an integrative model of ambivalence

21
The Social Science Journal 47 (2010) 609–629 Available online at www.sciencedirect.com An integrative model of ambivalence Young Min Baek The Annenberg School for Communication, University of Pennsylvania, 3620 Walnut Street, Philadelphia, PA 19104-6220, USA Received 6 July 2009; received in revised form 24 December 2009; accepted 1 February 2010 Abstract Ambivalence is a widely experienced psychological state, but inter-disciplinary studies, to a certain extent, define and conceptualize ambivalence independently. In spite of its lack of clarity, ambivalence has become an increasingly popular concept, utilized in hypotheses concerning a variety of social phenomena. This study provides an overview of extant studies on ambivalence, and summarizes the similarities and differences in how practitioners of social psychology, political science, and sociology have adopted the concept. A survey of literature from the three fields suggests four distinctive definitions of ambivalence or antecedents that have caused ambivalence: (1) co-activation of both positivity and negativity; (2) co-emergence of conflicting attitudes; (3) co-constraint of conflicting values; and (4) co-existence of conflicting reference groups. Some potential problems, such as inconsistent findings and lack of relevant measures or indices are indicated, and alternative methods are suggested. The paper concludes by suggesting a more sophisticated and precise integrative model of ambivalence. © 2010 Western Social Science Association. Published by Elsevier Inc. All rights reserved. 1. Introduction People frequently come across objects or events that lead them to report internal conflicts or mixed feelings on the objects/events, frequently called ambivalence. In studies on ambiva- lence, scholars define it as “resulting from the conflicting positive and negative thoughts and feelings that a person holds toward an object” (Priester & Petty, 2001, p. 19). Given the recent academic interest in diversity, disagreement, and social network heterogeneity, the con- cept of ambivalence has also become increasingly central to scholarship in social psychology (Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001; Priester & Petty, 1996, 2001), political Tel.: +1 215 573 1979; fax: +1 215 898 2024. E-mail address: [email protected]. 0362-3319/$ – see front matter © 2010 Western Social Science Association. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.soscij.2010.02.003

Upload: young-min-baek

Post on 28-Oct-2016

219 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: An integrative model of ambivalence

The Social Science Journal 47 (2010) 609–629

Available online at www.sciencedirect.com

An integrative model of ambivalence

Young Min Baek ∗

The Annenberg School for Communication, University of Pennsylvania, 3620 Walnut Street, Philadelphia, PA19104-6220, USA

Received 6 July 2009; received in revised form 24 December 2009; accepted 1 February 2010

Abstract

Ambivalence is a widely experienced psychological state, but inter-disciplinary studies, to a certainextent, define and conceptualize ambivalence independently. In spite of its lack of clarity, ambivalencehas become an increasingly popular concept, utilized in hypotheses concerning a variety of socialphenomena. This study provides an overview of extant studies on ambivalence, and summarizes thesimilarities and differences in how practitioners of social psychology, political science, and sociologyhave adopted the concept. A survey of literature from the three fields suggests four distinctive definitionsof ambivalence or antecedents that have caused ambivalence: (1) co-activation of both positivity andnegativity; (2) co-emergence of conflicting attitudes; (3) co-constraint of conflicting values; and (4)co-existence of conflicting reference groups. Some potential problems, such as inconsistent findingsand lack of relevant measures or indices are indicated, and alternative methods are suggested. The paperconcludes by suggesting a more sophisticated and precise integrative model of ambivalence.© 2010 Western Social Science Association. Published by Elsevier Inc. All rights reserved.

1. Introduction

People frequently come across objects or events that lead them to report internal conflictsor mixed feelings on the objects/events, frequently called ambivalence. In studies on ambiva-lence, scholars define it as “resulting from the conflicting positive and negative thoughts andfeelings that a person holds toward an object” (Priester & Petty, 2001, p. 19). Given therecent academic interest in diversity, disagreement, and social network heterogeneity, the con-cept of ambivalence has also become increasingly central to scholarship in social psychology(Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001; Priester & Petty, 1996, 2001), political

∗ Tel.: +1 215 573 1979; fax: +1 215 898 2024.E-mail address: [email protected].

0362-3319/$ – see front matter © 2010 Western Social Science Association. Published by Elsevier Inc. All rights reserved.doi:10.1016/j.soscij.2010.02.003

Page 2: An integrative model of ambivalence

610 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

science (Alvarez & Brehm, 1997; Huckfeldt, Mendez, & Osborn, 2004; Mutz, 2006), andsociology (Lane, 1962; Merton, 1976). Given the concurrent increase in ambivalence researchin the three fields, this study reviews the theory and conceptualization that these fields pro-vide, suggests what is lacking, and aims to propose a more sophisticated and precise model ofambivalence to guide social scientific studies.

2. Ambivalence and other psychological states

In order to understand ambivalence, it is useful to examine other psychological states thatshould be distinguished from ambivalence. First, ambivalence is different from indifference,although both frequently lead to the same behavior. Take, for example, two White citizens,“A” and “B,” who did not vote for Barack Obama in the 2008 presidential election. “A”was indifferent about the election (i.e., had no interest in politics), while “B” felt ambivalenttowards Obama (i.e., “B” wanted a Democratic president but did not want a Black president).In other words, “A,” who felt indifferent about an attitude object (i.e., Obama) did not expe-rience the internal conflict that “B” did because “A” had no psychological attachment to theobject (Kaplan, 1972, p. 368). Psychological indifference does not entail any psychologicalinvolvement or arousal (Abelson, 1988; Cacioppo & Berntson, 1994), while ambivalence does.

Second, ambivalence is also distinct from uncertainty. According to Downs (1957), uncer-tainty is a psychological state indicating a “lack of knowledge” (Downs, 1957, p. 77). Forexample, let us imagine two new people, “A” and “B,” who indicate in an opinion survey thatthey do not have a fixed preference (i.e., neither support nor oppose) regarding the healthcarereform issue. Although their present policy preferences seem similar, their psychological statescould be different. “A” is uncertain about the reform because “A” does not know much aboutthe issue. On the other hand, “B” is ambivalent towards the issue because “B” agrees it isnecessary to increase health insurance coverage but is highly concerned about the possibilitythat universal healthcare may cause an enormous government deficit. In other words, “A” feelsuncertain about an attitude object (i.e., healthcare reform) but does not experience internalconflict as “B” does because “A” does not have enough information to form a preference aboutthe object. Therefore, uncertainty is likely to disappear as knowledge increases (Alvarez &Brehm, 1997), but ambivalence is not resolved by increased information or knowledge.

Third, ambiguity – defined as “the subjective experience of missing information relevant toa prediction” (Frisch & Baron, 1988, p. 152)1 – is also a different psychological state fromambivalence. Put differently, ambiguity arises from a person’s ignorance of key social phe-nomena, whereas ambivalence emerges with conflicting evaluations even when an individualsufficiently understands the object. For example, let us imagine two citizens, “A” and “B,” whooppose the construction of nuclear power plants. Despite their shared attitude, the underlyingpsychological mechanism may be different in the two people. “A” has an ambiguous attitudeabout the utility of nuclear power plants because “A” does not know the scientific rationalebehind nuclear energy generation, but “B” is ambivalent due to his/her acknowledgment ofboth its benefits and risks (such as the potential for another disaster like Chernobyl of 1986).

Fourth, ambivalence should be distinguished from dissonance. Whereas ambivalenceemerges from an individual’s contrasting evaluations (e.g., positive and negative) of one attitude

Page 3: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 611

object (Cacioppo & Berntson, 1994; Priester & Petty, 1996, 2001), dissonance is imbalancedevaluation resulting from conflicting cognitions of two (or more) different attitude objects(Festinger, 1957; Heider, 1958). For example, dissonance exists when person “A” (i.e., a per-ceiver, P) likes slasher films (i.e., an object, X) as well as his girlfriend (i.e., another person,O), but “A”’s girlfriend does not like slasher films. According to Heider (1958), dissonanceleads to the reconstruction of cognitive components (e.g., giving up X to avoid conflicts withO) because the imbalanced relationship between the two attitude objects (X and O) causes theperceiver (P) some psychological discomfort. In this example, “A” may give up slasher filmsto avoid conflict with his girlfriend. While dissonance is observed when two different atti-tude objects are competing, ambivalence arises when a person makes contrasting evaluationsof one attitude object. Thus, dissonance and ambivalence are theoretically related only whenconflicting attitudes lead to conflicting evaluations of one object.

Fifth, ambivalence is different from inconsistency. When people are ambivalent about anobject, they show inconsistent attitudes or evaluations. Conversely, people might have incon-sistent attitudes toward an object without feeling ambivalent because they might be unawareof their inconsistent attitudes or evaluations (Converse, 1964; Zaller, 1992). For example, acitizen who identifies him/herself as a political liberal might support anti-immigration pol-icy, which in general is perceived as conservative, because the citizen wants “job security.”These inconsistent attitudes indicate psychological ambivalence only if the citizen is awarethat his/her generalized political ideology conflicts with his/her anti-immigration position.Thus while inconsistency is a condition of ambivalence, it is not a necessary one.

To summarize, ambivalence is distinguished from indifference, uncertainty, ambiguity,dissonance and inconsistency. Although dissonance and inconsistency are related with theemergence of ambivalence, they cannot be equated with ambivalence itself. Throughout thepaper, I define ambivalence as a “psychological state caused by contrasting evaluative ori-entations toward an object, which influences an individual’s decision-makings or behaviors”(Converse, 1964; Kaplan, 1972; Merton, 1976; Priester & Petty, 2001; Zaller, 1992). Basedon this definition, this study will examine a wide range of ambivalence studies and suggest amodel that integrates those studies.

3. Theoretical overview of ambivalence

This section will survey how studies conceptualize or define ambivalence in three socialscientific disciplines—social psychology, political science, and sociology. The differentiationbetween the three might seem quite arbitrary because ambivalence is a widely experienced phe-nomenon, and it is consequently an inter-disciplinary research topic. However, distinguishingbetween these disciplines will be helpful to understanding the reasons why there are differentdefinitions of ambivalence in different academic orientations.

3.1. Social psychological perspective2: what ambivalence is and how it happens

Ambivalence studies began with the criticism of Thurstone’s definition of an attitude aslocated on a bipolar continuum ranging from negative to positive, with neutral located in

Page 4: An integrative model of ambivalence

612 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

Fig. 1. Types of attitude.

between (Thurstone, 1928). While nearly all attitude scales (Thurstone, Guttman, Likert, andsemantic differential scale) assume a bipolar continuum (Eagly & Chaiken, 1993, see chapter2), empirical findings point to several problems inherent in the so-called bipolar attitude scale(Cacioppo & Berntson, 1994). Specifically, while some individuals are indifferent and othershave mixed feelings toward an object, both report the same neutral attitude. For example, manysurvey studies have found that the inclusion of “Don’t know” or “Refusal” choices significantlyreduces theportion of neutral attitudes reported (Schuman & Presser, 1980). In other words, theneutral point on an attitude scale is the safest and most frequently selected choice for peoplewho have indifferent or ambivalent attitudes.

Social psychologists have thus suggested revising methods for measuring attitudes by sep-arating positive and negative components (Kaplan, 1972; Thompson, Zanna, & Griffin, 1995).In contrast with a bipolar continuum, a bivariate concept of attitude can expand understandingof attitude further. By placing negativity and positivity on an X- and Y-axis (see Fig. 1), severaltypes of attitude are defined as follows (Priester & Petty, 1996):

• A univalent (positive or negative) attitude, in which negativity overrides positivity or viceversa;

• An indifferent attitude, in which neither positivity nor negativity are activated;• An ambivalent attitude, in which both positivity and negativity are co-activated.

Studies with a social psychological perspective will be defined as having an intra-attitudinalapproach because they focus purely on individual evaluation of an object without consider-ing theoretical antecedents or outcomes. Studies of an intra-personal approach (Cacioppo &Berntson, 1994) focus on ambivalence in human psychological processes. Studies in cross-cultural psychology also clarify that positivity and negativity are evaluated separately andfrequently co-activated (Baumeister et al., 2001; Rozin & Royzman, 2001). Additionally, brainMRIs (magnetic resonance imaging) reveal that both positive and negative events stimulateseparate areas in the human brain (Cacioppo, Petty, Losch, & Kim, 1986). Studies of facialexpression also report that the facial muscles of humans reveal different reactions to different

Page 5: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 613

emotional experiences. For example, the corrugators supercilii muscle moves when one facesnegative events, but the zygomaticus major muscle moves when one experiences positive events(Dimberg, Thunberg, & Elmehed, 2000; Larsen, Norris, & Cacioppo, 2003; Steiner, Glaser,Hawilo, & Berridge, 2001).

One of the most important findings comes from arousal studies, which suggest that peoplereact differently to positivity and negativity, based on their level of arousal. While at lowerarousal levels, people are more likely to evaluate relatively neutral situations as positive (i.e.,positive offset), those who are highly aroused pay more attention to negative stimuli (i.e.,negative bias). People’s asymmetric reactions are widely studied (for reviews, see Baumeisteret al., 2001; Rozin & Royzman, 2001). The literature explains that these two biases are basedon asymmetric experiences of both positive and negative events in the person’s life. Humanlives are generally dominated by positive experiences, or at least by unharmful experiences.However, when the situation becomes critical or new, negativity gets more attention becauseof its rarity. For example, impression management theory includes the idea that impressionformation depends more upon negative than positive evaluation (Peeters & Czapinski, 1990).Prospect theory (Tversky & Kahneman, 1981) also emphasizes the human propensity forloss aversion, based on the notion that “losses loom larger than corresponding gains” (p.1039). Taken together, the intra-attitudinal approach focuses on people’s different reactionsto negativity and positivity, and ambivalence arises when both positivity and negativity aresimultaneously activated at similar arousal levels.

Another approach to ambivalence in social psychology can be found in studies on inter-attitudinal inconsistency (Eagly & Chaiken, 1993; Katz & Hass, 1988; Wilson, Lindsey, &Schooler, 2000). Inter-attitudinal inconsistency studies emphasize that ambivalence resultsfrom a person’s unstable psychological structure, which is made up of dynamic or contrastingcomponents. Stated differently, inter-attitudinal inconsistency is a theoretical antecedent thatcauses intra-personal ambivalence. I will divide the inter-attitudinal approach into two groups:(1) vertical inconsistency (e.g., dynamics between implicit and explicit attitudes, especiallytoward stigmatized or minority groups), and (2) horizontal inconsistency (e.g., inconsistentattitudes toward different attributes of an object).

The vertical inconsistency approach emerges from the psychological conflict betweenexplicit (or controlled) and implicit (or automatic) attitudes (Fazio, 1995; Katz & Hass, 1988).I label this approach vertical because it assumes that the explicit or controlled attitude is con-scious, but the implicit or automatic attitude is largely unconscious.3 For example, peoplemay hold positive attitudes toward Afro-Americans because they sympathize with disadvan-taged Blacks as a minority underdog (i.e., their attitude stems from values of communalism);but simultaneously they could have negative attitudes towards Blacks because they think thatBlacks are poor due to a lack of self-reliance (i.e., their attitude stems from a protestant workethic). If a Caucasian American experiences these two opposing attitudes simultaneously,then he/she will feel internal conflict, which may lead to subjective ambivalence. Facing racialissues, most Americans no longer have explicit negative attitudes toward stigmatized or minor-ity groups, but negative attitudes still survive as implicit attitudes (Katz & Hass, 1988). Fazio(1995) reports that people experiencing conflicting attitudes take a longer response time incompleting attitude questionnaires than those who do not (also see Bassili, 1996; Bassili &Fletcher, 1991). Similarly, Wilson et al. (2000) argue: even though an explicit attitude change

Page 6: An integrative model of ambivalence

614 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

may occur, old attitudes do not disappear, but rather become implicit. Thus, in some contextswhere new attitudes do not control people’s behavior, old attitudes sometimes subvert the newones. This mechanism of ambivalence creation is not easily measured via conventional surveytechniques because implicit attitudes are stored latently.

The horizontal inconsistency approach was similarly conceptualized, but it does not assume ahierarchical structure among attitudes. Horizontal inconsistency is characterized by conflictingattitudes toward multiple attributes of an object (Campbell, 1983; Holbrook & Krosnick, 2005;Lavine, 2001; Tetlock, 1986). To exemplify this, let me assume that the final evaluation ofa presidential performance is based on attitudes toward the president’s handling of (1) theeconomy, and (2) national security. Citizens strongly favoring the president place differentweight on the two issues (e.g., President Obama is good at handling national economy, whichis the most important presidential task). Intra-personal ambivalence, however, is more likelyto arise when a citizen has a good attitude toward one issue and simultaneously holds a badattitude toward the other issue (e.g., President Obama is good at handling the national economybut bad at guarding national security). The final evaluation of an object is vulnerable to externalinfluence, for example news reporting on the prime specific attributes of an evaluative object(Iyengar & Kinder, 1987; Price & Tewksbury, 1997).

Although both the vertical and horizontal approaches are based on inter-attitudinal inconsis-tency, their consequences are strikingly different. Vertical inconsistency causes more extremebehaviors (i.e., amplification effects, Katz & Hass, 1988), while horizontal inconsistency leadsto delayed and hesitant behaviors (i.e., moderation effects, Tetlock, 1986). Due to the ampli-fication effects of ambivalence (Katz & Hass, 1988), if a White supervisor is dissatisfiedwith a Black worker’s performance, the supervisor might experience more internal conflictbecause he/she must choose only one of his/her self-concepts: (1) as a “humane person” caringfor a minority underdog, or (2) as an “objective person” treating every employee as equal.Thus, in order to protect either self-concept, the supervisor might treat the worker either “verypolitely” (i.e., protecting him/herself as a “humane” person caring for the underdog) or “veryharshly” (i.e., protecting the self-concept as an ‘objective’ person who does not favor anyone).In short, people with vertically inconsistent attitudes must select from polarized behavioralchoices.

However, the moderation effects of ambivalence on behaviors (Tetlock, 1986) reveal anotherside of the story. Ambivalence fosters more complex thinking about an object, which neutral-izes the extreme evaluation. Ambivalence triggered by horizontal inconsistency thus preventspeople from making quick decisions and delays their behaviors.

Why are these explanations contrasting? Two differences deserve to be noted. First, thevertical inconsistency approach focuses on one attitude’s dominance over the other, whereasthe horizontal inconsistency approach emphasizes balancing the accounts of relative advan-tages or disadvantages. With vertical inconsistency, therefore, one attitude is more likely toinfluence behaviors in ways that deny the other hidden attitude. With horizontal inconsistency,different attitudes coexist and people deliberate in an attempt to balance them, provided thatthere is no external influence, such as the media’s priming effects (Iyengar & Kinder, 1987;Price & Tewksbury, 1997). Second, the former approach emphasizes the desire for appropri-ate self-presentation in social situations (e.g., I want to be perceived by others “as a humaneperson”), but in the latter, the individual tries to maximize self-interest through careful eval-

Page 7: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 615

uations (e.g., how can I make a better decision?). In other words, the perceived self is notthe same. Because of vertical inconsistency in attitudes, one’s self-image is divided (e.g.,public self vs. private self), and one’s public self-presentation may be strategically differ-ent depending on the situation (e.g., a White American’s expression about racial issues maybe different if his/her conversation partner is White or non-White). In the vertical inconsis-tency approach, one self-image should be suppressed to make the other self-concept moresalient. However, the horizontal inconsistency approach assumes a unified self who does thebest to maximize his/her own gains (or to minimize his/her own loss). For example, one maybelieve, “My decision should be well-considered and balanced in order to maximize my inter-ests.” Thus, ambivalence caused by horizontally inconsistent attitudes creates more cautiousbehavior.

Before closing the examination of ambivalence from a social psychological perspective,it should be noted that although the intra-attitudinal and inter-attitudinal approaches aredifferent, they could be integrated. The intra-attitudinal approach examines ambivalencecaused by two contrasting evaluative components (Cacioppo & Berntson, 1994; Kaplan,1972; Priester & Petty, 1996), but the inter-attitudinal approach highlights the fact that intra-attitudinal ambivalence is caused by conflicting evaluations in different attitude domains.This could be multiple attitudes based on different value systems (e.g., Katz, 1981; Katz& Hass, 1988); cognitive attitudes toward complex realities (e.g., Tetlock, 1986); or sepa-rate attitudes toward multidimensional attributes of an object (e.g., Holbrook & Krosnick,2005; Lavine, 2001). If it is possible to assume that conflicting attitudes can be reducedinto bivariate evaluative components (i.e., positivity or negativity), inter-attitudinal inconsis-tency can be considered as the origin of intra-attitudinal ambivalence (Wegener, Downing,Krosnick, & Petty, 1995). This can be exemplified with attitudes toward abortion. It is rea-sonable to assume that pro-choice values lead to a positive evaluation of abortion, whereaspro-life values lead to a negative evaluation of abortion. Conflicting attitudes toward thetwo different values lead to simultaneous activation of both positive and negative evalua-tions of abortion. In other words, inter-attitudinal inconsistency is the cause of intra-attitudinalambivalence.

Here it is necessary to ask: “Where does inter-attitudinal inconsistency come from?” Asimplied earlier, inter-attitudinal inconsistency has supra-individual origins, such as values (e.g.,a protestant work ethic or communalism), worldviews (e.g., pro-choice or pro-life), or ideolo-gies (e.g., Liberalism or Conservatism). These supra-individual origins are traditional topicsin political science. Although the political scientific approach to ambivalence springs from adifferent angle of research than the social psychological approaches, the findings or theoreticalimplications are similar. In short, the political scientific approach identifies the supra-individuallevel of conflict as the psychological origin of an individual’s ambivalence.

3.2. Political Scientific perspective4: supra-individual causes of ambivalence

Ambivalence studies by political scientists can be traced to two theoretical legacies: (1) sym-bolic racism (Kinder & Sears, 1981) and (2) non-attitude thesis of citizens’ opinion (Converse,1964). The two traditions share the same political concept (i.e., ideology), but their definitionsand implications are qualitatively different.

Page 8: An integrative model of ambivalence

616 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

First, the ideology (or value system) of symbolic racism is conceptually higher-order. Kinderand Sears’s (1981) definition of symbolic racism is similar to false consciousness in the Marxistdefinition of ideology. For example, they argued,

Since the explicitly segregationist, white supremacist view has all but disappeared, it no longercan be a major political force. What has replaced it, we suggest, is a new variant that mightbe called symbolic racism. . . .Symbolic racism represents a form of resistance to change inthe racial status quo. . .. The stereotypical symbols of blacks’ violation of traditional values,which are in the media and informal communication all the time, may be more important thanthe realities of the more occasional actual damage blacks do to whites’ own lives (p. 416, Italicemphasis original, Underlined emphasis added).

As the authors argued, symbolic racism is a disguised form of White supremacists’desire/attempt to preserve racial discrimination. At face value, symbolic racism seems to beembedded in traditional American values (e.g., a protestant work ethic or individualism), butits “true” purpose is to preserve racial discrimination. Symbolic racism scholars argue thattraditional American values are illegitimately appropriated to justify racial attitudes aimedat preserving/maintaining racial discrimination. Therefore, White citizens’ have dual racialattitudes—one is formal and conscious and the other is informal and racially prejudiced. Thus,White citizens’ opinions may differ depending on the nature of the questionnaire that assessesthese opinions. For example, on the question measuring overt racism, citizens’ opinions mayseem free from racial prejudice, but their opinions become racially charged when applied toracially loaded social policies, such as Affirmative Action.

Kinder and Sears’s (1981) arguments are quite similar to Katz (1981); Katz and colleagues’arguments (1981, 1988) on inter-attitudinal inconsistency (especially, the vertical inconsistencyapproach). Symbolic racism is characterized by contrasting attitudes that cause ambivalentopinions on racial issues, like the power dynamics between explicit and implicit attitudes.However, while social psychological explanations mainly concern ‘individual’ desires (i.e.,positive public self-presentation), political scientific explanations focus on the “political” con-sequences of ambivalence on public policies (e.g., policy implementation or resistance) or the‘political’ socialization process (e.g., learning of racial prejudice in childhood).

The role of ideology, however, is different in the non-attitude thesis. The main argument ofthe non-attitude thesis is that ideology—frequently referred to as “ideological constraint”—failsto take an integrative role in forming a stable attitude that addresses multiple political issues.The non-attitude thesis has been widely studied in public opinion research since the advent ofopinion polling. Converse (1964) found that most citizens have no stable opinion about socialor policy issues. He argued that political ideology (e.g., Liberalism–Conservatism) does notwork well to unite ordinary citizens’ opinions on multiple issues. According to Converse’s(1964) famous finding,

It posits a very sharp dichotomy within the population according to process of change that arepolar opposites. There is first a “hard core” of opinion on a given issue, which is well crystallizedand perfectly stable over time. For the remainder of the population, response sequences overtime are statistically random (p. 242, emphasis added).

Why do ordinary citizens respond randomly? Although his findings have sometimes servedas telling evidence that the public’s opinion cannot be fully trusted and thus cannot be considered

Page 9: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 617

seriously by policy makers, political scientists explain that the public’s random response is dueto the plurality of political values (Lane, 1962, pp. 364–380; 1973, p. 102). Value—signalingwhat is right or wrong—guides a citizen’s opinion formation (Feldman & Zaller, 1992; Zaller,1992). However, ordinary citizens maintain multiple values simultaneously, but do not paymuch attention to integrating them (Tetlock, 1986). When opinion question items force citizensto report opinions or attitudes, ordinary citizens have to choose an answer after thinking overconflicting values accessible in their minds. The result is that citizens do not report stable,pre-existing, and well-organized opinions, but “manufacture opinions on the spot” (Converse& Presser, 1986, p. 36).

In public opinion research, many empirical studies report citizens’ tendency to claim that theyhave experienced objects that they have not in fact experienced and thus have attitudes towardspurely fictitious issues or organizations (e.g., Bishop, Oldendick, & Tuchfarber, 1982). Becauseopinion expression is activated by interviewing contexts and specific values or ideologies(Schwartz, 1999), the same person may present conflicting or inconsistent opinions towardthe same object, depending on the type of values or ideologies that are activated during theinterview (Feldman & Zaller, 1992). Random response is, therefore, the reflection of conflictingvalues or inconsistent political information ordinary citizens hold (Campbell, Converse, Miller,& Stokes, 1960; Holbrook & Krosnick, 2005).

Although social psychological and political scientific perspectives developed relatively inde-pendently, theories from these fields that explain the emergence of ambivalence is quite similar.Symbolic racism has an isomorphic explanatory structure like vertical inconsistency, but thenon-attitude thesis resembles horizontal inconsistency because the non-attitude thesis does notassume any hierarchical order among multiple attitudes or competing values.

In short, symbolic racism and the non-attitude thesis can be integrated with inter-attitudinalapproaches in a social psychological perspective. The political scientific approach, however,emphasizes macrolevel consequences and more causally distal5 origins to create ambivalencevia a political socialization process. Symbolic racism emerges as a psychological conflictbetween (a) positive attitude formation through formal education, and (b) negative attitudeformation via informal stereotyping of a stigmatized group. This implies that an individual’ssubjective ambivalence is a reflection of different socialization processes. Despite these dif-ferent conceptualizations (i.e., “differentiation” and “integration”; Tetlock, 1986), horizontalinconsistency relating to ideological constraints explains the reasons why a citizen’s lack of cog-nitive effort to unite opinions on multiple issues also originates from a disconnection betweenhis/her private life and the institutional political process (Converse, 1964). In sum, a citizen’sambivalence is begot from his/her lifeworld (Habermas, 1991) and also is the by-product ofsocialization via his/her social relationships (Sherif, 1936).

While the social psychological perspective provides explanation for subjective feelings ofambivalence (i.e., intra-personal conflicts), the political scientific perspective focuses moreon the higher-order generic process whereby intra-personal conflicts spring from the conflictsof the person’s existential base (Lane, 1962, 1973). Thus, this perspective emphasizes theimportance of the existential base and its conflicting values in driving people to experience intra-personal ambivalence. The next sociological perspective tries to explain the social structuralattributes—for example, social networks or identified group memberships—that generate intra-personal ambivalence.

Page 10: An integrative model of ambivalence

618 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

3.3. Sociological perspective: origins of intra-personal ambivalence

Ambivalence studies utilizing the sociological perspective emphasize two components: (1)socialization of values through; (2) interpersonal contacts or social memberships (i.e., net-works). While both components emphasize the group membership(s) that a person holds andare closely related, the two mechanisms explaining the emergence of ambivalence are some-what different because the network approach starts with an emphasis on a person’s objectivesocial condition, whereas the socialization approach focuses on an individual’s subjectiveidentification with the individual’s social relationships.

First, the socialization approach emphasizes how a person becomes accustomed to groupnorms and comes to act like a member of a group. Reference groups are “some identifiablegrouping to which an actor is related in some manner and the norms and values shared inthat group” (Shibutani, 1955, p. 562). Reference groups provide cognitive, affective, and/orbehavioral frames from which an individual can choose. Following Shibutani’s (1955) argu-ments, the plurality of values is the result of a “multiplicity of channels” (p. 562) that providesa “perspective [that] constitutes the frame of reference of the actor” (p. 563). Socializationthus refers to a social psychological process showing that an individual voluntarily accepts thevalues a social group holds (Durkheim, 1897/1997; Turner, 1991). According to Sherif (1936),“the psychological basis of the established social norms” is “the formation of common framesof reference as a product of the contact of individuals” (Sherif, 1936, p. 85), which impliesthat an individual’s perceived social norms are a sum of the values of the reference groups withwhich the individual identifies.

Second, the network-oriented approach emphasizes objective conditions (i.e., the humannetworks an individual maintains), which influence an individual’s behaviors or decision-making. Despite the human tendency to form homophonous relationships, the human networksa person belongs to could be heterogeneous. An extensive social network, for example, mightenhance the chances for a person to be exposed to dissimilar people (Huckfeldt & Sprague,2000; Huckfeldt et al., 2004). Because the different social groups with which an individualis affiliated expect that the individual would fulfill the specific role the group demands, theindividual is more likely to act “in groups” to meet the expectation of their endorsed groups(Lazarsfeld, Berelson, & Gaudet, 1948, p. 137). The cross-pressure hypothesis is the typicalexplanation for why people became ambivalent toward an object under the co-existence ofconflicting expectations from different memberships. Although Lazarsfeld et al. (1948) didnot adopt “ambivalence,” their arguments summarize the social origins of an ambivalent mind.They posit,

By “cross-pressures” we mean the conflicts and inconsistencies among the factors which influ-ence vote decision. (p. 53) . . .. Whatever the source of the conflicting pressures, whether fromsocial status or class identification, from voting traditions or the attitudes of associates, theconsistent result was to delay the voter’s final decision (p. 60, emphasis added).6

The cross-pressure hypothesis was almost forgotten (Horan, 1971), but the recent interestin network-oriented analysis has instigated the revival of the hypothesis (Huckfeldt et al.,2004; Mutz, 2002, 2006). The hypothesis posits that as social beings, people have multiplestatuses conferred from their ascribed and attained groups. Consistent with the mechanism

Page 11: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 619

of the cross-pressure hypothesis, Merton (1976) also defines sociological ambivalence as the“incompatible normative expectation from the unstable social structure composed by multiplerelations” (Merton, 1976, p. 4). As a result, those who have multiple networks or heterogeneoussocial memberships are more likely to be exposed to competing information about an objectand have a higher chance of experiencing different normative pressures.

The emergence of an individual’s subjective ambivalence cannot be fully explained when(1) the network approach only clarifies the potentially contrasting social norms cross-pressuredby contrasting human networks, or (2) socialization is only concerned about the process ofhow an individual recognizes him/herself as a member of a social group. While belonging tocompeting human networks is an objective condition, it does not necessarily cause subjectiveambivalence without the socialization process. On one hand, an individual may maintain con-trasting human networks, which provide potentially contrasting influences on the individual’sevaluation of a social event or object. Despite the existence of contrasting human networks,subjective ambivalence does not emerge unless an individual identifies his/her dual member-ship in contrasting reference groups as forcing the individual to form both positive and negativeevaluations of an event.

For example, let us imagine an adolescent who has friends who like rock-n-roll music butwhose parents dislike it. When it comes to the adolescent’s rock-n-roll preference, his/herfamily network and peer-group network could put potentially contrasting influences on theadolescent’s preference. These contrasting expectations from two human networks may createsubjective ambivalence toward rock-n-roll evaluation in the adolescent’s mind. Accordingto the reference group perspective, the degree of ambivalence can vary depending on theadolescent’s degree of identification with the two human networks. If an adolescent selects‘friends,’ not ‘parents,’ as a reference group in determining his/her preference for rock-n-roll,then the adolescent does not feel ambivalent toward rock-n-roll music, and thus will show apreference for rock-n-roll consistent with his/her peer group.

To summarize, only if a person is forced to make a decision, as in voting in an election(Lazarsfeld et al., 1948), or behaves under perceived group influence (Turner, 1991), socio-logical ambivalence is perceived, and subjective (i.e., intra-personal) ambivalence might beactivated. And only after intra-personal ambivalence is ignited do people become confused orhesitant, and some may delay their decisions or even give up on making decisions. In this sense,sociological ambivalence is latent and should be activated in order to trigger intra-personalambivalence.

4. Confusions in ambivalence studies

Ambivalence scholars have tried to integrate different approaches to study the formationof ambivalence and its consequences. For example, Priester and Petty (2001) examined thepossibility of a theoretical relationship between intra-personal ambivalence and interpersonalconflicts, finding that the perceived gap between individual perception and meta-perception(i.e., an individual’s perception of others’ perceptions of the attitude object) creates higherlevel of intra-personal ambivalence. Under a different perspective, Huckfeldt et al. (2004),Mutz (2002, 2006) and Nir (2005) link sociological origins of ambivalence (i.e., conflicting

Page 12: An integrative model of ambivalence

620 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

social networks) with the emergence of inter-attitudinal inconsistency. In social psychology,Wegener et al. (1995) suggest that intra-personal ambivalence is a sub-component of attitude-strength, arguing that conflicting attitudes toward an object can be reduced into one largerattitude containing two evaluative components.

Despite the proliferation of empirical studies that attempt to link various approaches toambivalence, extant research has not tried to integrate the diverse definitions of ambivalence.This lack of theoretical effort is critical for two reasons. First, it prevents clear and consis-tent integration of existing evidence. For example, the ambivalence-amplification hypothesis(Katz & Hass, 1988) and the ambivalence-moderation hypothesis (Tetlock, 1986) differ-ently predict the role of ambivalence on future behaviors. As mentioned above, the conceptof ambivalence used here is not the same: one type of ambivalence is cognitive and orig-inates from horizontally conflicting attitudes in individual decision-making, whereas theother type is affective and starts from vertically conflicting attitudes in social interaction.Another example of inconsistent findings is the effects of sociological ambivalence on politi-cal participation. Some studies suggest that sociological ambivalence (crosscutting network inMutz, 2002; or network ambivalence in Nir, 2005) decreases political participation, whereasHuckfeldt and Sprague (2000) and Huckfeldt et al. (2004) found that an ambivalent atmo-sphere promotes more active participation. These differences are easily understood whenaccounting for different definitions of ambivalence and measures. While Mutz (2002) andNir (2005) compare a citizen’s self-report of disagreement between his/her political posi-tion and his/her perceived network at the individual level, Huckfeldt and Sprague (2000) andHuckfeldt et al. (2004) used less subjective measures at the group level. The former studiesexplain ambivalence using “perception of disagreement between me and others,” but the lat-ter conceptualized ambivalence as “heterogeneity of a human relationship in which a personlives.” In sum, without identifying the subtle differences between these two qualitatively dis-tinguishable definitions of ambivalence, more studies will only result in more “ambivalentfindings.”

Second, the lack of research precludes the improvement of measures and analytic meth-ods. The popular measure of ambivalence is the so-called “Griffin index” (Kaplan, 1972;Thompson et al., 1995). I do not deny its validity or reliability because the index is a theory-driven formula to conceptualize ambivalence. However, the formula is developed from theintra-attitudinal approach, which assumes specific conditions. One of the main problems inpractical settings is that the formula is only applicable when the strength of two evaluativecomponents is the same or similar (Priester & Petty, 1996, p. 432; also see Thompson etal., 1995, p. 263). As shown in inter-attitudinal studies, it is easy to observe three or moreevaluations of multiple attributes in practical research settings. For example, Mutz (2002)chose only two candidates to calculate citizens’ political ambivalence in the 1992 Ameri-can Presidential Election in which Clinton, Bush, and Perrot competed, and examined theeffect of ambivalence on citizens’ level of political participation. Considering that all threecandidates were prevalent in the 1992 national election, it might be inappropriate to selectthe “top two” candidates and assume the third candidate was an error component with nullinfluence.

I argue that the two problems (i.e., inconsistent findings and lack of relevant ambivalenceindices) can be solved only by theoretical reasoning. Addressing the problem of inconsistent

Page 13: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 621

findings, as I point out in the previous paragraphs, requires a theoretical distinction betweenambivalence and its antecedents. Similarly, the limitations of relevant ambivalence indicesalso creates a need for theoretically driven principles: (1) “intensity,” indicating the strengthof aroused evaluation toward an object, and (2) “similarity” defined as the degree of simul-taneous consideration of positive and negative evaluation toward an object (Thompson et al.,1995).

5. Discussion: an integrative model of ambivalence

As discussed, ambivalence is loosely defined, and its use within each research traditionis not consistent. Although employing the same word, the working definition of ambiva-lence each approach adopts is not unanimous and simple measures seem to be excessivelyexploited without thoughtful theoretical reasoning. Before proposing an integrative model ofambivalence, I will first summarize the scope and diversity of ambivalence studies under thethree outlined perspectives (i.e., social psychological, political scientific, and sociological)(Table 1).

As the summary table implies, the right-side approaches emphasize social, macrolevel, andexternal conflicts that lead to ambivalence, whereas the left-side approaches show microleveland biological human psychological processes and internal experiences of ambivalence towardan object. As I mentioned, these two approaches are sequentially linked to each other. In orderto provide a theoretical integration, I propose a model that consists of four components: (1)the psychological state of ambivalence (intra-personal ambivalence); (2) antecedents causingambivalence (inter-attitudinal conflicts, value conflicts, or competing reference groups); (3)consequences caused by ambivalence (decision-making or behaviors); and (4) moderatingconditions that influence the relationships (asymmetry between positivity and negativity). Fig. 2summarizes the relationship among the different approaches to ambivalence.

The psychological state of ambivalence only happens on an intra-personal level, that is,according to a bivariate conceptualization of attitude. This is the psychological stage imme-diately preceding behavior or decision-making. When people feel ambivalent, they are morelikely to delay decisions or behaviors, or withdraw from the decision-making situation alto-gether (e.g., Campbell et al., 1960; Lane, 1973; Lazarsfeld et al., 1948). This relationship couldbe moderated by human biological tendency (i.e., negativity bias or positivity offset, Cacioppo& Berntson, 1994) and/or conflict-avoidance propensity (i.e., interpersonal discrepancy reduc-tion, Priester & Petty, 2001; face-to-face conflict aversion, Mutz, 2006). For example, healthcampaign studies (for review, see Rothman, Bartels, Wlaschin, & Salovey, 2006) found: whencampaigns target disease prevention (e.g., washing hands), people are more likely to focus onwhat is gained (i.e., positivity offset predominates); when campaigns target disease detection(e.g., mammograms for breast cancer detection), people pay more attention on what would belost (i.e., negativity bias dominates).

When it comes to the antecedents causing ambivalence, a person holds multiple attributestowards an object. As shown in studies of political priming (see Iyengar & Kinder, 1987; Price& Tewksbury, 1997), people’s evaluation of an attitude object (e.g., candidate Obama in the2008 American Presidential Election) varies depending on which attribute (e.g., ‘economy’ or

Page 14: An integrative model of ambivalence

622Y.M

.Baek

/The

SocialScienceJournal47

(2010)609–629

Table 1Summary table for the different approaches to ambivalence.

Perspective Social psychological Political scientific perspective Sociological perspective

Approach Intra-attitudinal approach Inter-attitudinal approach Latent approach

Working definition Co-activation ofboth positivity andnegativity

Co-emergence of conflicting attitudes Co-constraints ofconflicting values

Co-existence of conflicting reference groups

Verticallyconflicting

Horizontally conflicting

Psychologicalprocedures

Evaluation Activation (priming) Differentiationintegration

Identification internalization

Main findings Positivity offsetNegativity bias

Ambivalence-amplificationhypothesisSymbolic racism

Attitude-strengthCognitivecomplexity

Non-attitude thesis Socialization viareference groups

Cross-pressureshypothesisNetworkconstraints

Page 15: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 623

Fig. 2. An integrative model of ambivalence.

‘national defense’) is considered more seriously. In other words, multiple attributes towardsan object antecede a citizen’s bipolar evaluation of the object.

In the inter-attitudinal inconsistency approaches, the popular ambivalence measure (theGriffin index) might not be appropriate because attributes are sometimes multiple (e.g., threeor more evaluations) and different attributes contribute differently to the positive or negativeevaluation (e.g., some items are perceived as more important than others). Here I suggest anextended Griffin index of ambivalence in the situation of multiple attributes. Following Kaplan’s(1972) theorizing, two principles for the ambivalence index can be extracted: (1) intensity and(2) similarity. The intensity can be theorized as an averaged positivity and negativity toward anobject. The similarity, however, is defined as a low value of standard deviation (i.e., variance)following Alvarez and Brehm’s (1997) idea of heteroskedastic probit model.7 Thus, a newindex will be extracted as follows,

Extended Griffin index =(∑n

i=1wiPi

2n+

∑ni=1wiNi

2n

)−

⎛⎝

∑ni=1wi

√(Pi − Ni)2

n

⎞⎠

where wi denotes perceived importance of the issue; P denotes positive evaluation; N meansnegative evaluation; i is individual attitudinal item; and n is the total number of items measured.

Page 16: An integrative model of ambivalence

624 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

The first component is the representation of intensity of evaluation on all attitudinal items,which aims to capture simultaneous activation of positivity and negativity. The second compo-nent is the mathematical representation of similarity, which is the summation of each attitudinalitem’s standard deviation. Larger standard deviation means that a person has a clearly distin-guished evaluation of an object (i.e., dominant reactions, Cacioppo & Berntson, 1994). If aperson has greater intensity and less similarity, then he/she will be more likely to feel intra-personal ambivalence.8 This formula is an extension of the original Griffin index with thespecial case when n = 1 and wi = 1, as shown in the formula.

More distal factors creating intra-personal ambivalence would be supra-individual values(e.g., constraints of ideology or social values), and people’s social networks (e.g., conditionsof human networks and degree of socialization). Theoretically, these may be long-term andcumulative effects. Therefore, while longitudinal panel surveys or participatory observationstudies would provide telling evidence about the importance of those factors (for good exam-ples, see Lane, 1962, 1973), it is rather difficult to test the impact of those supra-individualfactors in short-term empirical studies like experiments.

6. Conclusion

Ambivalence is an interesting human phenomenon that covers and connects biological traitswith social and cultural properties. Here I survey diverse literature dealing with ambivalencein social psychology, political science, and sociology. Although studies in different socialscientific fields investigate ambivalence under different perspectives, they can be integrated viathe suggested model with many implications for social scientists. The “integrative model ofambivalence” links the underlying reasons for ambivalence; the mechanism of how ambivalenceemerges; and the consequences of ambivalence in social contexts and individual decision-making. Based on what is discussed, this study suggests the following:

• Scholars should distinguish subjective ambivalence from its causal factors. As definedin the model, inconsistent attitudes toward an object, conflicting values, or competingreference groups precede subjective ambivalence, and thus should be treated as causalfactors and should not be equated with ambivalence itself.

• Scholars should distinguish the order of antecedents that cause subjective ambivalence.As shown before, ambivalence caused by vertically inconsistent attitudes leads to extremebehavior (amplification effects), whereas ambivalence forced by horizontally inconsistentattitudes causes hesitancy of behavior (moderation effects).

• More research adopting more objective methods should be planned to investigate thelong-term effects of heterogeneous social networks on subjective ambivalence. Insightfulnetwork studies are referenced here, but some of them (e.g., Nir, 2005) have a criticallimitation because measures of social networks are based on self-reported evaluations ofthe respondent’s own social networks, which are therefore contaminated by a respondent’ssubjectivity bias (Marsden, 2005).

• Future research may be directed to explain psychological or behavioral mechanisms forovercoming subjective ambivalence. As far as this study researched, most ambivalencestudies focused on (1) what ambivalence is (i.e., the definition), and (2) why it happens

Page 17: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 625

(i.e., causal mechanisms), but only a few studies (e.g., Katz & Hass, 1988) have examinedhow people overcome subjective ambivalence. In this respect, this study admits that the“Integrative Model of Ambivalence” is also limited.

The integrative model of ambivalence will hopefully provide social researchers with dif-ferentiated conceptualizations of ambivalence emerging from diverse perspectives in socialscientific studies and an integrative way to understand ambivalence in individual and/or socialsituations.

Notes

1. Some readers may be curious about the difference between ‘uncertainty’ and ‘ambigu-ity’ because both psychological states are caused by a lack of information. However,psychologists and/or economists distinguish ambiguity from uncertainty. Ambiguity iscaused by a lack of critical information whereas uncertainty is invited by insufficientinformation to form an evaluation of an event (Keren & Gerritsen, 1999). Thus uncer-tainty is overcome simply by absorbing more information (i.e., information quantity),but ambiguity is averted by acquiring key information (i.e., information type). For furtherdiscussion, see Keren and Gerritsen (1999).

2. There are other approaches to investigating ambivalence. For example, social schema the-ory implicitly defines ambivalence as ‘the inconsistency between schema and a particularevent’ (Fiske, 1998). The most relevant example from social schema theory is subtyping(Allport, 1954). However, this paper does not discuss subtyping in detail because it ismore closely related to inconsistency than ambivalence. For similar reasons, this paperwill not include an examination of theoretical legacy in cognitive dissonance theory(Festinger, 1957). As discussed earlier, ambivalence and dissonance (or inconsistency)should be distinguished conceptually.

3. The vertical inconsistency approach resembles the Freudian psychology model (i.e., themodel of dynamics among the id, ego, and superego). In fact, Katz (1981) traced itstheoretical origin to the Freudian Model.

4. There are other relevant traditions related to “ambivalence” in political science. Forexample, game theorists may explain ambivalence as a game-player’s or social agent’spsychological hesitancy caused by insufficient information about a partner’s or a com-petitor’s attitudes (Raiffa, 1968). However, I do not explore game theoretic implicationsfor two reasons. First, ambivalence in game theory is usually conceptualized by uncer-tainty. In other words, providing more information can resolve the uncertainty aboutthe decision-making situation. As shown earlier in this paper, uncertainty is a conceptdistinct from ambivalence (see Alvarez & Brehm, 1997). Second, game theory assumessymmetry between good and bad values. People usually behave as rational actors (Raiffa,1968), and thus negativity and positivity are assumed to have the same strength. How-ever, people have a psychologically biased tendency towards different reactions—forexample, negativity bias or positivity offset (Cacioppo & Berntson, 1994)—and theirdecision-making process is influenced by subjective risk perception (e.g., risk-taking orloss aversion; see Tversky & Kahneman, 1981).

Page 18: An integrative model of ambivalence

626 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

5. “Distal” means the causal factors having a distant relationship with the final outcome vari-able. For example, in the path model of A → B → C, A is defined as a distal variable. Ina political scientific perspective, ideology is related with inter-attitudinal inconsistency,which causes intra-personal ambivalence.

6. Scholars distinguish two traditions of election studies—the Columbia and MichiganSchools (for review, see Shaw, 2001). While the Columbia School emphasized theinfluence of socio-demographic factors, the Michigan School investigated political psy-chological reasons, such as party affiliation, attitudes toward issues, and affect towardspoliticians (e.g., feeling thermometer). Similar arguments were suggested in the Michi-gan School of election studies, which I include under the political scientific perspective,but their explanation suggests how the sociological approach can be related to the inter-attitudinal approach in either a political scientific or a social psychological perspective.Campbell et al. (1960) posited.

Our measurements of the individual’s evaluations of political objects may lead us todescribe a type of psychological conflict that is linked less closely to politically heteroge-neous memberships than are other types of conflict, such as perceptions of different partisanstandards in two or more primary groups to which the individual belongs. Neverthelessthe hypothesis that social cross-pressures contribute part of the conflict we observe in thesystem of partisan attitudes is one deserving careful exploration. . . .These cross-pressuresshould be reflected in a less consistent field of partisan attitudes (p. 87, emphasis added)utsimply, Campbell et al. (1960) do not deny the plausibility of the cross-pressures hypothe-sis. In Campbell et al’s explanation, the psychological conflict is the mediating concept thatlinks sociological ambivalence to delay in decision-making or political/civic participation.

7. Alvarez and Brehm (1997) treat the inter-attitudinal differences that are not explainedby ideological constraint as “the degree of ambivalence.” Conceptually, the standarddeviation of the second component of the extended Griffin index quantifies the degreeof heteroskedasticity that is inter-attitudinal difference.

8. Another strategy to capture the inter-attitudinal latent ambivalence is the applicationof “multi-trait and multi-method” (MTMM, Campbell & Fiske, 1959) via structuralequation modeling. In other words, overall evaluation (positive vs. negative) can bedefined as trait, and multiple items can be treated as method. In my survey of the literature,there are no applications of MTMM to measure ambivalence.

Acknowledgement

The author would like to thank Dr. Joseph N. Cappella and anonymous reviewers for valuablecomments and suggestions.

References

Abelson, R. P. (1988). Conviction. American Psychologist, 43(4), 267–275.Allport, G. W. (1954). The nature of prejudice. Reading, MA: Addison-Wesley.

Page 19: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 627

Alvarez, R. M., & Brehm, J. (1997). Are Americans ambivalent towards racial policies? American Journal ofPolitical Science, 41(2), 345–374.

Bassili, J. N. (1996). Meta-judgmental versus operative indexes of psychological attributes: The case of measuresof attitude strength. Journal of Personality and Social Psychology, 71(4), 637–653.

Bassili, J. N., & Fletcher, J. F. (1991). Response-time measurement in survey research: A method for CATI and anew look at nonattitudes. Public Opinion Quarterly, 55(3), 331–346.

Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review ofGeneral Psychology, 5(4), 323–370.

Bishop, G. F., Oldendick, R. W., & Tuchfarber, A. J. (1982). Effects of presenting one versus two sides of an issuein survey questions. The Public Opinion Quarterly, 46(1), 69–85.

Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space: A critical review,with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 115(3), 401–423.

Cacioppo, J. T., Petty, R. E., Losch, M. E., & Kim, H. S. (1986). Electromyographic activity over facial muscleregions can differentiate the valence and intensity of affective reactions. Journal of Personality and SocialPsychology, 50(2), 260–268.

Campbell, A., Converse, P. E., Miller, W., & Stokes, D. (1960). The American voter. Chicago, IL: University ofChicago Press.

Campbell, J. E. (1983). Ambiguity in the issue positions of presidential candidates: A causal analysis. AmericanJournal of Political Science, 27(2), 284–293.

Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethodmatrix. Psychological Bulletin, 56(2), 81–105.

Converse, J. M., & Presser, S. (1986). Survey questions: handcrafting the standardized questionnaire. Beverly Hills,CA: Sage Publications.

Converse, P. E. (1964). The nature of belief systems in mass publics. In D. E. Apter (Ed.), Ideology and discontent(pp. 206–261). New York, NY: Free Press.

Dimberg, U., Thunberg, M., & Elmehed, K. (2000). Unconscious facial reactions to emotional facial expressions.Psychological Science, 11(1), 86–89.

Downs, A. (1957). An economic theory of democracy. New York: Harper & Row.Durkheim, E. (1897/1997). The division of labor in society. New York: The Free Press.Eagly, A. H., & Chaiken, S. (1993). The Psychology of Attitudes. Fort Worth, TX: Harcourt Brace.Fazio, R. H. (1995). Attitudes as object-evaluation associations: Determinants, consequences, and correlates of

attitude accessibility. In R. Petty, & J. Krosnick (Eds.), Attitude strength: antecedents and consequences (pp.247–282). Hillsdale, NJ: Erlbaum.

Feldman, S., & Zaller, J. (1992). The political culture of ambivalence: Ideological responses to the welfare state.American Journal of Political Science, 36(1), 268–307.

Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson.Fiske, S. T. (1998). Streotyping, prejudice, and discrimination. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.),

The handbook of social psychology (4th ed., Vol. II, pp. 357–414). New York: The McGraw-Hill Companies,Inc.

Frisch, D., & Baron, J. (1988). Ambiguity and rationality. Journal of Behavioral Decision Making, 1(3),149–157.

Habermas, J. (1991). The structural transformation of the public sphere: an inquiry into a category of bourgeoissociety. Cambridge, MA: MIT Press.

Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.Holbrook, A. L., & Krosnick, J. A. (2005). Meta-psychological vs. operative measures of ambivalence: Differenti-

ating the consequences of perceived intra-psychic conflict and real intra-psychic conflict. In S. C. Craig, & M.D. Martinez (Eds.), Ambivalence and the structure of public opinion. New York: Palgrave Macmillan.

Horan, P. M. (1971). Social positions and political cross-pressures: A re-examination. American SociologicalReview, 36(7), 650–660.

Huckfeldt, R., Mendez, J. M., & Osborn, T. (2004). Disagreement, ambivalence, and engagement: The politicalconsequences of heterogeneous networks. Political Psychology, 25(1), 65–95.

Page 20: An integrative model of ambivalence

628 Y.M. Baek / The Social Science Journal 47 (2010) 609–629

Huckfeldt, R., & Sprague, J. (2000). Political consequences of inconsistency: The accessibility and stability ofabortion attitudes. Political Psychology, 21(1), 57–79.

Iyengar, S., & Kinder, D. R. (1987). News that matters: television and American opinion. Chicago: University ofChicago Press.

Kaplan, K. J. (1972). On the ambivalence-indifference problem in attitude theory and measurement: A suggestedmodification of the semantic differential technique. Psychological Bulletin, 77(5), 361–372.

Katz, I. (1981). Stigma: a social psychological analysis. Hillsdale, NJ: Erlbaum.Katz, I., & Hass, R. G. (1988). Racial ambivalence and American value conflict: Correlational and priming studies

of dual cognitive structures. Journal of Personality and Social Psychology, 55(6), 893–905.Keren, G., & Gerritsen, L. n. E. M. (1999). On the robustness and possible accounts of ambiguity aversion. Acta

Psychologica, 103(1–2), 149–172.Kinder, D. R., & Sears, D. O. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life.

Journal of Personality and Social Psychology, 40(3), 414–431.Lane, R. E. (1962). Political ideology: why the American common man believes what he does. New York: Free

Press.Lane, R. E. (1973). Patterns of political belief. In J. N. Knuntson (Ed.), Handbook of political psychology (pp.

83–116). San Francisco, CA: Jossey-Bass.Larsen, J. T., Norris, C. J., & Cacioppo, J. T. (2003). Effects of positive and negative affect on electromyographic

activity over zygomaticus major and corrugator supercilii. Psychophysiology, 40(5), 776–785.Lavine, H. (2001). The electoral consequences of ambivalence toward presidential candidates. American Journal

of Political Science, 45(4), 915–929.Lazarsfeld, P. F., Berelson, B., & Gaudet, H. (1948). The people’s choice: how the voter makes up his mind in a

presidential campaign. New York: Columbia University Press.Marsden, P. V. (2005). Recent Development in Network Measurement. In P. J. Carrington, J. Scott, & S. Wasserman

(Eds.), Models and methods in social network analysis (pp. 8–30). New York: Cambridge University Press.Merton, R. K. (1976). Sociological ambivalence and other essays. New York: The Free Press.Mutz, D. C. (2002). The consequences of cross-cutting networks for political participation. American Journal of

Political Science, 46(4), 838–855.Mutz, D. C. (2006). Hearing the other side: deliberative versus participatory democracy. New York: Cambridge

University Press.Nir, L. (2005). Ambivalent social networks and their consequences for participation. International Journal of Public

Opinion Research, 17(4), 422–442.Peeters, G., & Czapinski, J. (1990). Positive-negative asymmetry in evaluations: The distinction between affective

and informational effects. In W. S. M. Hewstone (Ed.), European review of social psychology (pp. 33–60). NewYork: Wiley.

Price, V., & Tewksbury, D. (1997). News values and public opinion: A theoretical account of media priming andframing. In G. Barnet, & F. J. Border (Eds.), Progresses in communication sciences (pp. 173–212). Greenwich,CT: Ablex.

Priester, J. R., & Petty, R. E. (1996). The gradual threshold model of ambivalence: Relating the positive andnegative bases of attitudes to subjective ambivalence. Journal of Personality and Social Psychology, 71(3),431–449.

Priester, J. R., & Petty, R. E. (2001). Extending the bases of subjective attitudinal ambivalence: Interpersonal andintrapersonal antecedents of evaluative tension. Journal of Personality and Social Psychology, 80(1), 19–34.

Raiffa, H. (1968). Decision analysis: introductory lectures on choices under uncertainty. Reading. MA: Addison-Wesley.

Rothman, A. J., Bartels, R. D., Wlaschin, J., & Salovey, P. (2006). The strategic use of gain-and loss-framedmessages to promote healthy behavior: How theory can inform practice. Journal of Communication, 56(s1),S202–S220.

Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion. Personality & SocialPsychology Review, 5(4), 296–320.

Schuman, H., & Presser, S. (1980). Public opinion and public ignorance: The fine line between attitudes andnonattitudes. The American Journal of Sociology, 85(5), 1214–1225.

Page 21: An integrative model of ambivalence

Y.M. Baek / The Social Science Journal 47 (2010) 609–629 629

Schwartz, N. (1999). Self-report: How the questions shape the answers. American Psychologist, 54(2), 93–105.Shaw, D. (2001). Communicating and electing. In R. P. Hart, & D. Shaw (Eds.), Communication in U.S. elections:

new agendas (pp. 1–17). Oxford: Rowman & Littlefield Publishers.Sherif, M. (1936). The psychology of social norms. New York: Harper & Brothers.Shibutani, T. (1955). Reference groups as perspectives. American Journal of Sociology, 60(6), 562–569.Steiner, J. E., Glaser, D., Hawilo, M. E., & Berridge, K. C. (2001). Comparative expression of hedonic impact:

Affective reactions to taste by human infants and other primates. Neuroscience & Biobehavioral Reviews, 25(1),53–74.

Tetlock, P. E. (1986). A value pluralism model of ideological reasoning. Journal of Personality and Social Psychol-ogy, 50(4), 819–827.

Thompson, M., Zanna, M. P., & Griffin, D. W. (1995). Let’s Not Be Indifferent about (Attitudinal) Ambivalence. InR. E. Petty, & J. A. Krosnick (Eds.), Attitude strength: antecedents and consequences (pp. 361–386). Mahwah,NJ: Lawrence Erlbaum Associates.

Thurstone, L. L. (1928). Attitudes can be measured. The American Journal of Sociology, 33(4), 529–554.Turner, J. C. (1991). Social influence. Pacific Grove, CA: Brooks/Cole.Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481),

453–458.Wegener, D. T., Downing, J., Krosnick, J. A., & Petty, R. E. (1995). Measures and manipulations of strength-related

properties of attitudes: current practice and future directions. In R. E. Petty, & J. A. Krosnick (Eds.), Attitudestrength: Antecedents and consequences. Mahwah, NJ: Erlbaum, pp. 455–487.

Wilson, T. D., Lindsey, S., & Schooler, T. Y. (2000). A model of dual attitudes. Psychological Review, 107(1),101–126.

Zaller, J. R. (1992). The nature and origins of mass opinion. New York: Cambridge University Press.