This action might not be possible to undo. Are you sure you want to continue?
and empirical evidence. We hQPI::Ibebcok ts of .inrerest both 1.0 graduate students initiating work on atdtudesas wellas 19 Iong-.stanctip_gscholars in the field. Inaddition, because oftbe· rmilly potential directions for application of work on attitude strength to ,am~liqratioD of social problems, t1ie book may be of interest [0 scholars in variousapplled disciplines studyiilg attlhldin·al phenomena.
Attitude Strength: An Overview
A. Krosnick Richard E. Petty CJhio State llniversi~
ACKNOWLEDGMENTS As usmil. mere are manypeopJe to thank fortbeji:contribirtiomHotlli& voiume. First; weare .grateful to ourehapter' authors who C:rreated asuperbconference atmosphere and wrioaecommodated our editorial suggestions. In addition we ar~ indebted ro OUT colleagues and students at Ohio·Stat~ Who provide astimue Jatmg atmosphere for the S~dy of attitudes and social psychology I11.Ore generally, our supportstaff.let] by Shirley Bostwick, and OUT editors at Lawrence Erlbaem Associates, who facilitated productien of this book in many ways. Finally, we are grateful to eurwonderfuf and amazing wives, Lynn and Cathy, who, as always, were bcrhpatientamt supportive as the work on this book was completed. Richard E. Petty Jon A. Krosnick
Reprinted from: Petty, R. E., & Krosnick, J. A. (Eds.) Attitude Strength: Antecedents and Consequences. Mahwah, NJ: Erlbaum Associates.
Throughout history, many of the most sensational events and changes have focused public attention on powerful attitudes. From the French Revolution to recent bombings of abortion clinics by right-to-life activists, the incidents that attract our attention are often those associated with strong sentiments. Furthermore, many of the most significant sea changes in U.S. society have involved the shifting of seemingly unmovable and highly consequential attitudes. Among the most notable of these transformations are the shift from the overt racist attitudes of the 1950s to the seemingly more tolerant stance of contemporary society, and the reshaping of traditionalist opposition to a significant role for women in positions of societal leadership into widespread acceptance of such a role. In these cases and various others like them, powerful attitudes were gradually transfigured as the result of intense social pressure and heated public debate. Such concerted efforts at inducing collective attitude change in these instances were inspired partly by the belief that people's attitudes (e.g., prejudice) were responsible for destructive behaviors (e.g, racial discrimination). Therefore, changing the attitudes would change behavior. Consistent with these informal observations, a number of studies conducted since the 1950s have made it clear that attitudes can be very stable, consequential, and very difficult to change. As Hovland (1959), Hyman and Sheatsley (1947), and others pointed out, most attitudes appear to change only rarely in the course of nonnal daily life, even when elaborate influence campaigns are mounted to induce such shifts. And some attitudes, such as those toward political candidates, are very powerful determinants of relevant behaviors (e.g., voting in elections; see Schuman & Johnson, 1976). 1
attitudes can influence information processing and judgments. it is a heuristic label we attach to certain attitudes as a way of efficiently noting that they possess certain characteristics. Treating attitude strength in this manner allows us to incorporate the most common meaning of the construct and to be consistent with the past work reviewed in this book. Bollen & Lennox. Of particular relevance are the notions that strength refers to "the power to resist attack . 1981). others are quite flexible and have few if any important effects. From this perspective. so too can it be that some attitudes are harder to change and have more powerful effects on peoples' lives than do others. 1801). we would need to specify its relation to these manifestations. the two defining features of strength could be said to combine with one another additively or multiplicatively to yield an overall level of an attitude's strength. because their apparent preferences are so flexible. when Raden reviewed the literature on attitude strength in 1985. or that certain decisions will be rendered. In this regard.g. the stronger it is. Strong attitudes presumably show persistence and/or resistance. just as physically strong people are hard to budge from where they stand and have powerful effects on others and the world around them. MacCallum & Browne. Two manifestations of durability have received the most conceptual and empirical attention in past research. strength itself is not presumed to be a latent psychological construct somehow represented in memory. ATTITUDE STRENGTH: AN OVERVIEW 3 At the same time. beginning with LaPiere's (1934) investigation of hotel and restaurant acceptance of Chinese patrons..2 KROSNICK AND PETTY 1. . " (p. 312). In addition. We begin by offering a working definition of attitude strength and outlining a series of attributes of attitudes related to strength that have been the focus of extensive empirical study and that are the focus of the chapters to follow. we could define attitude strength as a latent psychological construct that is presumably represented in memory by various attributes of the attitude. during the last few decades a great deal of research has demonstrated that whereas some attitudes are indeed stable and consequential. 1969). 1976) provides several meanings for strength. & Smith. Taking this approach. For example. If we were to take this approach. Strong attitudes are more likely to impart a bias to information processing activity and judgments than are weak ones..g. which refers to an attitude's ability to withstand an attack (Petty & Cacioppo. This refers to the degree to which an attitude remains unchanged over an extended period in the course of normal daily life. DEFINING A TIITUDE STRENGTH Although attitude strength has often been discussed in the social science literature over the years (see. see Petty. 1981). The first aspect of durability is the persistence of the attitude (or stability. e. In order to formally define attitude strength. who asserted that many people really have "nonattitudes" on major issues of the day. Thus. numerous studies have shown that attitudes are sometimes only very weakly associated with behavior (for a review.. we are inclined to treat attitude strength as the extent to which attitudes manifest the qualities of durability and impactfulness. durability . even if it were never challenged. and they have impact. One especially cogent and influential statement of this view was offered by Philip Converse (1964). 1993. 1991. he noted that "attitude strength has generally not been defined with any precision and it does not appear to have any agreed-upon meaning for attitude researchers" (p. and strong attitudes should be more likely to do so than weak ones. as it is often called). Likewise. 5.. Thus. the more durable and the more impactful).. e. In this case. Although these bodies of literature on attitudes may appear to be in conflict with each other. we would presume that an attitude's durability and impactfulness covary to at least some extent. Furthermore. 1985. force . which would be viewed as a phantom variable (Bollen & Lennox. see Wicker. Schuman & Presser. Eagly & Chaiken. e. in the sense that they make it more likely that certain information will come to mind. We also review evidence linking these attributes to the defining features of strength.. The various chapters in this volume examine attitude attributes related to attitude strength and the processes by which attitudes attain these attributes. social psychologists have accumulated evidence that suggests that attitudes sometimes can be anything but stable and consequential. and strength would be said to exist only when both of these attributes are present in an attitude. 1991). many of the attitude change studies conducted in the laboratory during the last four decades can be viewed as documenting how easy it is to change people's opinions (see. it has been more of a vague metaphor than a formally defined social scientific construct. The primary goal of this book is to understand the intrapsychic processes responsible for this variation in the strength of attitudes. The more of each feature an attitude possesses (i. Our goal in this discussion is to place the following chapters in a historical and conceptual context. Therefore. First. we could assert that an attitude is strong to the extent that it manifests either durability or impactfulness or both.g. First. because these observable qualities would presumably be results of an attitude's strength (see. That is... A second aspect of durability is resistance. ch. 1986. it would seem that strong attitudes are ones that possess these two features: They are durable. we could treat durability and impactfulness as causal indicators of attitude strength. durability and impactfulness would be viewed as effect indicators of an attitude's strength... Rather. For example.e. Petty & Cacioppo. this volume). attitudes can guide behavior. we have at least two choices. But what does it mean for an attitude to have strength? Webster's unabridged dictionary (McKechnie. 1993). Haugtvedt. the power to produce a reaction or effect . as a working definition. Alternatively.. two manifestations of attitudinal impact have been the focus of extensive research.. however. Yet the notion that some attitudes are stronger than others has powerful intuitive appeal. Raden. and we consider the relations among these attributes.
this volume) argued that attitudes influence our behavior in part by shaping our perceptions of the world around us. latitudes of acceptance and rejection.. there are a number of reasons to expect that attitudes possessing any one of the four aspects of strength that we have identified will often possess all Numerous studies have investigated the attributes of attitudes (e. the more knowledgeable a person becomes about an attitude object. the more committed he or she will be to the attitude (see Kiesler. liking of loud parties) may be tied to inherited individual differences (e. (b) aspects of the cognitive structure associated with the attitude and attitude object in memory. Yet remarkably little research has been done to date exploring the degree of empirical overlap among these features of attitudes or the degree to which they can reasonably be thought of as reflecting a single underlying construct that might be called strength. flexibility. persistence. These specific properties were of interest mostly because they were assumed to relate to an attitude's durability and/or impactfulness.. Raden (1985) expanded Scott's list by examining accessibility. an attitude assessed at one time is unlikely to predict behavior at a later time if the attitude does not persist over the time interval. because there is so much support for the existing viewpoint. & Abraham. Also.. attitudes such as liking to work may have biological origins that are common to all humans (Arvey. consistent with common dictionary definitions of strength. Other attitudes (e. and (d) cognitive processes by which an attitude is formed. That is. The more committed an individual is to an attitude. the ability of an attitude to predict behavior is dependent in part on the attitude's ability to bias perceptions of the attitude object and the behavioral context In addition to the relatively direct influence of some features on others. contributing to persistence. The chapters in this book focus on some of the strength-related dimensions of attitudes that have been addressed in empirical research extensively since the 1950s.g. an attitude's stability over time is likely to be a joint function of its resistance to overt challenges as well as the degree to which its representation in memory fades naturally over time. some of these properties have been the subject of extensive empirical research. affective salience. That is.g. Since Scott's chapter appeared. positive or neg a'Some researchers have recently begun to explore the genetic determinants of attitude strength as well (e. ambivalence. I Aspects of Attitudes Attitudes are presumed to vary along an evaluative continuum ranging from a strongly positive orientation to a neutral orientation to a strongly negative orientation. Furthermore. and guiding behavior) as the defining features of strength. Eysenck. some of the features appear to influence others in a rather direct manner. 1971). salience. evaluative-cognitive consistency. Thus. a variety of strength-related attributes have been proposed and investigated. for example. resistance.g. an attitude's strength is the degree to wbich if possesses these features. embeddedness.g. RELATIONS AMONG THE DEFINING OF AlTITUDE STRENGTH FEATURES of them. The primary goal of this book is to understand how and why attitudes come to have these strength properties. resistance to change). for a number of reasons. they do seem likely to co-occur. ch. the more a person performs behaviors toward an object that are consistent with his or her attitude toward it. and thus might be more easily modifiable. 10..g. This continuum can be decomposed into valence (i. STRENGTH-RELATED DIMENSIONS OF ATTITUDES Although the various defining features of strong attitudes can be separated conceptually and empirically. a number of other attributes related to strength have been proposed and explored.. see Tesser. some attitudes may be durable and impactful because they have an inherited component. importance.e. knowledge) that are correlates of each of the four strength features (e. impact on information processing and judgments. 1967). For example. ATIITUDE S1RENGTH: AN OVERVIEW 5 We refer to these four aspects of attitudes (i. attitudes that are expressed frequently in behavior are likely to be reinforced in memory. in his chapter on attitude measurement in the Handbook of Social Psychology. the harder it will likely be to change his or her attitude toward the object. That is. Campbell.4 KROSNICK AND PETTY 1. factors that might enhance an attitude's impact on information processing also seem likely thereby to enhance its resistance to change. As we describe in the following. In his literature review in the mid-1980s.e. Bouchard. we define attitude-strength in terms of postulated marker characteristics or consequences that strong attitudes are thought to possess. For example. First. the ability of an attitude to predict a subsequent behavior is dependent on the attitude's stability. Yet another possible source of commonality among the four features of strong attitudes is shared origins. . 1989). & Brock.. whereas other properties have been largely ignored in the empirical literature. cognitive complexity. For instance. as well as sources of relations among them. Similarly. as Schwarz (1978) noted. Thus. Scott (1968) described 10 such properties: magnitude (extremity). these strength-related attributes can be viewed as falling into four categories: (a) aspects of the attitude itself. Since then. 1957). the more likely he or she is to resist attempts to change it (Hovland. having more knowledge about an object is also likely to enhance the impact of the attitude on information processing. So. Thai is. The chapters in this book focus on aspects of strength that are presumably learned. In addition. there are also likely to be indirect influences of one feature on another. Segal. (c) subjective beliefs about the attitude and attitude object. and consciousness. overtness. We therefore adopt the definition of strength that makes no presumptions about these relations. direct behavioral experience. certainty. Since the late 1960s. intensity. introversion! extraversion. 1993). because the person will have a greater ability to interpret events as consistent with the attitude. For example. Fazio (1986.. and vested interest. For example.
' 2Because links to the self can sometimes be structural (e.. However.. ~Some other strengjh-relaiedanlmde attributes nave been discussed at 'length in the lirerature. little knowledge (e. ch.. 1981) and latitudes of aceeptance. people vary ill the extent of their confidence that their attitudes toward an:y given object accurately represent theiroverall orientations toward it. negative attitudes might be more durable or impactful than positive ones. people perceive some attitude objects to be closely connected tQ their important personal goals. Bergida. Johnson & EagJy . 6 this volume).hey are discussed )1) various chapters. One particular basis of personal relevance or involvement is vested-interest (see. attitudes vary in the degree to which there is consistency between evaluations of the object and the information associated with it. & Berent. 5. The most notable of these is elaboration (e. conceiving of attitudes as singleevalnatiens of objects. Furthermore. & Dittus.1:1. '8. Krosnick. and the degree to which it is evaluated riegatively (e. in the form of structural linkage. 1988a). this volume. 1972). Tesser. 1986). most research attention has focused on extremity. Attitudes are alscthougnt be linked 10 knowledge about the attitude 96ject. in both of these senses (see Gross. Wood. this volume).. this volume). Judd & Brauer. Attitude researchers generally believe that people are motivated to hold "correct" attitudes (see.. & Miller.ence~. the extent to ·which jhe animdecbject is perceived be instrumental to ones tangible outcomes (see also. Subjective Beliefs About Attitudes Aspects of Attitude Structure Many theorists assume chat attitudes exist in memory wifuin a network of associative links connecting them to other cognitive elements . Rhodes. 14. to People hold a number of beliefs about the attributes of their own attitudes and about the attitude object. Attitude importance is thus the degree of psychological significance people attach to anaterude (see Beninger. ~d nonccmmirmem (Sherif.. this volume. Holtz. Berent. Wilson. ch.pf perceived self-relevance rlllh¢! tha:n actual ~clf-rel~vaui. & Giner-Sorolla. Wood. Yet people are more confident in thecorrecmess of some attitudes than others. 198~). Krosnick. 1954. are to of Processes Rather than focusing on a particular characteristic of the attitude or its structure. 7.g.rila). . & Haugtvedt. Haugtvedt. & Griffin. . Petty & Cacioppo. That is. Fabrigar. <::. Petty. this volume.g. Bo_1h valence and extremity might be related to strength in that. ch.(e. Elaboration refers to the degree of thinking one does and has done about an attitude object's attributes. &. whereas others are associated with.g. Ambivalence refers to the degree of cOi:!:fliotbetween these components (see Thompson. 1981. and the content of that information (see Davidson. Breckler. be a manifestation of the degree of personalrelevance tile attitude object (Berringer. This information can include memories of emotions and past behaviors that are evoked by the object as well as more specific attributes of the object. ATTITUDE STRENGTH: AN OVERVIEW 7 and Attitude Objects tive) and extremity (degree of favorabflity). the more an individual likes or dislikes the object (see Abelson. 'such IIsarre¢. AlIIl._Fesiinger. it is possible to view an attitude as the summary of two distinct components: the degree LO which one evaluates an object positively.sify personal relevance as a structural feature Qf attitudes.e.g. Kaplan. & Petty. Radecki. 2. AttitlJdes low in ambivalence involve either mos~lJ pOSitive evaluation or mostly negativeevaluariou.!e certainty refers to the degree to which an individual is confident In hisor her attitude toward an object. Extremity is the extent to which the attitude deviates from neutrality. this volume).ua6veaffective consistency. 1989). Martin.this volume). this volume). Petty & Cacioppo. eh. Some attitudes are accompanied by a great deal of attitude-relevant knowledge. 1984' Crites.1 In a similar vein. ch.oughthis is reesonable. ch. Petlf & Cacioppe. 'and consequently they care deeply and areespecially concerned about it (Krosnick. haldfug exlrcnii!y constant. Crane. desires and wishes (e. ch. Attitude accessibility is defined as the strength of the object-evaluation link and is most directly manifest as the ease with which an attitude comes to mind in the course of social perception (see Faz-io. two . 19:.e. this volume). this volume). 15. 19~).seeChaiken. !. and consistency of the attitude with emotions associated with the ebjecL(eval.6 KROSNICK AND PETIY 1. 12. Shertl\ & Nebel"gall. f_or example. Most research attention bas focussed on two kinds of conststency: 'consistency of the attitude with beliefs about the object's attributes (evaluative-cognitive c0Rsistency).. a person's subjective perception of the amount of information he or she has. The more extreme an attitude is. 9. 1986). Fazio (l986) proposed that an attitude can be thought of as a link between the representation of an attitude obje~t and its evaluation in memory. Pratkanis & Greenwald.t!J. ch. Rather than. 1994. Furthermore. whereas highly ?IDbiv. Cacioppo. knowledge has been examined in terms of the size of the body of information one has about an object. 4. Attilu. 1989). I 0. an attitude that is connected in .c. this volume). 1990.56. some theorists have focused on the cognitive processes by which an attitude is formed. ch. Pomerantz. Lavine.g. Attitude importance is thought to. 13. Jaccard. ch.memery to a self·~.sorne attitude objects high iuper-sandi relevance and produce a sense of personal involvement with the-issue (see Thomsen. & Mendolia. this volume. 'One might· olas. its merits and drawbacks (see Petty.g.~ent attitudes involve Doth positive and negariveevaluations.qu. Rosenberg.researchers have emphasized lheeonse.. & Smith. some people consider an attitude to be very important to them personally. ch. Oe~avim:atexperience (Fazip' & Zanna. ~995). 11.g. i"qiltlion. Zanna. ch. & Fabrigar. In the' attitude strength literature. '1992). ch. For example. AI!-hougn thesedlrnensions are not central fncl of lhls bOok. this volume). 3. & Biek.
people were instructed to think about the reasons for their attitudes. instead of observing people's behavior. based on the assumption that attitudes play a role in determining these judgments. at the very least.. Chen. In addition. resistance has been gauged most often by exposing people to persuasive messages in laboratory settings and assessing how much attitudes change as a result (e. such that people are more likely to remember attitude-consistent information than attitude-challenging information (e. allowing researchers to estimate attitude stability while correcting for the distorting impact of random and systematic measurement error (e. & Wilson.. Furthermore.g. In other studies. Chaiken. We presumably like others who share our attitudes more than those who do not (e. Festinger. Wood. Krosnick. Raden. 1982. 1979). 1979). Conventional paradigms assessing these effects. In still other studies. multiple indicators have been collected at each time point.g. McDonel. As distinct as these constructs appear. 1988b). For example. which might ultimately yield extremity. strong attitudes were expected to be more consistent with behavior than weak ones. 1983. they may share a small set of common causes. Haugtvedt & Petty. & Carnot. 1989.g. Studies of attitude persistence have typically assessed an attitude at one time point.. Haugtvedt & Petty. Byrne. In still other studies. Hodges. Studies to date have not yet assessed the relations of all strength-related dimensions to all four defining features in these ways. either in experimental laboratory settings or via survey questionnaires. 1992). expanded knowledge. subjects' behaviors performed in the course of everyday life were measured directly and compared with attitude reports (Ajzen & Madden. 1957. attitudes are presumed to influence attraction to others. 1980. 1942). Such engagement could be sparked initially. & Lepper.. for example. attitudes can bias the evaluation of scientific data: Evidence supporting our attitudes is seen as more compelling than evidence that disagrees with our attitudes (Lord. have all been used to examine whether strong attitudes have more powerful effects than weak ones. Krosnick.g.. Also. Finally. those relations that have been assessed are remarkably consistent: Each dimension has been shown to be positively associated with one or more of the four defining features (for reviews. But other studies have employed different change-inducing methodologies. In some. For example. RELATIONS AMONG THE DIMENSIONS Why might there be such strong similarity among these dimensions in terms of their correlations with the defining features of attitude strength? One possibility is that all these dimensions reflect a single underlying construct. Ross. and we presume that they do not share our attitudes if we dislike them (e.g. A wide diversity of approaches have been taken to assessing the impact of attitudes on information processing. Boninger. & Dunn. Roberts. 1986). 1986). One might therefore think of the confluence of these dimensions as constituting emotional and intellectual engagement in an attitude. see Erber. In each case. researchers have relied upon subjects' reports of their past behavior or upon reports of behavioral intentions regarding the future (e. 1982). Furthermore.. As gauged in any of these ways.. attitudes are presumed to shape memory for relevant information. Chuang. reassessed it at a later time point.g.8 RELATIONS OF THE DIMENSIONS KROSNICK AND PEITY 1. For example. Fazio. certainty. ATIITUDE STRENGTH: AN OVERVIEW 9 TO STRENGTH These attitude dimensions have typically been examined in separate investigations that related each one individually to durability and/or impactfulness (e.. Berent. the stability of strong attitudes is expected to exceed that of weak attitudes. Schuman & Presser. the valence of attitudes (ranging from favorable to unfavorable) has been used to predict valenced judgments or biases of some sort. 17. they all share one feature in common: They appear to be related to the four defining features of strong attitudes. only in very few instances has one of the dimensions been found to be unrelated to one of these features. some studies have measured people's attitudes toward an object and then given them an opportunity to perform a behavior that is either favorable or unfavorable toward the object. or attitudes have been measured at three or more time points.. That is. 1989.. which might then inspire extensive thinking and information-gathering. 1984). Boninger et . Weigel & Newman. 1981). Heider.. Feldman. and estimated the correlation between the two (e.g. ch. 1992. This might instigate a sense of importance at first. 1971). 1985). In other studies. Swann & Ely. people were asked leading questions about their attitudes that led some to change (e. and in no case has a dimension been found to be negatively related to one of the features.g. stability has been assessed by seeing how much a group's mean attitude is maintained over time (e.. In each case. 1992. All such manipulations are expected to have more impact on weak attitudes than on strong ones. 1976). however.g. Studies to date have used a variety of approaches to assess these features. a procedure that also induces change in some of them (Wilson.g. and so forth (e. Ewing. by recognizing the personal relevance of the issue (Petty & Cacioppo. Miller & Grush. Thus. Thus it is plausible that they may have unique origins and unique effects. Kraft. see Krosnick & Abelson. Attitudes are also presumed to influence our perceptions of other people's attitudes-we presume that others share our attitudes if we like them. However.g. attitude-behavior consistency has been gauged in a variety of different ways.g. all of these dimensions seem related to strength. Davidson & Morrison. 1985). such as signing a petition advocating a certain view (e. 1958). this volume). the key attributes of attitudes related to strength have been defined and operationalized in ways that may make them appear clearly distinct from one another. although the various dimensions are clearly conceptually and operationally distinct from one another. 1993. & Shennan. In some cases.
it is conceivable that attitude accessibility is the sole determinant of attitude-behavior consistency. Katz (1960). certainty and direct experience) may enhance an attitude's impact on behavior and information processing by strengthening the object-evaluation link in memory (i. and other attitude attributes.. 1988). On the other hand. this volume. and impact on cognition and behavior through unique processes. Another higher order construct about which there has been a great diversity in approaches is attitude centrality. and White (1956. who suggested that at least some attitude attributes (e.. Lewin (1951). Tannenbaum.g. Katz. Petty & Cacioppo. 1975. 7.g. it would be most parsimonious to think of them all as manifestations of a single latent construct. 1956). and empirical testing in this arena. there may be at least some overlap among some of the dimensions that reflect a smaller set of higher order constructs. resistance. One of the most studied higher order constructs. ATTITUDE STRENGTH: AN OVERVIEW 11 aI. Therefore if multivariate analyses were to be conducted predicting attitude-behavior consistency using accessibility. 1975. if any.. ann 'some groups of dimensions might reflect single underlying constructs. ultimately producing high levels in them all. 1980) and frequency of thought (Brown. thinking about an attitude object increases its accessibility (Rennier. Questions about importance have frequently been used to gauge this construct (Converse. Or. The great similarity between the dimensions in terms of their relations to the four defining features of strong attitudes justifies empirical examination of the associations among the dimensions. 1965). certainty. involvement. virtually all associations between these dimensions and attitude strength features would be redundant.. Yet direct measures of such structural linkage have rarely if ever been used in the attitude literature to operationalize this construct. leading to such high correlations among the dimensions that they might be viewed as reflecting a single underlying construct.e. Ultimately. 1977. they reflect. and others defined "centrality" as the extent of structurallinkage among attitudes. This sort of theoretical viewpoint has been offered most clearly by Fazio (1989). 1983. 1990). 1959. or the confluence of importance. Schuman & Presser. any departures from perfect association among the dimensions for any given attitude might then simply reflect the fact that these sequential processes have not yet completely unfolded. there are anumber of instances in the attitudes literature where higher order constructs have been measured with a wide variety of different specific measures. thinking about an attitude can increase its extremity (Tesser. such perfect overlapseems unlikely. Future research could then identify and measure or manipulate the most effective indicators of this underlying dimension. Attitudinal saLience has been 'measured by questions about importance (Hoelter. Judd & Krosnick. over a period of time. Borgida. More generally. 1974). Jackson & Marcus. thought. p. reverberate throughout the system. Support for the notion of reverberation comes from evidence that some of the attitude dimensions cause others. Even if the dimensions all represent independent but moderately overlapping constructs. Thus it is conceivable that once one such dimension is elevated (e.g. 1947. 1981). According to this view. and self-interest (Flora & Maibach. ch. Bern (1970). ch. various cognitive and behavioral processes are set in motion to elevate some or even most of the other dimensions (e. Powell. 1968. 1955. 1982. and social support (Miller. knowledge (Stember & Hyman. 1968. and future theory building could focus on that one dimension instead of on a multitude of higher order properties. this volume). Bruner. 182). 1978). 1975. if the . questions would arise about the extent to which relations between individual dimensions and defining features of attitude strength are spurious. all of the attitude dimensions might be uniformly shaped by a set of common causes. Borgida & Howard-Pitney. accessibility. because the array of dimensions currently being addressed by separate research programs could be reduced to a single dimension that could then be studied in a much more streamlined fashion. For example. Tedin. enhancing an attitude's persistence. For example. it is possible that each of the dimensions is uniquely related to the strength features. Also. 35) and Freedman (1964). For example. and that importance and evaluative-cognitive consistency are only associated with attitude-behavior consistency by virtue of their correlations with accessibility. However. And if this is true. 1950) or extremity (McDill. What might appear to be distinct qualities of attitudes could then be thought of as only one single property. Converse (1970) defined centraLity as the "proportion of 'mental time' that is occupied by attention to the attitude object over substantial periods" (p. Petersen & Dutton. Thus there is a fair amount of disagreement in the literature about the extent to which these dimensions can be used to measure each other and about which higher order constructs. has been assessed by measuring importance (Apsler & Sears. 1964. 1944. Consequently. Gom. & Omoto. information gathering. the attitude's accessibility).. 1979). Nevertheless. investigators have presumed that attitudinal intensity can be measured via certainty (Brim. Lemon. extremity). 1986). McDill. an individual comes to perceive the personal relevance of the attitude object and begins to think about it. 1986. with one set of causes and one set of effects. 1959. Given the clear distinctions between the dimensions in terms of definitions and operationalizations. the dimensions may be highly overlapping with one another and may therefore have essentially identical effects on attitude strength features. which seems closer to the notions of attitude salience and accessibility. On the one hand. the confluence of thought. If the dimensions are all perfectly overlapping with one another. an approach that comes closest to the definitions of it offered by Smith.10 KROSNICK AND PETrY 1. amount of thought (Bishop. some overlap seems quite plausible. 5. many of these correlates of consistency might turn out not to be causes of it. commitment. it would greatly simplify theory building. 1949-1950). Krosnick. Petty et al. Alternatively. 1985. thus resembling Scott's (1968) notion of embeddedness. Suchman. however. 1990). direct experience. At the same time. inducing a high level in anyone of these dimensions might. Howard-Pitney. Guttman & Suchman.. Indeed.
1985.. Therefore. On the basis of this sort of evidence. The literature Raden reviewed provided only a small subset of all possible correlations among these dimensions. Krosnick et al. but most of the associations he observed were relatively weak.1 displays the results involving the dimensions addressed in this book from a typical example. because they are likely to have been distorted by random and systematic measurement error. as well --0 . Krosnick et al. one for each of four attitude objects. Thus. Raden did not collect original empirical data. it is difficult to know exactly what to make of the zero order correlations that were the focus of Raden's study. Boruch & Wolins. (1993) collected multiple measures of each of ten strength-related dimensions..12 KROSNICK AND PETIY dimensions are completely nonoverlapping. Given Cote and Buckley's (1987) evidence that random and systematic measurement error typically account for more than 50% of the variance in psychological measures. 1988. In order to estimate these correlations more precisely. Raden (1985) concluded that these dimensions are not all reflections of a single underlying. Raden asserted that the one construct view of attitude strength should be abandoned in favor of a multiconstruct view in which the dimensions are essentially independent of one another. but rather reviewed some of the existing published evidence on associations between some strength dimensions. 0\00 0 I· II') 13 . it is useful to assess the overlap among the dimensions. •. Alwin & Krosnick. Table 1. and correlated measurement error due to shared method can make correlations either more positive or more negative (see. Green. Krosnick & Alwin.. However.S. 1988). These dimensions include measures presumably tapping accessibility (frequency of talking about the object. government (all variables were coded so that positive correlations would be expected).. e.g. Associations Among the Dimensions In fact. 1970. in order to understand which attitude attributes are directly responsible for attitude strength. In all.. it seems quite plausible that zero order correlations among attitude dimensions are misleadingly attenuated and that there is more overlap among them than Raden (1985) believed there to be. regarding attitudes toward defense spending by the U. Each dimension might have reasonably unique origins and effects. many correlational studies have assessed the relations among a subset of the various strength-related dimensions and have consistently documented only low to moderately positive correlations. Random measurement error attenuates correlations between indicators. . superordinate dimension. zero order correlations may either overestimate or underestimate correlations between attitude dimensions. These investigators then applied structural equation modeling techniques to the resulting data in order to correct for the impact of random and systematic measurement error. such multivariate analyses would likely leave all bivariate relations unaltered. (1993) estimated four correlation matrices. Consequently.
certainty. and many are near zero. Similarly. 1984. the amount of knowledge people had about the object (as assessed via self-perceptions as well as listings of everything people knew about the object). some cases of moderate overlap. (1993). the correlation between self-perceptions of knowledgeability and the amount of knowledge subjects listed was .14 KROSNICK AND PETrY as response latency). and extremity.2. but they were not. It is interesting that different means of assessing the same construct appear to have yielded distinct results. elaboration (frequency of thinking about the object). 1991. Krosnick et al.'s (1993) findings. others are moderate. Although some of the correlations shown here are relatively strong. one might expect them to be more confident in their attitudes. 1993) to correct for the impact of random measurement error. for some of these later variables. as well as additional ones were examined. importance. Krosnick et al. and cognitive-affective consistency. Thus self-perceptions of a construct might even have effects different than those r-r00 I• • . extremity. and Petty (1994) reanalyzed data origirially collected by Strathman (1991). (1994) found clear distinctions between self-perception and objective measures of the same construct.. Thus it seems that the general complexion of Krosnick et al. Jarvis. surprisingly strong negative correlations appeared (again. Some of the stronger correlations involve importance. In fact. and consistently weaker correlations involve evaluative-cognitive consistency. some measures that would be expected to be correlated were not. MacCallum & Browne. Consistent with Krosnick et al. First. (1993) estimated. if people possess considerable knowledge about an attitude object. 15 . Also consistent with Krosnick et al. Bollen & Lennox. The corrected correlations they obtained for one of their attitude objects (George Bush) are shown in Table 1. Strathman. it seems worthwhile to speculate about the meaning of such a distinction. and evaluative-cognitive consistency. . self-perceptions might reflect constructs different from those tapped by the direct measures but equally real and useful. most of them are moderate to weak. amount of thought. some of these correlations are relatively strong.cl~~3 ' .'s (1993) results were replicated. the variables were coded so that positive associations would be expected). self-perceptions of knowledgeability were only weakly associated with the amount of knowledge people were actually able to list about an object. Krosnick. evaluative-affective consistency.. These results are comparable to those in the other three corrected correlation matrices Krosnick et al. Because of the nature of some of the measures. Given the apparent distinction between subjective and objective measures of what might be considered the same construct. For example. This is particularly so for the amount of knowledge people could list about the object and for evaluative-cognitive consistency. certainty. highlighting some cases of strong overlap. (1994) estimated phantom variable structural equation models (Bollen. For example. and some cases of no overlap at all.'s (1993) results. Multiple measures of some of the dimensions examined by Krosnick et al.14. For example.
It is thus conceivable that both the self-perceptions and the objective measures of attitude dimensions may both contain useful. thinking. Do the Dimensions Reflect a Few Higher Order Constructs? It seems clear from the evidence we reviewed (i. it is possible for all of the dimensions that are at least somewhat correlated with one another to reflect only a single latent construct. desire to express one's opinion on the issue. but imperfectly.. The first. 1986) do seem to exert such regulatory influences. 1981) and extremity (Hippler & Schwarz. although attitude importance. Thus the available evidence seems to support the view that many of the various attributes of attitudes that may contribute to strength are best thought of as distinct from one another. resistance. these investigators conducted confirmatory factor analyses. and some were suggested by theoretical considerations. ATTITUDE STRENGTH: AN OVERVIEW 17 of direct measures of it. called "ego preoccupation. and interest." included items measuring knowledge. 1988). this imperfection cannot be attributable to it.2. the imperfection could be due to the presence of method-specific systematic error. Krosnick & Schuman. Similarly. They also factor analyzed a set of relevant indicators and found yet a different three factor structure. evaluative-cognitive consistency (Chaiken & Baldwin. Krosnick et al. that medical researchers have found that self-perceptions of health status contain some valid information that predicts later mortality. and independent information. Also consistent with the notion that many of these dimensions are distinct from one another is evidence that they can have nonoverlapping. Conceivably. which he called "emotional commitment. Furthermore. whereas the objective measures are less susceptible. . 1993. every one of the models Krosnick et al. intensity. talking. (1993) evaluated did not fit the observed data and was therefore empirically rejected. the self-perceptions and objective measures could tap the same construct. though. It is interesting to note. In fact. The second factor included measures of certainty and ego defensiveness. unexplained covariation may simply represent random or systematic measurement error. this evidence suggests that it is . "cognitive elaboration. Other investigations have uncovered clear evidence of more complex organizations.16 KROSNICK AND PETTY 1. And the third factor included importance and the relevance of one's social identity and values. Thus Lastovicka and Gardner's (1979) findings were quite different from Abelson's (1988). Krosnick et al.86. Because random measurement error has been purged from the corrected correlations in Tables 1. Second. and feeling involved in an issue. Taken together. 1994) that there is a considerable amount of independence among the individual strength-related dimensions.86. This suggests that a single factor did indeed account for a great deal of the covariation among the dimensions. and . That is. Bradburn. and mediated effects. and other such dimensions. Their first factor included items addressing frequency of talking. A similar investigation was conducted by Lastovicka and Gardner (1979). Furthermore. knowledge. And the third cluster. Rasinski. For example. 1991) reported one pair of studies consistent with the notion that a group of strength-related dimensions reflects a higher order construct. Krosnick et al. Biek (1992) showed that attitude-defensive biased processing of a persuasive message is most likely among respondents high in know ledge and intensity. Tourangeau. valid. because it is presumably subject to intentional manipulation by respondents. Some of the models they tested were suggested by previous investigators' presumptions regarding interchangeability. He examined the structure of measures of interest." included items assessing certainty and relevance to self-concept. Abelson's (1988) analysis of a series of attitude dimensions revealed three clusters of items across a set of issues. 1990. Because this approach involved explicit model testing and goodness-of-fit assessment. and those underlying constructs might produce associations with stability. this evidence suggests that the one factor view of strengthrelated attributes may not be viable. Verplanken treated single indicators of each of these dimensions as if they were all indicators of a single underlying construct. it is nonetheless possible that two or more of them reflect common. or it could represent the existence of a more complex underlying factor structure. and impact on cognition and behavior. interactive. and D'Andrade (1989a.1 and 1. The remaining. For example. It is tempting to view the self-perception data as the more flawed. Krosnick et al. which raises questions about the reliability of these factor structures. and certainty generally do not regulate the magnitude of question wording effects (Bishop." included items measuring frequency of thought and importance. However.. as well as others not directly linked to the dimensions of interest here. over and above the information contained in objective measures of health status (Idler & Angel.'s (1993) findings can explain the inconsistency of previous exploratory factor analyses: That method was apparently being used to uncover latent structure that did not exist. For example.e. independent. Instead of conducting exploratory factor analyses. Verplanken (1989. (1993) offered a resolution of this confusion by taking a somewhat different approach to assessing the latent structure of these dimensions. however. the set of dimensions may be reflections of a smaller set of underlying latent constructs. higher order constructs. Specifically. frequency of reading about the issue. Despite this. 1989b) showed that question order effects are especially likely to occur among people who are high in ambivalence and consider an attitude to be highly personally important. The second cluster. these dimensions sometimes interact with one another. it seems quite informative. Indeed. (1993) specified and tested the adequacy of a series of models that presumed two or more dimensions reflected a single underlying construct while controlling for the impact of random and systematic measurement error. Cronbach's alphas for indices derived from sets of them for three issues were . even if these dimensions are in fact distinct from one another.75. 1990).
Journal of Experimental Social Psychology. Consequently. Zanna. Previously. Attitudes. we suspect. and Mendolia explore the impact of mere thought on extremity and attitude-behavior consistency. Haugtvedt. Borgida. we sought to bring together researchers exploring attitude strength to share their findings and perspectives. These techniques can be easily adapted for use in future investigations. encounters. Public Opinion Quarterly. In future years. J. Krosnick. Berent. Tesser. and Smith. some investigators will continue to concentrate on single dimensions. F. and Biek.18 KROSNICK AND PETIY 1. J. L. The measurement of values in surveys: A comparison of ratings and rankings. 267-275. . Radecki. Rhodes. There was not a great deal of cross-citation taking place. The chapters in this book attempt to provide a state of the art discussion of many of the most researched dimensions of attitudes related to strength. THIS BOOK Although individual researchers have not concentrated exclusively upon single dimensions of attitudes related to strength. Boninger. (1986). All of these chapters describe the existing bodies of research documenting the relations of these dimensions to the four defining features of strength. A. and the conditions under which they are most likely to appear. probing their idiosyncratic features and building models of their causes and effects. We hope that this volume helps to facilitate both sorts of work. Yet the fact that these dimensions all appear to be related to the four defining features of strength suggested a common core to these individual research agendas. (1985). Hodges. Our hope was that some direct exchange of ideas among ourselves and integration of our perspectives might be productive. Dittus. His emphasis is on the social and interpersonal sources of extremity. Alwin. and Wilson. Ajzen. Prediction of goal-directed behavior. Judd and Brauer explore how mere thought. (1988). T. In the next three chapters. First is Fazio on accessibility. D. attempting to understand the relations among the strength-related attributes and to identify their unique contributions to attitude strength. Pomerantz. & Madden. CONCLUSION Focus on one strength-related dimension at a time has been the predominant approach employed in attitude strength research in the past. and Lavine on personal involvement. REFERENCES Abelson. 453-474. and Griffin then report on ambivalence. and expressions can influence attitude extremity. In organizing a 1991 conference on attitude strength that set this book in motion. and Wilson report multivariate analyses exploring the relations of numerous strengthrelated attitude attributes to attitude stability and resistance to change. there has been a strong tendency for scholars to focus on just one dimension at a time. valuable new insights can be produced by multivariate studies considering many dimensions simultaneously. In it. 43. and Eagly and Chaiken review implications of interattitudinal and intraattitudinal structure. 49. The book's final chapter is intended to provide practical help for investigators interested in conducting studies of strength-related attitude dimensions. Thomsen. Next are three chapters on knowledge: Wood. The next few chapters address both the origins and consequences of dimensions defined as subjective perceptions of attitude qualities: Crano reviews work on vested interest. the next chapters address the causes and effects of structure-related dimensions. overarching issues being addressed in a coordinated fashion from multiple perspectives. In particular. the strength consequences of these dimensions. the findings of individual researchers studying individual attributes had been published in relative isolation from one other. 22. Krosnick. we asked chapter authors to identify the antecedents of the attitude strength dimension(s) examined in their research programs. resistance. Martin. the focus shifts to thought processes and their impact on strength. and there were not central. American Psychologist. and Fabrigar on importance. examines not mere thought but rather the effects of elaboration of persuasive messages on attitude persistence. Davidson. the one strength-related dimension that is an attribute of the attitude itself. Downing. Holtz. A second important goal of this book is to catalogue and systematize the many diverse findings on attitude strength and related attitude dimensions. Finally. and Petty provide an overview of the various ways in which strength-related attributes have been measured and manipulated. and perceived behavioral control. by Petty. 535-552. it seems that a full understanding of attitude strength will require in-depth investigations of each of the attitude dimensions. Following this group. Wegener. Increasing numbers of investigators may well employ this sort of metadimensional approach in the future. Relatively few review articles have been written on this topic to date. This book is intended partly to help make this common core more explicit in ways that might have constructive impact on future studies. and Gross. R. The book begins with Abelson's chapter focused on attitude extremity. Chaiken. ATTITUDE STRENGTH: AN OVERVIEW 19 difficult to reduce the set of dimensions to a smaller set of latent constructs and that doing so may be counterproductive. and Giner-Sorolla consider evaluative-cognitive and evaluative-affective consistency. however. Conviction. intentions. Erber. so it seemed useful for leading researchers to present their perspectives on the accumulated findings and to suggest directions for future studies. and Jaccard. The next chapter. Thompson. the mechanisms by which these consequences are produced. P.. and Miller on certainty. At the same time. & Krosnick. and impact on behavior.
organization. & Schwarz. Gorn. The attraction paradigm. Journal of Personality arlii: SQcw. Heider. A. Fazio. Chaiken. (1990). A. The structure of self-conception: Conceptualization and measurement.. 17. Public Opinion Quarterly. Adolescent attitudes and lawful behavior. 61-80. Higgins (Eds. 339-357. (1971). Janis. R. New Haven. (1982).I"Ps). personal involvement. NJ: Lawrence Erlbaum Associates. Political Analysis. P. Personal ihvolverrt. H.. R. Attitude structure and function (pp. L. H. Auitudes :'u1'd. Journal of Peiulilalrry and.. H. The origins of attitude importance: Self-interest. McDonel.sijcge CU~ [f! pCr~~a$ion. A procedure for cstimafionef trait. A theory oj cognitive dissonance. H. (1970).15-31&. R. American Journal oj Political Science. ATTITUDE STRENGTH: AN OVERVIEW 21 Apsler. Journal qfP"Ts.:CoiHipttlaltindll1ethod()logical r issues. R. D. Attitude accessibility. 63.997-1009.&: Baldwin.tcli. . 18. IL: Row. Political competence and ideological constraint. Cote. Apter (Ed. On the power and functionality of attitudes: The role of attitude accessibility.nan/s. N. EillicaliannJ'arid Piycholi. 1.chology. (J 987). CT: Yale University Press. Psychological Bulletin.. & Maibach. Generalizing. E. E. and change. & Krosnick.20 KROSNICK AND PETIY c 1. (1959). M. 110. Peterson. R. & Sherman. Direct experience and attitude-behavior consistency. (1990). 14. New York: Academic. Reconciling conflicting results derived from experimental and survey studies of attitude change..ali!)nsrudie:s. A theory of social comparison processes. Corill:qual1bnof a'. 446-452. P. F. Green. Skim Psyq. Personal/i)' and SotiaJPs}'cli'OI"8'Y Bull~lll!. M. (1991). Pratkanis. & Brock. (198. Psychological Bulletin. Idler. 436--447. Boruch. Personal involvement: An examination of processing differences. Communication Research. J. CA: Academic. & Sheatsley. Advances in experimental social psychology (Vol.:qnd ajJl!ctas delOl'm. Breckler. 204-243). and source prestige on reactions to communications on separatism.. Attitude content-intensity and probability expectations. attitudes. Johnson. J. Cognitive responses to AIDS information: The effects of issue involvement and message appeal. Hyman. Predicting contraceptive behavior from attitudes: A comparison of within. E. Journal of Applied Psychology. C. 7.·F. TX: Harcourt. (1974). American Psychologist. andbeha~io. Conventional wisdom on measurement: A structural equation perspective. A. R~ding. '14. SlrrVeys . Ganv!:rse. F. as. . 98-106. K. I. B.). 8-17..308-319. Hoelter. L. A. 187-192. (1995). Multiple indicators: Internal consistency or no necessary relationship? Quality and Quantity.. Monterey. L R. 290-295. Judd. Festinger. Ji~uc:.45. 305-314. Journal of Abnormal and Social Psychology. 74. R. (1980) Heuristic versus systematic information processing and the use of source versus m:e:. J. Journal of Personality and Social Psychology. M. Freedman. & Suchman. 4. pp. Flora. 87-96. (1955). & Howard-Pitney. Howard-Pitney. E. (1957). (1983).eplioo of artltudes . Chaili. G.. R. L. (1947). (1967). (1968). (1989). Involvement. 11. l. W. Chen. 39-57. D. (1986). Campbell. Borgida. Hovland. R. Self-rated health and mortality in the NHANES-I epidemiologic follow-up survey. & Lennox. A. N..oll Quaflerl}" 54. C. Boninger. S. (1988). R. J.. 106. L.. K.vmlly. 206--261).(. Public Opinion Quarterly.ve una. N. (1982). 39. American Sociological Review. P. Biek. 4 Brim. H. & Berent. E. & Sears. Journal of Personality and Social Psychology.. W. J. E. E. Guttman. Social Science Research. Festinger. Mandell. McGuire. A. (1954).amJoQgnilion. ':f09-2IS_ .. and personal affairs.ny. D. The order oj presentation in persuasion (pp. & Zanna.Thfle (Ed. S. I. S. Q'''"ilitaJ. Bollen. 7. L. (1984). social identification. S. In C. E (11170).ly~i~ of social problems (pP' l6S~189).. R. C. 38.. 30. Journal oj Personality and Social Psychology. method. 162-166. Cohen. Texas.. . H. 57-67. 752-766. 1392-1407. 69. 560-570. M.1994). J. 16. J. R. O. C.. Ideology arid discontent (pp. H. H. & Marcus. Sorrentino & E. Brock.. T. 9.). Springfield.. H. e-'illl~go: Starlon. 42. R. Jackson. The effects of personal involvement. Anderson. 32.( 1990). R. A. and the strength of the object-evaluation association. I. 18. San Diego.. Hillsdale. Some reasons why information campaigns fail.. Journal oj Marh!lit(8 Reseo. Journal oj Personality and Social Psychology. Bern. 80. across 10 C(!r)SI1:UCl va-lid. (. 7. A. S. T.. M. ttal~ method. & Morrison.ln:E. 20. i\Jfetdv~-cognllhecQiLSistency and the effect of salient bcl1(lYIo'!'aJ infarrITa. H. EmpIrtci\J validatlon of meCl. E. R.' & Buckle.w:l~ respenseefferxs' 10 public epiBiQli. L. 45. 50. D. Fazio. Byrne. J. J. J..en. (1993).en1. . G. 93--111. (1958). R. & Eagly. Arvey. 20. E. 547-574. B. A study of certain factors involved in changes of opinion. Crites. Krosnick. Segal. T. The psychology of attitudes. Hovland.\'. Personality and persuasion: Need for cognition moderates the persistence and resistance of attitude changes. S. (1989). L. and value-relevance. & Pe. T.. F.dlijJogue. 758-780. L. Journal of Personality and Social j'syaholtJgy. (1975). D. G. Borgida. E. E. Not forbidding isn't allowing: The cognitive basis of the forbid-allow asymmetry.. Ewing. Te~li!! A & M Uni. (1989)..r . L.Wisley. 63--88. A..distinct components of attitude. & Angel. R. 49. S. K. (1989).. The effects of 'commitment' on opinion change following communication. W. A. H. Public Opinion Quarterly. . Brace. J. 412-423.oi. Jovanovich. Journal of Personality and Social Psychology. M.<>gy.. 290-314. J. Journal oj Personality and Social Psychology. In L. L.).e!. 14.swemeT!l. M.eIlt and the robustness of perceptual salience effects. S. 68. (1992). The biological basis of personality.l dissertation. A.rgicaIMg(!.. On the dimensionality of public sentiment toward panisan and ideological groups.. Measuring issue preferences: The problem of response instability. E. (1947). & Abraham. Attitude centrality. 1. M. (1986). (1942). 68-76. Bishop. Journal of Personality and Social Psychology. Berkowitz (Ed. The handbook of motivation and cognition: Foundation of social behavior (pp. altiiU!ie JIa. Human Relations.759-774.-i-2 Fazio.191-. T.versus across-subjects procedures.. (1983). (19'84). Intensity and a zero point for attitude analysis. 117-140. Hippler.ncllllledg.. New York: Guilford. and attitude change. The nature of belief systems in the mass public. Beliefs. Bollen. B. American Journal oj Public Health.lliliry... W. J. 1. G. (1964). L. D. Journal of Social Psychology. How do attitudes guide behavior? In R. communication discrepancy. 4. & Chaiken.. Campbell. Canadian Journal of Behavioral Science. . Luchins. Job satisfaction: Environmental and genetic components. 12.( f9'92). 019--634. & Omoto. Warning. 3. Hovland. Greenwald (Eds. O. (1964). 25-60. In A.ESlimw-ing. Feierabend. 153--179). The psychology of interpersonal relations. P.' ah(l errervarieoee. 377-385.onaliry und So~iol Psychology. 161-202). I. and error variances attributable te a m'e·a5U1il. (1981). P. W... P. T. i~v91vem. Feldman. New York: Wiley. G.M~urlngthe affiidve and cegnltlvepropenjes of iluitucl. Fabrigar. H. E. R.). Evanston. on Converse. T. Eagly. Addison. J. D. (1957). Breckler.1205. & Petty. Fort Worth. C. Fazio. J.. Social Cognition. A. 41:. C. Pitb/ic ()pln. (1970). D. M. (1986). Unpublished doctQi<i. M. B. W.Thf.. Effects of involvement on persuasion: A meta-analysis. In D. (1985). & A. bl!ltavic&'. Davidson. & Wolins. and measurement.. discrepancy. Haugtvedt. MA.).. (1975). Brown. & N.Utl:tiri'ies. American Sociological Review. IL: Thomas. CA: Brooks/Cole. A.llon the ~df-pe!i!. H. New York: The Free Press. Bouchard. C. Eysenck.1). attitude-behavior consistency. 23-32).369-386. S.S.~'oi1.offrrformafion prou~Sii!g..
L. New York: Collins-World. (1968). J. P. V. (1988). New York: Springer-Verlag. M. R. confidence and salience of an opinion. (1990). the attitude-behavior relationship. (1949--1950). Questions about survey questions (pp. H. Advances in experimental social psychology (Vol..).). (1956). K. Roberts. & Petty.). E. Cantril (Ed. Initial attitude toward source and concept as factors in attitude change through communication. wording. (1956). D. A comparison of three measures of attitude intensity. C. (J 984). (1951). Schuman. Smith. R. E. J.• & Browne. Star. Social Forces. Social Psychological Quarterly. New York: Wiley. & Hyman. 1132-1151. H. 54. 151-162. 22. M. In D. P. Columbus.. (Ed. Tedin. University of Missouri. 245-285). R. Brown. (1991). Powell. K.). Katz. International Journal of Opinion and Attitude Research. Self-generated attitude change.. (1980). & Schuman. D. The psychology of commitment: Experiments linking behavior to belief San Diego. A. Issue involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses.. Columbus. Kiesler. 533-541. New York: Springer-Verlag. (1979). Questions and answers in attitude surveys: Experiments on question form. J. R. Krosnick. (1986)..22 KROSNICK AND PETrY 1. and the Elaboration Likelihood Model of persuasion. E. (1%0). In L. Assessing peer and parental influence on adolescent political attitudes. H. Sherif. J. Annual Review of Sociology. L. A. P. A. Philadelphia: Saunders. Berent. 1915-1926. A. 77. J. P. Petty. 20.. K. Southern Speech Communication Journal. A. 55. A model of the extremity.. 2. G. J. 393-414. II).2098-2109. W. 6. S. A. 54. OH. (1993). McKechnie. Lord. (1992). New York: Harper & Row. S. Princeton. CA: Academic Press. Journal of Abnormal and Social Psychology. 53-73). & Presser. J. A. 24.). L. R. B. & Gardner.. J.... 121-132. 36. Attitude research plays for high stakes (pp. Attitudes and behavior. 38. E. C. Suchman. J. (1986). S. Individual differences in attitudinal versus normative determination of behavior. & Lepper. Attitude strength: One construct or many related constructs? Journal of Personality and Social Psychology. 95-99.. S. Basic and Applied Social Psychology. Krosnick. R. 48. Petty. E. (1985). Public Opinion Quarterly. Social Forces. (1988b). Journal of Experimental Social Psychology. A.. 37. CA: Academic. (1988). & Alwin. A. & Cacioppo. L. The functional approach to the study of attitudes. E. Krosnick. In S. 24. Attitude intensity. The measurement of intensity. Attitudes versus actions. A. Temporal instability as a moderator of the attitude-behavior relationship. A sociocognitive model of attitude structure and function. 161-207. Ego-involvement and persuasion: An appreciative look at the Sherifs' contribution to the study of self-relevance and attitude change. extremity. The intensity component in attitude and opinion research. Rennier. L. 526-538. Jr. J. Unpublished doctoral dissertation. 204--273). Aronson (Eds. (1976). 52. Reading. A. A. (1988a). A.). (1993).. & Greenwald. MO. Webster's new twentieth century dictionary of the English language (2nd ed. G.. & 1.. The case for measuring attitude strength in surveys. 1. pp. E. 1287-1302. M. Advances in Experimental Social Psychology (Vol. T. M. (1968). Relations among dimensions of attitude strength: Insights from induced variable models. C. pp. 22. Petersen. . (1971). J. The use of causal indicators in covariance structure models: Some practical issues.. 106-114.). A test of the form-resistant correlation hypothesis: Ratings. (1950). F. J. Attitude measurement. (1978). C. New York: Academic. 107. J. 7. 493-512. Journal of Personality and Social Psychology. R.& G. R. Cognitive structure and attitudinal affect. K. Petty. University of Michigan. R. J. Tannenbaum. Chuang. Princeton. New York: Academic. (1978). (1934). Policy voting in American Presidential elections: An application of psychological theory to American politics. G . Journal of Personality and Social Psychology. and certainty and susceptibility to response effects. & Grush. (1972). Social Forces. 177-203). 230-237. (1989).. (1977). and voting behavior. Handbook of social psychology (Vol. 51-65). In G. Chicago: American Marketing Association. M. 1. A.312-330. T. H. Public Opinion Quarterly. & Abelson. Strathman. Journal of Personality and Social Psychology. H. Ann Arbor. Rosenberg. R. Krosnick. A. Suchman. Lastovicka. The strength of the object-evaluation association. Attitude importance and attitude change. 65. (1979). presidential candidate evaluations. P. Krosnick. Lindzey & E. (1944). Y.). Journal of Personality and Social Psychology.190-202. The role of attitude importance in social evaluation: A study of policy preferences. R. 240-255.). A. Journal of Experimental Social Psychology. Stouffer. Tanur (Ed. S. E. Columbia. rankings.. (1965). G. E. & Cacioppo. NJ: Princeton University Press. 221-241. Biased assimilation and attitude polarization: The effects of prior theories on subsequently considered evidence. (1985). Guttman. 163-204. (1976). 136-154. Silverman (Eds. & Haugtvedt. L. J. 361-372. & Cacioppo. Measurement and prediction (pp. 367-374.. Components of involvement. McDill. A. In J. Public Opinion Quarterly. Ohio State University. A. In L. 1. M.. 37. 24. (1992). & Nebergall. & Johnson. C. Miller.).. Strength-related attitude dimensions. Swann.. W. Strathman. and context. D. 13. 147-174). R. H. D. Krosnick. Maloney & B. M. 96-210. Granberg . Berkowitz (Ed. E. IA: Wm. (1979). Bruner. Sternber. E. American Journal of Political Science. Schuman. Krosnick. Attitudes and persuasion: Classic and contemporary approaches. C. D. Unpublished doctoral dissertation. C. Miller. D. N. R. W. (1986). Pratkanis. (1965). W. (1959).• Ross. Ohio State University.. British Journal of Social Clinical Psychology. Journal of Personality and Social Psychology. T. 367-372. intensity: Neglected variables in research on attitude-behavior consistency. & Ely. Certainty. Lewin. On the ambivalence-indifference problem in attitude theory and measurement: A suggested modification of the semantic differential technique. (1981). 413-425. Petty. H.. (1956). ATTITUDE STRENGTH: AN OVERVIEW 23 Kaplan. B. 114. Involvement and dogmatism as inhibitors of attitude change. T. Tesser. Opinions and personality. F. T. (1981). Involvement and persuasion: Tradition versus integration. J. How interviewer effects operate through question form. MA: Addison-Wesley.. Gauging public opinion (pp. Berkowitz (Ed. Psychological Bulletin.. E. Sarup (Eds. Journal of Experimental Social Psychology. A. Investigation of the influence of need for cognition on attitude-behavior consistency. Psychological Bulletin. L. 940-952. J. J. In H. Sherif. A. New York: Russell Sage. W. 715-724. Boninger. & Carnot. 1. Lemon. L. Scott. The attitude-memory relationship after 40 years: A meta-analysis of the literature. Lazarsfeld. (1975). 213-276). Attitude and attitude change. Petty. Clausen (Eds. Katz. K. Unpublished doctoral dissertation. A battle of wills: Self-verification versus behavioral confirmation. Field theory in social science. J. In J. Krosnick. Raden. 2. San Diego. C. Unpublished manuscript. C. & Cacioppo. & White. Dubuque. Social judgment and intergroup relations: Essays in honor of Muzafer Sherif(pp. Jarvis. and the measurement of values. 46. Communication and persuasion: Central and peripheral routes to attitude change. 3. MacCallum.. N. 53. importance. Schwartz. (1988). NJ: Princeton University Press. Psychological Bulletin. (1994). B. M. A. 1. Satirical persuasion and topic salience. E. R. & Dutton.. LaPiere.. E. Journal of Personality and Social Psychology. F. 42. Cacioppo.
Wilson. Weigel. Kraft. (1976). & Dunn. attitude extremity.. Carryover effects in attitude surveys. 25. 793-802. usually labeled at the ends with the respective designations. 379-400. Personality and Social Psychology Bulletin. 25. & D' Andrade. The importance of heritability in psychological research: The case of attitudes. attitude accessibility for its theoretical coherence. Rasinski. N.. extremely unfavorable and extremely favorable.. 115-122. Krosnick and Abelson (1991) recommended three measures: attitude importance for its simplicity and sheer frequency of appearance in a variety of relationships... A. & Newman. W. Some measures of attitude strength command attention because of ease and frequency of empirical use. Journal of Social. 42. 798-810. T. Wicker. Persuasive communication of risk information: A test of cue versus message processing effects in a field experiment. (1991). D.. Bradburn.. (1982). Belief accessibility and context effects in attitude measurement. Verplanken. 25. British. 401-421. 129-142. Psychological Review. B. some because of theoretical and empirical coherence. & D' Andrade. R. Journal of Experimental Social Psychology. A. Public figures and ordinary folks alike often cling tenaciously to beliefs and attitudes that we. Journal of Personality and Social Psychology. R. 188-193. 25 . 53. Verplanken. 33. but the meaning of placements at less-thanextreme positions is left to the judgment of the subject-an option whose consequences are discussed later in this chapter. H. 2 Attitude Extremity Robert P. The disruptive effects of explaining attitudes: The moderating effect of knowledge about the attitude object. 41-78. This chapter focuses on the last named. Journal of Experimental Social Psychology. because conflict can breed attitude extremity. Neutral may also be labeled. -. Wood. lOa. R. (1993). K. Tourangeau. Journal of Personality and Social Psychology. K. and some because they are relevant to important social phenomena In a recent review of a large number of candidate measures of attitude strength. A. (l989b).24 KROSNICK AND PETrY Tesser. Public Opinion Quarterly. Rasinski. which in turn serves to maintain or increase conflict. I analyze later several interesting social psychological processes that may be implicated in this vicious cycle. Journal of Social Issues. Extremity has been conventionally measured by selfplacement of subjects along a numerical scale of attitude position. A. B. W. Psychology. (1989a). R. (1989).. One social arena in which extreme attitudes are consequential is that of intergroup conflict. D. Attitudes versus actions: The relationship of verbal and overt behavioral responses to attitude objects. Tourangeau. (1969). 495-524. and attitude extremity for its practical importance. S. (1989).. Involvement and need for cognition as moderators of beliefs-attitude-intention consistency. L. N. R. S. BACKGROUND Throughout my academic career I have been fascinated by the capacity of holders of very strong attitudes to resist persuasive attempts at change. Retrieval of attitude-relevant information from memory: Effects on susceptibility to persuasion and on intrinsic motivation. Bradburn. D. Abelson Yale University Different meanings of attitude strength attach to different measures of what is almost certainly a multidimensional concept. 17. Increasing attitude-behavior correspondence by broadening the scope of the behavioral measure. 28.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.