicited under the affective instruction set. Results provide support for the tripartite model of attitude structure. Future research should examine the relative importance of affect and cognition in less emotion-laden domains. Key words: affect vs. cognition, attitude structure, blood donation Historically, attitude has been defined in terms of an acquired behavioral disposition (Campbell, 1963), degree of positive or negative evaluation, or “a men- tal and neural state of readiness . . . exerting a . . . dynamic influence upon” behavior (Allport, 1935, p. 810). One widespread conceptualization of attitude is the tripartite model (Eagly & Chaiken, 1993). Ac- cording to the tripartite model, an attitude is com- prised of three correlated, but distinct, components: affect, cognition, and behavior. Affective measures of attitude include self-report measures of feelings about attitude objects and physiological measures such as blood pressure and heart rate. Cognitive mea- sures may include beliefs about attitude objects and judged evaluative favorability toward attitude ob- jects. Behavioral indices typically involve self-report measures of past behavior, behavioral intentions, or observations/reports of actual behavior (Fishbein & Ajzen, 1975). The tripartite model of attitudes has been a popu- lar and enduring conceptualization of attitude, though it has not been endorsed uniformly in the lit- erature. Perhaps the best-known criticisms of the DOI: 10.1027//1618-3169.50.1.55 ” 2003 Hogrefe & Huber Publishers Experimental Psychology 2003; Vol. 50(1): 55Ð62 model have focused on doubts surrounding the exis- tence of strong links between the affective and cogni- tive components on the one hand and the behavioral component on the other (LaPiere, 1934; Wicker, 1969). Wicker’s (1969) paper sparked the consis- tency controversy, challenging the assumption that people possessed stable, underlying attitudes that in- fluence behavior. Wicker maintained that attitudes were, at best, only weakly related to overt behavior based on an average correlation of .15 between atti- tudes and behavior in the 42 studies reviewed. Subse- quent work illustrated that correspondence between measures is an important moderator of attitude-beha- vior consistency (Fishbein & Ajzen, 1974; 1975), with larger correlations between attitudes (affective or cognitive) and behavior when both measures refer to the same action, target, context, and time (Fish- bein & Ajzen, 1975). More recent models such as the MODE postulated by Fazio (1990) have articulated conditions under which attitudes are good predictors of decisions and behaviors. Similar to the ELM (Petty & Cacioppo, 1986), the MODE model main-