Social Cognition, Vol. 25, No. 5,2007, pp. 582-602 THE ADVANTAGES OF AN INCLUSIVE DEFINITION OF ATTITUDE Alice H. Eagly Northwestern University Shelly Ghaiken Berkeley, CA In The Psychology of Attitudes, we provided an abstract—or umbrella—definition of attitude as "a psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor" (Eagly & Chaiken, 1993, p. 1). This definition encompasses the key features of attitudes—namely, tendency, entity (or attitude object), and evaluation.
This conception of attitude distinguishes between the inner tendency that is attitude and the evaluative responses that express attitudes. Our definition invites psychologists to specify the natureof attitudes by proposing theories that provide metaphors for the constituents of the inner tendency that is attitude. We advocate theoretical metaphors that endow attitudes with structural qualities. New efforts to contemplate the definition of attitude are welcome in light of innovations in attitude theory and research.
Researchers have the burden of figuring out whether the phenomena that they have discovered are compatible with definitions of attitude that emerged in the field in earlier years. After all, Allport (1935) may have believed that he had provided the definition for all time, and this definition lingered for decades in social psychology textbooks. However, his definition became too diffuse as attitude research developed in the second half of the 20th century.
If contemporary researchers suspect that the attitudinal phenomena that they have identified may not be compatible with established definitions, there are two possible outcomes: the definition should change, or researchers should think harder to understand how We thank Wendy Wood for comments on a draft of this article. Please address correspondence to AUce H. Eagly, Department of Psychology, Northwestern University, 2029 Sheridan Rd. , Evanston, IL 60208; E-mail: eagly@northwestem. edu or sc4@nyu. edu. 582
AN INCLUSIVE DEFINITION OF ATTITUDE 583 these phenomena are compatible with the established definition. So, given important theoretical and empirical developments in attitude research, the present is an excellent time for determining how phenomena ordinarily identified as attitudinal relate to current and alternative defirutions of attitude. Why should social psychologists bother to settle on a definition of attitude? Perhaps scientists should just do research and not worry so much about abstractions and labels.
Not so. A science without definitions of basic constructs would be chaotic. Defirutions identify fields of inquiry by setting their boundaries and distinguishing their questions from questions that deal with other phenomena. Precise definitions also foster valid measurement. They provide a framework that enhances theory development and empirical research in a community of scientists. We should therefore all welcome this latest exchange of ideas on the definition of the attitude concept.
THE ABSTRACT, UMBRELLA DEFINITION OF ATTITUDE We offer an abstract—or umbrella—definition of attitude that posits three essential features: evaluation, attitude object, and tendency (Eagly & Chaiken, 1993). Together these elements refer to an individual's propensity to evaluate a particular entity with some degree of favorability or unfavorability. Evaluation refers to all classes of evaluative responding, whether overt or covert, or cognitive, affective, or behavioral.
Evaluation thus encompasses the evaluative aspects of beliefs and thoughts, feelings and emotions, and intentions and overt behavior. None of these reactions need be consciously experienced by the holder of an attitude, although they may be conscious. This evaluative responding is directed to some entity or thing that is its object—that is, we may evaluate a person (George W. Bush), a city (Chicago), an ideology (conservatism), and a myriad of other entities. In the language of social psychology, an entity that is evaluated is known as an attitude object.
Anything that is discriminable or held in mind, sometimes below the level of conscious awareness, can be evaluated and therefore can function as an attitude object. Attitude objects may be abstract (e. g. , liberalism, religious fiindamentalism) or concrete (e. g. , the White House, my green raincoat) as well as individual (e. g. , Condoleezza Rice, my sister-in-law) or collective (e. g. , undocumented workers, European nations). The entity, or attitude object, yields the stimuli that elicit the evaluative responses that psychologists identify as attitudinal.
In addition, such stimuli can be varied—George W. Bush may be denoted by his physical image as encountered on television, his 584 EACLY AND CHAIKEN statements as presented in newspapers and other media, his policies and actions as expressed by his staff, and so forth. This attitude object aspect of definitions of attitude is important because it distinguishes the concept of attitude from other psychological constructs such as mood that involve evaluative reactions that are more diffuse because they are not directed toward a circumscribed entity.
ATTITUDES AS ACQUIRED BEHAVIORAL DISPOSITIONS Our definition, like most definitions of attitude, places attitudes inside the mind of the individual. Campbell (1963) provided a particularly cogent statement of this general approach in his discussion of acquired behavioral dispositions, which he defined as states of the person that come into being on the basis of some transaction with the environment. These states then affect subsequent responding.
Consistent with Campbell's treatment, attitudes do not exist at all until an individual perceives an attitude object (on a conscious or unconscious basis) and responds to it on an explicit or implicit basis. This first response should not be considered entirely learned from the environment because it may be influenced by evolved predispositions or at least by an initial bias that directs attention to the stimuli that constitute the attitude object (Buller, 2005).
Some fearful responses, for example, to snakes or spiders, appear to have some innate aspects (Oehman & Mineka, 2001), and, more generally, some attitudes have heritable precursors (e. g. , J. M. Olson, Vernon, Harris, & Jang, 2001; Tesser, 1993). Nonetheless, in Campbell's and our view, attitudes even toward entities such as snakes do not exist until an individual first encounters an instance of the entity. An individual who had o (conscious or nonconscious) exposure to snakes could not be said to have a negative attitude toward them but instead likely has the potential for an easily acquired negative attitude. An individual's first reaction to an exemplar of a particular category, such as a negative response in the case of a snake or a spider, leaves a mental residue that predisposes the individual to respond consistently with that residue on subsequent encounters with the same or other exemplars. The initial negative response to a snake or spider makes negative responses more likely in the future.
The evaluative mental residue of past experience, or attitude, exists as a hypothetical construct, or in Campbell's terms, as an acquired behavioral disposition—that is, an intervening tendency that hypothetically accounts for the covariation that scientists observe between stimuli denoting the attitude object and the evaluative responses elicited by these stimuli. An attitude is inside the person, not directly observable, and is manifested by covert and overt AN INCLUSIVE DEFINITION OF ATTITUDE 585 responses.
Some of these responses are observable by researchers—or can be made observable through special operations or instrumentation—and can thereby give evidence of the presence of the attitude. EVALUATIVE MENTAL RESIDUE AS A TENDENCY TO EVALUATE What is this residue of past experience that constitutes an attitude? We have advocated and continue to advocate an abstract definition of attitude that labels this residue as a tendency to evaluate. Along with attitude object and evaluation as two elements of our definition, tendency is its third key feature.
An individual's past experience establishes a tendency to respond with some degree of positivity or negativity to an attitude object—for example, negatively in the instance of spiders. Why do we invoke the term tendency rather than disposition or state or any of the other terms that Campbell (1963) identified as denoting acquired behavioral dispositions? The term tendency has appropriate connotations because it does not imply that the residue of past experience exists necessarily on an enduring basis or on a temporary basis. In sychology the term state has come to imply temporariness, and the term disposition to imply greater permanence. Therefore, neither term is suitable to refer to attitude. Our stance that attitudes can be short-term or long-term departs from earlier theorists such as Allport (1935) and Krech and Crutchfield (1948), who defined attitudes as enduring. Although many attitudes are enduring, others are not. Some encounters with attitude objects are so fleeting and trivial that the mental residue left by the encounter fades away with time.
To allow for such possibilities, it is unwise to restrict the concept of attitude in a temporal sense. Another advantage of the term tendency is that it does not necessarily imply that attitudes are accessible to consciousness. Although researchers traditionally dealt mainly with attitudes assumed to be consciously experienced, from a contemporary perspective it is important to acknowledge that the mental residue that constitutes attitude can exist anywhere on a continuum that extends from unconscious to fully conscious.
As these various considerations suggest, our goal is to provide a nonrestrictive definition that can serve as a broad umbrella for attitude research. Our definition thus is the following: Attitude is "a psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor" (Eagly & Chaiken, 1993, p. 1). Some may object to this definition because it does not specify the inner tendency beyond labeling it as evaluative. We purposefully avoid fur- 586 EAGLY AND CHAIKEN ther specification of the inner tendency, not because it has not, carmot, or should not be described in more detail.
On the contrary, our reasoning is that the description of this inner tendency inevitably changes as attitude research develops and different theoretical positions emerge, become popular, and then may erode. We thus envision that our definition of attitude provides an umbrella under which multiple conceptualizations of attitudes' inner tendency can flourish. In fact, we believe that our definition challenges psychologists to develop descriptions of that irmer tendency. The generality of our abstract definition of attitude thereby complements the specificity of the many models of the mental residue that is attitude.
Some of these models accord considerable complexity to this mental residue, whereas others propose a simple, unitary residue that merely conveys some degree of positive or negative evaluation. Efforts to model the psychological and physiological events that constitute this inner tendency will never end and should never end. THE IMPORTANCE OF DISTINGUISHING BETWEEN ATTITUDE AND ITS EXPRESSIONS In agreement with many other theorists (e. g. , Zanna & Rempel, 1988) and consistent with our treatment of attitude as an inner tendency of the person, we maintain that attitudes can be expressed through many different types of responses.
However, we disagree with some attitude theorists by objecting to definitions of attitude as a response per se. For example, some attitude researchers have defined attitude as evaluative judgments or affective or evaluative responses (e. g. , Albarracin & Wyer, 2001; Kruglanski & Stroebe, 2005). Evaluative judgments, and, more generally, overt or covert evaluative responses are best regarded as expressions, or manifestations, of the irmer tendency that constitutes attitude. Although all evaluative responses are of course attitudinal in the sense that they express attitudes, they are not synonymous with attitude itself.
We are also reluctant to define attitude as a categorization of the attitude object on the evaluative continuum (Zanna & Rempel, 1988) because categorization pertains to a particular process that is critical to forming attitudes rather than to the end result that is the evaluative mental residue of past experience. Attitude is thus a tendency or latent property of the person that gives rise to judgments as well as to many other types of responses such as emotions and overt behaviors.
Attitude theories have as one of their main goals the prediction of these responses. This theoretical separation between the inner tendency that constitutes attitude and evaluative responses is important because it fosters AN INCLUSIVE DEFINITION OF ATTITUDE 587 understanding of the relation between (a) evaluative tendencies, which are mental residues of past experience with the attitude object, and (b) current evaluative responding, which reflects a whole range of influences in addition to those that emanate from the inner tendency.
This distinction between attitude and its expression is fundamental to theory development concerning attitude change, attitude-behavior relations, and other attitudinal phenomena. The failure of some psychologists to distinguish between attitudinal judgments and attitude itself has instigated a debate about whether most, if not all, attitudes are unstable, emerging anew in each specific situation. Those who have made such constructionist arguments for instability have equated instability in expressions of attitudes with inconsistency in the evaluative tendency that constitutes attitude itself (e. . , Schwarz & Bohner, 2001; Wilson & Hodges, 1992; see also Schwarz, this issue). Our constructionist colleagues are thus entirely correct to argue that evaluative judgments may well be constructed on each occasion of encountering an instance of an attitude object. For example, the presence of an audience w^ith known views on an issue might cause a person to render relatively thoughtless, superficial evaluative judgments that have the goal of pleasing audience members (Prislin & Wood, 2005).
Such judgments may appear to be inconsistent with a person's earlier evaluative responses because they reflect the demands of the current situation as much as or even more than the influence of the preexisting mental residue that constitutes the person's attitude. We agree that context effects are pervasive. Context effects due to the presence of an audience or other cues are ubiquitous because evaluative judgments are not pure expressions of attitude but outputs that reflect information in the current situation as well as at least some aspects of the preexisting evaluative tendency.
The contemporaneous setting contains not only cues that elicit the inner attitude but also a wealth of information that provides new inputs to the attitude, activates an individual's goals, and provides standards against which to judge the current instantiation of the attitude object. Evaluative judgments and other evaluative responses such as behaviors and emotions emerge from this array of influences. The person's internal evaluative tendency, or attitude, is but one influence on evaluative responding, not he only influence. Moreover, aside from the influences that derive from the external situation, the relation between the evaluative tendency itself and a given evaluative expression is not one-to-one because each response only imperfectly maps evaluation. Like the items on a Likert scale, each response that expresses evaluation reflects other tendencies and states of the person (e. g. , personality traits, moods) as well as random error. 588 EAGLY AND CHAIKEN
Whereas judgments and other evaluative responses may be very labile depending on the judgment context, the inner tendency or latent construct that constitutes attitude may often be relatively stable, notwithstanding our definition's inclusion of temporary as well as enduring tendencies to evaluate. Instability of attitudes may seem to be present because people sometimes report what appear to be substantially different attitudes in new situations. Although instability could arise because of a genuine change in the inner tendency, some of this inconsistency may reflect more ephemeral contextual factors.
Therefore, apparent attitude change can erode over time, resulting in a minimally changed evaluative tendency. Due to the ir\fluence of contextual cues and new information, evaluative responding to an attitude object may merely vary around an average value that reflects the inner tendency that constitutes the attitude (e. g. , N. H. Anderson, 1971). To understand this variability, psychologists model the psychological processes that mediate between the person's evaluative tendency and the evaluative responses that are elicited in varied circumstances.
In other words, one of the main goals of attitude research is to provide theories that account for the effects of contextual variables on attitudes and their expression. If attitudes were always constructed anew on the basis of current inputs and whatever information happens to be salient at a given moment, psychologists would not have uncovered evidence of long-term stability of many attitudes. Demonstrations of stability include considerable consistency in the sociopolitical attitudes of civil rights activists over a 20-year period (Marwell, Aiken, & Demerath, 1987).
Also, people's attitudes toward their jobs have proven to be moderately consistent across a five-year time span in nationally representative survey data (Staw & Ross, 1985). In addition, a meta-analytic examination of a set of broadly defined attitudes that included self-esteem, work satisfaction, and life satisfaction found moderate stability over five to ten years, albeit less consistency over longer time periods ranging up to 40 years (Conley, 1984).
Also uncongenial to a strict constructionist position is the moderate degree of consistency that surveys have revealed across measures of political attitudes and ideologies when these variables are assessed, not by single items, but by measures of latent constructs that cumulate across items (e. g. , Moskowitz & Jenkins, 2004). A constructionist interpretation of consistency between attitudes across time points and between ideologically related attitudes would hold that the situational contexts in which these attitudes were constructed remained constant, thus yielding consistent constructions.
Such a high level of consistency of contexts seems quite implausible. AN INCLUSIVE DEFINITION OF ATTITUDE 589 THE METAPHORIC NATURE OF DESCRIPTIONS OF THE EVALUATIVE TENDENCY As we have indicated, researchers' quest to understand the nature of the inner tendency that constitutes attitude is an important tradition in research. Theorists have defined this construct in varying ways, depending on their particular theoretical frameworks.
Most notably, Fazio (1989) defined attitude as an association in memory between an attitude object and an evaluation (see also Fazio, this issue). This specification of the latent property that constitutes attitude follows from associative learning or network models of learning (e. g. , J. R. Anderson, 1983). This way of thinking about the inner tendency that constitutes attitude has been influential. A colleague of ours thus remarked several years ago that he could not imagine that the future would offer any other way of thinking about attitudes.
From his perspective, the object-evaluation association seemed to be the transcendental definition of the mental residue that is attitude. Our colleague was not prescient in his views of attitude theory. New metaphors continually emerge. Most notably, connectionism has fostered several provocative metaphors (Conrey & Smith, this issue). In some applications of this general approach, attitude objects are represented as nodes that are linked by implicational relations to unipolar valence nodes, as in Van Overwalle and Sieber's (2005) research.
In this model, the mental residue that is attitude consists of these positive and negative nodes and their links to the attitude object node. In another connectionist metaphor, Bassili and Brown (2005) proposed an attitudinal cognitorium that represents attitude as a module that contains atomistic elements representing memories and feelings that are associated with the attitude object. In any particular situation, a pattern of activation develops among these elements, and an overall evaluation may emerge from this activity.
These connectionist metaphors yield varied insights about the nature of attitudes. The classic metaphor for understanding the evaluative mental residue that constitutes attitude is the tripartite model whereby cognition, affect, and behavior are three omnipresent components of attitude (e. g. , Katz & Stotland, 1959; Krech & Crutchfield, 1948; Rosenberg & Hovland, 1960). This way of thinking about attitude surely has contributed to theorizing and has roots in earlier traditions of understanding the human psyche (McGuire, 1969).
Nonetheless, from a contemporary perspective, it seems misleading to maintain that that the mental residue of past experience with an attitude object invariably has three components, as stated in many textbook descriptions of attitudes. Attitudes may indeed reflect cognitive, affective, and behavioral expe- 590 EAGLY AND CHAIKEN riences with attitude objects, but there is little justification for claiming that these experiences produce three separable and omnipresent components of evaluative tendencies.
Many other possibilities exist for specifying the inner tendency, including, for example, the metaphors inherent in various cognitive consistency theories, such as the balance theory representation of attitude as a sentiment relation between two cognitive elements and the dissonance theory representation of attitude as a cognitive element (see Eagly & Chaiken, 1993). No matter which specifications of the inner tendency may be widely embraced at a particular point in time, we resist defining attitude in terms of any one of them.
Our reluctance stems in part from the fact that aU ways of specifying the tendency that constitutes attitude have been metaphoric because they do not have an inherent reality that allows them to be directly viewed or verified. Therefore, associative network models produce a mere metaphor, albeit a highly generative one, as do connectionist models, the three-component model, cognitive consistency models, and all other models. Researchers cannot directly observe object-evaluation associations or any of the other psychological entities that they have proposed as descriptions of attitudes.
The varied metaphors that have emerged for describing attitudes' inner tendency are important because they guide theorizing, yield a useful vocabulary for thinking about attitudes, and bring certain types of testable hypotheses to light. Moreover, the fact that a particular specification loses popularity does not negate its contribution to attitude theory. For example, Doob (1947) defined attitude as a learned, implicit anticipatory response, a treatment that adopted language from the Hullian learning theory framework, which was very popular in the 1940s.
Although contemporary attitude research is not guided by this particular theoretical metaphor, it enhanced understanding of attitudes at the time. In fact, as testimony to the enduring importance of prior theorizing, the implicit response aspect of Doob's definition anticipated the contemporary interest in implicit attitudes. Our abstract definition of attitude encompasses the various distinctive conceptualizations of the latent tendencies that constitute attitudes. A definition that provides an overall framework for attitude theory and research would ideally not change as new theories emerge and become popular.
Most theories of attitudes describe, with certain words and images, the mental residue that constitutes attitude. Each approach can promote important insights about attitudes, as long as it inspires testable hypotheses that yield sufficient levels of confirmation. The consistency o^ our umbrella definition with these metaphors derives from its f breadth, which allows it to transcend particular theoretical preferences, as well as from its precision, which identifies attitudes' three key ele- AN INCLUSIVE DEFINITION OF ATTITUDE 591 ments of evaluation, attitude object, and tendency.
This definition embraces psychologists' invention of new and shifting metaphors for understanding the mental residue that constitutes attitude. Although all such conceptions do not have equal value in furthering knowledge about attitudes, a multiplicity of metaphors produces far more understanding than adherence to a single metaphor. THE ANTECEDENTS AND CONSEQUENCES OF ATTITUDES Despite the difficulties of arguing that the inner tendency that is attitude necessarily has three components, the tripartite model, in a revised form, has continued to be important (e. . , Breckler, 1984; Zarma & Rempel, 1988). As theory and research developed, interpretations of the three-component idea gravitated toward meaning that the mental residue of experience with an attitude object can be formed or expressed through cognitive, affective, and behavioral processes. Why is it useful to retain a revised tripartite analysis? In terms of attitudes' antecedents, the labeling of them as cognitive, affective, and behavioral acknowledges differing assumptions concerning how attitudes come into being and change (see Eagly & Chaiken, 1993,1998).
For example, theories of message-based persuasion have traditionally assumed a cognitive learning process that follows from exposure to persuasive communications. Affective or emotional processes are inherent in several theories of attitude formation, including Zajonc's (1980, 1984) proposition that "preferences" (i. e. , evaluations) are based primarily on affective responses, unmediated by thinking. The idea that attitudes derive from behavioral responses has appeared in varying forms, including dissonance theory (Festinger, 1957) and self-perception theory (Bem, 1972).
In describing the manner in which attitudes are expressed, this revised tripartite, or neotripartite, analysis also points to cognition, affect, and behavior, but as three types of evaluative responding that attitude objects elicit. The cognitive aspect of attitudes consists of associations that people establish between an attitude object and various attributes that they ascribe to it (Fishbein & Ajzen, 1975). The affective aspect of attitudes consists of feelings and emotions and the as yet only partially mapped physiological responses that may accompany affective experience (Schimmack & Crites, 2005).
The behavioral aspect of attitudinal responding refers to overt actions toward the attitude object as well as to intentions to act. Cogrutions, affects, and behaviors all express positive or negative evaluation of more or less extremity (and sometimes they are neutral in their evaluative implications). 592 EAGLY AND CHAIKEN Despite the insights that have followed from thinking about cognitive, affective, and behavioral antecedents and consequences of attitudes, the idea of three components has often been overstated.
One concern is that the three components have frequently failed to appear as neatly separable in straightforward factor analytic tests (see Eagly & Chaiken, 1993). Regardless of their empirical separability, there is no necessity that attitudes include all three aspects—cognition, affect, and behavior—either at the point of their formation or at the point of attitudinal responding. Attitudes can be formed or expressed primarily or exclusively on the basis of any one of the three tj^es of processes or some mix of these processes.
The formation of attitudes through affective, cognitive, or behavioral processes establishes associations that are linked to the attitude object and can become part of the mental residue that is attitude. These associations can reflect one or a mixture of the affective, cognitive, and behavioral precursors of attitude. Many psychologists have argued that affect, understood as feelings and emotions, surely is a necessary component of attitudes.
Some of these psychologists have merely equated evaluative emotions and feelings with abstract evaluation (a view that we do not endorse), but others think that emotions and feelings are truly the core of attitudes. However, contrary to the idea that affect is necessarily primary to attitudes (e. g. , Zajonc, 1984), research has shown that attitudes may be based mainly on cognitions or on affects (see Chaiken, Pomerantz, & Giner-Sorolla, 1995). Once the evaluative extremity of the cognitive and ffective constituents of attitudes is controlled, affective constituents of attitudes appear to be more accessible in memory for attitudes that are primarily affectively based, and cognitive constituents for attitudes that are primarily cognitively based (Giner-Sorolla, 2004). Given such findings, affect may not be a more essential component of attitudes than cognitions or behaviors. Affect may well be special, but this matter is an empirical and theoretical issue, not a definitional one.
In summary, the neotripartite cognitive-affective-behavioral analysis is best regarded as providing a convenient terminology for referring to differing aspects of the processes that are involved in forming and expressing attitudes. Such an analysis favors metaphoric representations of attitudes' inner tendency as potentially multifaceted and not merely taking the form of a unitary abstract evaluation. IMPLICIT AND EXPLICIT ATTITUDES Some recent elaborations of attitude theory distinguish between implicit and explicit attitudes. In current parlance, explicit attitudes are evalua-
AN INCLUSIVE DEFINITION OF ATTITUDE 593 tions that are reported by the person who holds the attitude. The person who self-reports an attitude is aware of his or her attitude or at least aware of the expression of the attitude that is elicited by the measuring instrument. Until recent years, the predominant view in research, although not generally stated, was that attitudes are explicit in the sense that people are largely aware of their own evaluative tendencies. In contrast, implicit attitudes are those that people do not consciously recognize (e. g. Greenwald & Banaji, 1995). Although earlier theorizing was not devoid of acknowledgment that people can be unaware of their attitudes, this recognition has come to the forefront of theoretical and empirical efforts to understand attitudes. This direction represents an important step beyond the strands of theorizing about implicit attitudes that were embedded in some earlier perspectives. For example, in addition to Doob's (1947) introduction of the idea of implicit anticipatory evaluative responses, Zajonc (1980, 1984) contributed seminal ideas about evaluative preferences ithout conscious inferences, and the unobtrusive measures movement advocated assessing attitudes without people being aware that their attitudes are under scrutiny (Webb, Campbell, Schwartz, & Sechrest, 1966). There is general agreement among contemporary attitude researchers that, even when a person does not have conscious access to an attitude, it may be automatically activated by the attitude object or cues associated with the object (e. g. , Bargh, Chaiken, Govender, & Pratto, 1992).
Attitudes that are implicit in this sense can direct responding, perhaps mainly spontaneous behaviors, as some researchers have argued (e. g. , Dovidio, Brigham, Johnson, & Gaertner, 1996). In contrast, explicit attitudes, to which one has conscious access, are presumed to be activated in a more deliberative marmer that requires cognitive effort. Such attitudes may better predict behaviors that are under volitional control (Ajzen & Fishbein, 2005).
These differing implications of implicit and explicit attitudes for spontaneous and volitional behaviors illustrate the potential utility of studying implicit attitudes. One manifestation of this distinction is Wilson, Lindsey, and Schooler's (2000) conception of dual attitudes, by which people may have an implicit attitude and an explicit attitude toward the same attitude object. Wilson and his colleagues assumed that new information often changes an existing attitude, creating a new explicit attitude.
Yet, the prior attitude may continue to be present, but often in implicit form. This approach acknowledges the constructionist point that attitudinal expressions are often unstable. By adding the principle that an "old" attitude may endure in implicit form, this theory also takes account of the common tendency for newly changed attitudes to erode and revert to a value more typical of the average of past expressions of the attitude. 594 EAGLY AND CHAIKEN Consistent with some aspects of the Wilson et al. 2000) position, Gawronski and Bodenhausen (2006) have proposed that implicit attitudes derive from affective associations that are activated automatically by stimuli associated with an attitude ohject (see also Gawronski & Bodenhausen, this issue). They have assumed that these (not necessarily unconscious) associations are activated, regardless of whether a person would regard them as accurate or inaccurate, true or false. These associations thus do not possess a subjective truth value.
Gawronski and Bodenhausen have also assumed that explicit attitudes derive from evaluative judgments that take the form of propositions that the holder of the attitude regards as accurate, or true. For example, an individual might have automatic negative affective associations to an illegal immigrant but derive a more positive evaluation from activating the propositional knowledge that most illegal immigrants are merely trying to provide for their impoverished families. Despite this theory's ties with particular distinctions between more spontaneous and more deliberative processing (e. . , Gilbert, 1991; Strack & Deutsch, 2004), the idea that nonendorsed associations, or extrapersonal associations, can underlie attitudes remains a point of contention (Fazio & Olson, 2003). An implicit measure that reduces their influence provides estimates of attitude that are more in agreement with those produced by other explicit and implicit measures (M. A. Olson & Fazio, 2004; Han, Olson, & Fazio, 2006). Measuring instruments that exclude nonendorsed (automatic or otherwise) associations may well yield more valid estimates of the mental residue that constitutes attitude.
Even though we agree that attitudes need not be conscious to influence responding, we maintain that there is only one underlying evaluative residue, regardless of whether it takes a urutary or multifaceted form or whether the holder of the attitude is aware or unaware of its presence. Regarding attitudes as dual or multiple or divisible into implicit and explicit attitudes captures certain phenomena but may not provide the most productive direction for theorizing about the mental residue that constitutes attitude.
It is tempting to take such a direction, given the dissociations that sometimes occur between estimates of attitudes from implicit and explicit measures. However, such dissociations can reflect a variety of differing influences. Postulating two types of attitudes does not relieve the burden of understanding the theoretical processes that underlie these dissociations. What might produce dissociations between estimates of attitudes from implicit and explicit assessments? Most obviously, an ttitude may not be validly expressed on explicit measures due to the influence of situational pressures such as norms fostering politically correct respond- AN INCLUSIVE DEFINITION OF ATTITUDE 595 ing. Moreover, both implicit and explicit measures impose certain situational and methodological constraints on responding. These constraints differ, given that implicit measures are ordinarily based on response latencies, usually of dichotomous responses, and explicit measures on deliberative endorsement of evaluative statements, generally on paper-and-pencil rating scales.
In addition, as we have noted, some implicit measures are constructed so that they elicit nonendorsed as well as endorsed associations—most notably, the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998). Also, although the fundamental assumption of implicit measures is that they assess only automatic responding, there is no assurance that currently available measures satisfy this criterion (Conrey, Sherman, Gawronski, Hugenberg, & Groom, 2005).
In addition, there is no assurance that respondents are necessarily unaware of implicitly assessed attitudes (Olson & Fazio, 2004). Given that implicit measures, like explicit measures, merely allow investigators to infer the evaluative content of the mental residue that is attitude, more research is needed before psychologists can decide whether the two types of measures reveal different aspects of attitude. These issues, while complex, should inspire researchers to explore the implications of more and less consciously experienced attitudes.
THE STRUCTURE OF ATTITUDES To better understand why differing t3^es of measuring instruments and situations can produce different estimates of attitudes, we suggest that researchers develop metaphorical accounts of the inner tendency that address its potential complexity. According to such theories of the latent attitudinal tendency, an attitude is a repository of the individual's past experience with the attitude object. Depending on external cues, this mental residue is often activated as a unitary overall evaluation of the attitude object but, given ts potentially multifaceted nature, is sometimes activated only in part. Imputting structure to attitudes fosters understanding of their complexity. The neotripartite analysis, by which attitudes may have different types of antecedents, provides one important metaphor describing the mental residue that constitutes attitude. The differing types of inputs to attitudes can be represented in memory as mental associations liriked to the attitude object (Eagly & Chaiken, 1998).
When positive and negative experiences become attached to attitude objects in people's minds, they acquire mental associatior\s that join the attitude object to relevant prior experience, which may have taken the form of cognitive, affective, or be- 596 EAGLY AND CHAIKEN havioral responding. In this sense, the inner tendency that is attitude may be saturated with associations involving cognitions, affects, or behaviors, but not necessarily with all three types of associations. Cognitive associations, or propositional knowledge, may be present (e. . , illegal immigrants are trying to support their impoverished families) along with affective associations (e. g. , fear of immigrants) and memories of behaviors (e. g. , signing of an anti-immigrant petition). These mental associations may have regularities that lead psychologists to ascribe various structural properties to them. We have termed some of these properties intra-attitudinal structure (Eagly & Chaiken, 1993, 1998). For example, mental associations may be more or less evaluatively inconsistent with one another (i. e. , ambivalent).
Although individuals often do have an abstract overall evaluation of many attitude objects, perhaps in the form of an easily retrieved evaluation of, for example, the war in Iraq, our neotripartite metaphor would also endow this attitude with many more specific mental associations representing, perhaps, knowledge of this war, prior emotional reactions to it, and prior behaviors and behavioral intentions in relation to it. In addition, because people form attitudes toward many different entities that themselves are interrelated, there may also be a larger structure whereby attitude objects are connected to other ttitude objects. For example, the war in Iraq may be associated in many people's minds with George W. Bush and Dick Cheney. We have termed these more global structures that encompass more than one attitude inter-attitudinal structure. Many important properties derive from inter-attitudinal structure (e. g. , ideologies formed by linked attitudes) as well as from intra-attitudinal structure (e. g. , ambivalence, evaluative-cognitive consistency). Consistent with the neotripartite analysis of attitudes' antecedents, intra-attitudinal and inter-attitudinal structure reflect differing ways that attitudes have been formed.
As this model suggests, people can form attitudes experientially based on direct or indirect cognitive, affective, or behavioral responding to the attitude object. This intra-attitudinal mode of attitude formation entails storing the information produced by one's responses as associations between the attitude object and these responses. As evaluative meaning is abstracted from these associations, an overall abstract attitude may be formed as a generalization from these more elementary associations.
Alternatively, one can form an attitude in a top-down manner by forging linkages between the attitude object and other attitude objects. These links are stored, along with the target attitude itself. This mode of attitude formation entails an inference by which a new attitude is deduced from an attitude that has already been formed, often a more abstract or general attitude (Eagly & Chaiken, 1995). For example, in one demonstration of such a process, participants deduced AN INCLUSIVE DEFINITION OF ATTITUDE 597 heir attitude toward a news item concerning sex discrimination from their existing attitude on the general issue of equal rights for women (Prislin, Wood, & Pool, 1998). Aspects of attitude structure that are available in memory and therefore could potentially be activated are not necessarily accessible at any particular point in time so that they influence evaluative responding. The accessibility of attitude structure reflects the demands of the perceiver's situation as well as a variety of other factors such as personality traits and chronic goals.
The aspect of a person's attitude toward, for example, gay marriage, that is activated might or might not include an abstract evaluation and might or might not include the differing intra-attitudinal and inter-attitudinal associations to gay marriage that the individual has stored in memory. How do ideas about the complexity of attitude structure account for findings that have led researchers to postulate that people hold more than one attitude in relation to many attitude objects—for example, attitudes that are old and new, implicit and explicit, or derived from associative and propositional processes?
If the evaluative mental residue has been laid down by many encounters with the attitude object, it no doubt has complex structure. Different aspects of that residue of past experience may form the basis of attitudinal responding in particular circumstances. To understand complex structure, consider, for example, people's attitudes toward their mothers. An affect-laden attitude is ordinarily formed by the young child, and this attitude is elaborated and changed by numerous experiences as the child matures.
For example, a rebellious teenager may form a negative attitude in response to a mother's restrictions. The attitude of the mature son or daughter becomes more complex because of the acquisition of knowledge about the mother in a wide range of settings. Given such a complex structure, the adult child may appear sometimes to revert to a childish or adolescent attitude when interacting with the mother, perhaps without awareness of the activation of this aspect of the attitude.
Because the residue of past experience that constitutes this attitude is multifaceted, it can be activated in various forms, depending on situational cues and their relatioris to the stored mental associations linked to the mother. Attitudes toward richly experienced attitude objects such as family members and one's own nation can be particularly vulnerable to context effects because contents of memory with differing evaluative implications may be activated in varying circumstances.
However, some complex attitudes are coherently organized in the sense that most mental associations are consistent with the person's overall evaluation of the attitude object. If so, the attitude may not be especially vulnerable to con- 598 EAGLY AND CHAIKEN text effects because the person's evaluation would be similar, regardless of whether the overall evaluation or only some mental associations are accessed (see Eagly & Chaiken, 1998; Chaiken et al. , 1995).
Nonetheless, even for richly experienced attitude objects, the person has just one evaluative residue of past experience, albeit one that encompasses varied mental associations. Attitudes that are fairly important but not so complexly experienced in the past may ordinarily be more unitary because they are based on a stable set of attributes that are chronically associated with an attitude object (Van Harreveld & Van der Phght, 2004; Van Harreveld, Van der Plight, De Vries, & Andreas, 2004).
For example, attitudes toward current social policies pertaining to issues such as taxes or immigration may be based on relatively few attributes that are often accessed and therefore foster a similar abstract evaluation. Such an attitude may nevertheless be expressed somewhat differently under the influence of varying contextual stimuli. Still, attitudes of this general type, such as prejudices toward ethnic groups and evaluations of political parties, have proven to be difficult to change and often endure over substantial time periods.
We suspect that the majority of the attitudes historically of interest to social scientists are of this type. In contrast to these moderately important, often sociopolitical, attitudes commonly studied by social scientists, attitudes that are more important or less important are likely to be associated with the appearance of greater attitudinal volatility. As we have indicated, attitudes accompanied by an especially rich repertoire of associations could yield evaluative responses of inconsistent valence because of the activation of only some of these associations in certain circumstances.
In addition, trivial attitudes or attitudes newly formed in laboratory experiments would also appear to be volatile because the mental residue of past experience is relatively impoverished in terms of prior knowledge, emotional experiences, or memories of overt behaviors. Appropriate analysis of the underlying structure of attitudes could clarify these regularities. For example, some connectionist metaphors for attitudes' inner tendency (e. g. , Bassili & Brown, 2005; Van Overwalle & Sieber, 2005) could be easily elaborated to model critical features of attitude structure.
ONCE MORE, THE DEFINITION OF ATTITUDES Finally, we ask once more whether our definition of attitude as a psychological tendency that is expressed by evaluating a particular entity with some degree offavor or disfavor remains viable given current specifications of this tendency. The answer to this question is yes. As we have shown in AN INCLUSIVE DEFINiriON OF ATTirUDE 599 this article, the generality and precision of our definition are congenial to the various ways of thinking about attitudes' inner tendency that theorists have proposed.
Moreover, this mental residue of experience with an attitude object may take any of a variety of simple or more complex forms because it can include a wide range of cognitive, affective, and behavioral associations. The challenge for attitude theorists is to propose models of attitude that map these forms in ways that further the understanding of attitudinal phenomena. REFERENCES Ajzen, I. , & Fishbein, M. (2005). The influence of attitudes on behavior. In D. Albarracin, B. T. Johnson, & M. P. Zanna (Eds. ). The handbook of attitudes (pp. 173-221). Mahwah, NJ: Erlbaum. Albarracin, D. , & Wyer, R. S. (2001).
Elaborative and nonelaborative processing of a behavior-related persuasive communication. Personality and Social Psychology Bulletin, 27, 691-705. Allport, G. W. (1935). Attitudes. In C. Murchison (Ed. ), Handbook of social psychology (pp. 798-844). Worcester, MA: Clark University Press. Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA: Harvard University Press. Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78,171-206. Bargh, J. A. , Chaiken, S. , Govender, R. , & Pratto, F. (1992). The generality of the automatic attitude activation effect.
Journal ofPersonality and Social Psychology, 62, 893-912. Bassili, J. N. , & Brown, R. D. (2005). Implicit and explicit attitudes: Research, challenges, and theory. In D. Albarracin, B. T. Johnson, & M. P. Zanna (Eds. ), The handbook of attitudes (pp. 543-574). Mahwah, NJ: Erlbaum. Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed. ), Advances in experimental social psychology (Vol. 6, pp. 1-62). New York: Academic Press. Breckler, S. J. (1984). Empirical validation of affect, behavior, and cognition as distinct components of attitude. Journal of Personality and Social Psychology, 47,1191-1205.
Buller, D. J. (2005). Adapting minds: Evolutionary psychology and the persistent quest for human nature. Cambridge, MA: MIT Press. Campbell, D. T. (1963). Social attitudes and other acquired behavioral dispositions. In S. Koch (Ed. ), Psychology: A study of a science (Vol. 6, pp. 94-172). New York: McGraw-Hill. Chaiken, S. , Pomerantz, E. M. , & Giner-SoroUa, R. (1995). Structural consistency and attitude strength. In R. E. Petty & J. A. Krosnick (Eds. ), Attitude strength: Antecedents and consequences (pp. 387-^12). HiUsdale, NJ: Erlbaum. Conley, J. J. (1984).
The hierarchy of consistency: A review and model of longitu- 600 EAGLY AND CHAIKEN dinal findings on adult individual differences in intelligence, personality and self-opinion. Personality and Individual Differences, 5,11-25. Conrey, F. R. , Sherman, J. W. , Gawronski, B. , Hugenberg, K. , & Groom, C. J. (2005), Separating multiple processes in implicit social cognition: The quad model of implicit task performance. Journal of Personality and Social Psychology, 89,469-487. Conrey, F. R,, & Smith, E. R. (2007). Attitude representation: Attitudes as patterns in a distributed, connectionist representational system.
Social Cognition, 25(5), 718-735, Doob, L, W. (1947), The behavior of attitudes. Psychological Review, 54,135-156, Dovidio, J, F,, Brigham, J, C, Johnson, B. T,, & Gaertner, S, L, (1996), Stereotyping, prejudice, and discrimination: Another look. In N, Macrae & C, Stangor & M, Hewstone (Eds. ), Stereotypes and stereotyping (pp. 276-319). New York: Guilford. Eagly, A. H,, & Chaiken, S. (1993), The psychology ofattitudes. Fort Worth, TX: Harcourt. Brace, Jovanovich, Eagly, A, H,, & Chaiken, S. (1995), Attitude strength, attitude structure, and resistance to change. In R, E, Petty & J, A. Krosnick (Eds. , Attitude strength: Antecedents and consequences (pp. 413^32). Mahwah, NJ: Erlbaum. Eagly, A, H. , & Chaiken, S, (1998), Attitude structiire and function. In D. Gilbert, S. Fiske, & G. Lindzey (Eds. ), The handbook ofsociai psychology (4th ed. , pp. 269-322). New York: McGraw-Hill. Fazio, R. H, (1989), On the power and functionality of attitudes: The role of attitude accessibility. In A, R. Pratkanis, S. J, Breckler, & A. G. Greenwald (Eds,), AthfMdesfrMcfureflndyMnrt! on(pp, 153-179). Hillsdale, NJ: Erlbaum, Fazio, R, H, (2007), Attitudes as object-evaluation associations of varying strength.
Social Cognition, 25(5), 603-637, Fazio, R H & Olson, M, A. (2003). Implicit measures in social cognition research: Their meaning and uses. Annual Review of Psychology, 54,297-327, Festinger, L, (1957), A theory ofcognitive dissonance. Evanston, IL: Row, Peterson, Fishbein, M,, & Ajzen, I, (1975), Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Gawronski, B. , & Bodenhausen, G, V, (2006), Associative and propositional processes in evaluation: An integrative review of implicit and explicit attitude change.
Psychological Bulletin, 132, 692-731, Gawronski, B. , & Bodenhausen, G. V. (2007), Unraveling the processes underlying evaluation: Attitudes from the perspective of the APE Model. Social Cognition, 25(5), 687-717. Gilbert, D. (1991). How mental systems believe. American Psychologist, 46, 107-119. Giner-Sorolla, R. (2004), Is affective material in attitudes more accessible than cognitive material? The moderating role of attitude basis, European Journal of Social Psychology, 34, 761-780, Greenwald, A, G. , & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes.
Psychological Review, 102, 4-27, AN INCLUSIVE DEFINITION OF ATTITUDE 601 Greenwald, A, G,, McGhee, D, E,, & Schwartz, J, K, L, (1998). Measuring individual differences in implicit cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74,1464-1480, Han, H, A,, Olson, M. A. , & Fazio, R, H, (2006), The influence of experimentally created extrapersonal associations on the Implicit Association Test, Journal of Experimental Social Psychology, 42, 259-272. Katz, D,, & Stotland, E. (1959). A preliminary statement to a theory of attitude structure and change.
In S. Koch (Ed. ), Psychology: A study ofa science (Vol. 3, pp. 423-475). New York: McGraw-Hill. Krech, D. , & Crutchfield, R. S. (1948), Theory and problems ofsociai psychology. New York: McGraw-Hill. Kruglanski, A, W,, & Stroebe, W. (2005). The influence of beliefs and goals on attitudes: Issues of structure, function, and dynamics. In D, Albarracin, B, T. Johnson, & M. P, Zanna (Eds. ). The handbook of attitudes (pp. 323-368). Mahwah, NJ: Erlbaum. Marwell, G. , Aiken, M. T,, & Demerath, N, J, (1987). The persistence of political attitudes among 1960s civil rights activists.
Public Opinion Quarterly, 51, 359-375, McGuire, W, J. (1969). The nature of attitudes and attitude change. In G, Lindzey & E, Aronson (Eds,), Handbook of social psychology (2nd ed,, Vol, 3, pp. 136-314). Reading, MA: Addison-Wesley. Moskowitz, A, N,, & Jenkins, J, C, (2004), Structuring political opinions: Attitude consistency and democratic competence among the U. S. mass public. Sociological Quarterly, 45, 395^19. Oehman, A. , & Mineka, S. (2001), Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108,483-522.
Olson, J, M,, Vernon, P, A,, Harris, J, A. , & Jang, K. L. (2001). The heritabiUty of attitudes: A study of twins. Journal of Personality and Social Psychology, 80, 845-660. Olson, M, A. , & Fazio, R. H, (2004), Reducing the influence of extrapersonal associations on the Implicit Association Test: Personalizing the IAT, Journal of Personality and Social Psychology, 86,653-667, Prislin, R,, & Wood, E. (2005), Social influence in attitudes and attitude change. In D, Albarracin, B, T, Johnson, & M, P, Zanna (Eds,), The handbook of attitudes (pp. 671-706).
Mahwah, NJ: Erlbaum. Prislin, R. , Wood, W. , & Pool, G, J. (1998), Structiiral consistency and the deduction of novel from existing attitudes. Journal of Experimental Social Psychology, 34, 66-89, Rosenberg, M, J,, & Hovland, C, I. (1960). Cognitive, affective, and behavioral components of attitudes. In C. I. Hovland & M, J. Rosenberg (Eds. ), Attitude organization and change: An analysis of consistency among attitude components (pp. 1-14). New Haven, GT: Yale University Press, Schimmack, U,, & Grites, S, I. , Jr. (2005). The structure of affect. In D. Albarracin, B.
T, Johnson, & M, P, Zanna (Eds,), The handbook of attitudes (pp, 397-435). Mahwah, NJ: Erlbaum. 602 EAGLY AND CHAIKEN Schwarz, N, (2007), Attitude construction: Evaluation in context. Social Cognition, 25(5), 638-656. Schwarz, N. & Bohner, G. (2001). The construction of attitudes. In A, Tesser & N. Schwarz (Eds. ), Blackwell handbook of social psychology: Intraindividual processes (pp. 436-457). Maiden, MA: Blackwell. Strack, F. , & Deutsch, R, (2004). Reflective and impulsive determinants of social behavior. Personality and Social Psychology Review, 8, 220-247. Staw, B. M. & Ross, J. (1985). Stability in the midst of change: A dispositional approach to job attitudes. Journal of Applied Psychology, 70,469^80, Tesser, A, (1993), The importance of heritability in psychological research: The case of attitudes. Psychological Review, 100,129-142. Van Harreveld, F,, & Van der Pligt, J. (2004), Attitudes as stable and transparent constructions. Journal of Experimental Social Psychology, 40, 666-674. Van Harreveld, F. , Van der Pligt, J. , De Vries, N. K,, & Andreas, S, (2004), The structure of attitudes: Attribute importance, accessibility and judgment.
British Journal of Social Psychology, 39, 363-380, Van Overwalle, F,, & Sieber, F, (2005), A connectionist model of attitude formation and change. Personality and Social Psychology Bulletin, 9, 231-274. Webb, E, J. , Campbell, D. T. , Schwartz, R, D,, & Sechrest, L. (1966), Unobtrusive measures: Nonreactive research in the social sciences. Chicago: Rand McNally, Wilson, T. D. , & Hodges, S, D. (1992). Attitudes as temporary constructions. In L. L, Martin & A, Tesser (Eds,), The construction ofsocial judgments (pp, 37-65).
Hillsdale, NJ: Erlbaum, Wilson, T, D. , Lindsey, S,, & Schooler, T, Y, (2000), A model of dual attitudes. Psychological Review, 107,101-126. Zajonc, R. B, (1980), Feeling and thinking: Preferences need no inferences, American Psychologist, 35,151-175, Zajonc, R, B, (1984). On the primacy of affect. American Psychologist, 39,117-123. Zanna, M. P. , & Rempel, J. K, (1988), Attitudes: A new look at an old concept. In D, Bar-Tal & A. W, Kruglanski (Eds. ), The social psychology ofknowledge (pp. 315-334). Cambridge, UK: Cambridge University Press.