Rabbie, J. M., Schot, J. C., & Visser, L. (1989). Social identity theory:A conceptual and empirical critique from the perspective of abehavioral interaction model.European Journal of Social Psy-chology, 19,171–202.

Rescher, N. (1969).Many-valued logic.New York: McGraw-Hill.Rosenberg, M. J. (1956). Cognitive structure and attitudinal af-

fect. Journal of Abnormal and Social Psychology, 53,367–372.

Rosenberg, M. J. (1960). An analysis of affective-cognitive consis-tency. In C. I. Hovland (Ed.),Attitude organization and change(pp. 15–64). New Haven, CT: Yale University Press.

Runkel, P. J., & Peizer, D. B. (1968). The two-valued orientationof current equilibrium theory.Behavioral Science, 13,56–65.

Schlenker, B. R. (1982). Translating actions in attitudes: An iden-tity-analytic approach to the explanation of social conduct. In L.Berkowitz (Ed.),Advances in experimental social psychology(Vol. 15, pp. 194–247). New York: Academic.

Sedikides, C., & Skowronski, J. J. (1997). The symbolic self in evolu-tionary context.Psychological Inquiry, 1, 80–102.

Steele, C. M. (1988). The psychology of self-affirmation: Sustainingthe integrity of the self. In L. Berkowitz (Ed.),Advances in ex-perimental social psychology(Vol. 21, pp. 261–302). SanDiego, CA: Academic.

Steele, C. M., & Liu, T. J. (1983). Dissonance processes asself-affirmation.Journal of Personality and Social Psychology,45,5–19.

Tajfel, H. (1970). Experiments in intergroup discrimination.Scien-tific American, 223(5), 96–102.

Tajfel,H. (1981).Humangroupsandsocial categories:Studies insocialpsychology.Cambridge, England: Cambridge University Press.

Tashakkori, A., & Insko, C. A. (1979). Interpersonal attraction andthe polarity of similar attitudes: A test of three balance models.Journal of Personality and Social Psychology, 37,2262–2277.

Tashakkori, A., & Insko, C. A. (1981). Interpersonal attraction andperson perception: Two tests of three balance models.Journalof Experimental Social Psychology, 17,266–285.

Thurstone, L. L. (1928). Attitudes can be measured.American Jour-nal of Sociology, 33,529–544.

Tice, D. M. (1993). The social motivations of people with lowself-esteem. In R. F. Baumeister (Ed.),Self-esteem: The puzzleof low self-regard(pp. 37–53). New York: Plenum.

Wellens, A. R., & Thistlethwaite, D. L. (1971a). An analysis of twoquantitative theories of cognitive balance.Psychological Re-view, 78,141–150.

Wellens, A. R., & Thistlethwaite, D. L. (1971b). Comparison of threetheories of cognitive balance.Journal of Personality and SocialPsychology, 20,89–92.

Wiest, W. M. (1965). A quantitative extension of Heider’s theory ofcognitive balance applied to interpersonal perception andself-esteem.Psychological Monographs, 79(14, Whole No. 607).

Wyer, R. S. (1974).Cognitive organization and change: An informa-tion processing approach.Hillsdale, NJ: Lawrence ErlbaumAssociates, Inc.

Applying the Unimodel to Political Persuasion

Peter KerkhofDepartment of Social Psychology

Vrije Universiteit, Amsterdam

Kruglanski and Thompson’s unimodel of persua-sion provides a theoretical and empirical challenge forboth dual-route theorists and for the numerous scholarsapplying the Elaboration Likelihood Model (ELM;Petty & Cacioppo, 1986) and the Heuristic SystematicModel (HSM; Chaiken, 1980) to everyday phenomenalike advertising and political campaigning: Why usetwo routes of persuasion when, as Kruglanski andThompson claim, one is sufficient? In this article, Ifirst examine why dual-route models have had suchgreat appeal to those studying the dynamics of politicalpersuasion. I then contrast different predictions de-rived from dual-route models and the unimodel. Spe-cifically, I examine how the two models explainpersuasion in a situation where cue information ismore readily available than message arguments, a situ-ation that is quite common in the political context.

Political Persuasion

One of the core tasks of politicians is to persuade.Whether they aim their efforts to persuade at col-

leagues or at citizens, politicians continuously try toconvince others that their party has the best solutionsfor the country’s problems and the best people to im-plement those solutions. The question of how peoplecome to accept or reject these persuasive messages hastraditionally attracted a lot of attention from both so-cial psychologists and political scientists. Dual-routemodels like the ELM and HSM have been helpful inunderstanding political persuasion. Kruglanski andThompson’s attempt to replace the two routes by a sin-gle one is therefore a daring one.

One reason dual-route models have been helpful inexplaining political persuasion is the distinction that isusually made between politically aware and politicallyunaware citizens. In a political utopia, people arehighly politically aware: They hold political attitudesthat are relatively stable over time, that change as a re-sult of new and relevant message arguments, that showa certain consistency, and that are predictive of differ-ent kinds of political behavior (e.g., voting or takingpart in political protest). Before going to the electionbooth, these ideal citizens all read the party programs,they watch the political debates on television, read all

137

COMMENTARIES

Copyright © 2000 All Rights Reserved

the op-ed pages, and discuss their findings extensivelywith friends, family, and colleagues. In real life, how-ever, most citizens are not that politically aware: Theirpolitical attitudes are not that consistent and theychange over time in a seemingly random way (Con-verse, 1964). Voting preferences seem to depend oncandidates’ looks rather than on their political pro-grams. Citizens lack even the most basic factual politi-cal knowledge (see Page & Shapiro, 1992, pp. 10–11,for examples of the lack of political knowledge amongU.S. citizens).

One of the key questions in political persuasion re-search is how this large group of politically unawarecitizens differs from the politically aware in the waythey come to their political preferences. Both the ELMand HSM have shown large heuristic value in answer-ing this question. The cognitive and motivational de-terminants of the two ELM (or HSM) routes (e.g., lowvs. high need for cognition, low vs. high involvement)mirror many of the distinctions made in political psy-chology to separate the politically unaware from thepolitically aware: uneducated versus educated (e.g.,Phelan, Link, Stueve, & Moore, 1995), novice versusexpert (e.g., Fiske, Kinder, & Larter, 1983), and unin-volved versus involved (e.g., Judd, Krosnick, &Milburn, 1981).

Politically unaware citizens are usually comparedwith the students in the low-involvement condition inPetty and Cacioppo’s classical comprehensive examexperiment. As Zaller (1992) put it, “most politics, atleast in the contemporary United States, is notoriouslylow key and uninvolving. The stakes are high, but peo-ple find it hard to stay interested” (p. 47). Under suchconditions, one cannot expect citizens to make the ef-fort of processing all the message arguments during apolitical campaign. Why bother to examine the prosand cons of NATO expansion or higher minimumwages if you can judge a candidate by their looks, elo-quence, or the party they represent?

Politically aware citizens are typically compared tothe students in the high-involvement conditions in thePetty and Cacioppo experiments. McGraw and Hub-bard (1996) referred to Petty and Cacioppo (1986) andstated that the politically aware, whom they calledso-phisticates,“are both motivated and able to carefullyscrutinize a politician’s explanation” (p. 150). Part ofthe appeal of dual-route models thus seems to lie in thefact that the large body of research concerning theELM and HSM can be used to hypothesize about dif-ferences between the politically aware and unaware.Dual-route models make clear predictions about theway that people, given their different levels of involve-ment and knowledge, deal with political persuasive in-formation.

Apart from this heuristic value of dual-route mod-els, there is another reason for students of political per-suasion to resist substituting the two routes for one

single route. To many political theorists there is a qual-itative difference between political preferences basedon processing peripheral cues (e.g., the looks of thecandidate) versus preferences based on political mes-sage arguments. Specifically, they contend thatcue-based preferences are more prone to misleadingpersuasive attempts than argument-based preferences.This may not be a principled difference in the persua-sion process as such, as Kruglanski and Thompson ar-gue. It does, however, constitute a principleddifference in the way people form political prefer-ences. In this perspective, the (normative) aim has al-ways been to make the group of politically awarecitizens as large as possible: Democracy would func-tion in a better way if its citizens paid more attention tomessage arguments than to political cue information.The assumption is that argument-based preferenceswould make citizens less vulnerable to easy solutionsoffered by populist politicians.

Given the aforementioned reasons for applyingdual-route models to political persuasion, is there areason for applying the unimodel to political persua-sion? If both models were to make the same predic-tions about how the politically aware and unaware getpersuaded by political messages, I would argue thatthere is no compelling reason to prefer one model overthe other. However, Kruglanski and Thompson’s theo-rizing leads them to predict that, under certain condi-tions, dual-route models make predictions that arequite different from those made by the unimodel.These conditions, I argue in the remainder of this com-mentary, are very common in the context of politicalpersuasion and therefore highly relevant for studyingthe effects of political communication.

How to Persuade the PoliticallyAware?

The core of Kruglanski and Thompson’s argumentis that “once differences on persuasively relevantin-formational parametersare controlled for, cue-basedand message argument-based persuasion should be im-pacted similarly by various persuasively relevantpro-cessing variables.” Kruglanski and Thompson point tothe length, complexity, and ordinal position of the cuesand message arguments as possible causes of the ef-fects that are established within the ELM and HSMframework.

In their first study, Kruglanski and Thompsonshowed that when cue information is lengthy and com-plex, low-involvement participants are not influencedby cue information, whereas high-involvement partici-pants are. These conditions, namely cue informationthat is more lengthy and complex than message argu-ments, are quite common in the political informationthat people receive from mass media. Kruglanski and

138

COMMENTARIES

Copyright © 2000 All Rights Reserved

Thompson’s results thereby question the idea that thepolitically aware, typically compared tohigh-involvement participants, are persuaded by argu-ments rather than cues. Thus, in the context of lengthyand complex cue information, the unimodel predictsthat cue information influences the involved more thanthe uninvolved, whereas the dual-route models wouldpredict the opposite (McGraw & Hubbard, 1996; Petty& Cacioppo, 1986).

In mass media messages about politics, relativelymore attention is devoted to information about the po-litical sources and their relations with each other thanto information about the issues. Kleinnijenhuis,Oegema, De Ridder, and Bos (1995) conducted a de-tailed content analysis of the news during the cam-paigning for Dutch parliamentary elections in 1994.Their results showed that about one third of the newscontained issue information (i.e., message arguments),information about the different parties’ stands on theissues that are at stake. The remaining two thirds of in-formation concerned the success a political party hadduring the campaign or its relationships with other par-ties (i.e., cue information).

Another example of the dominance of cue informa-tion in political news can be found in a study conductedby Klandermans and Goslinga (1996). These authorsstudied the news media content after a highly contro-versial measure proposed by the Dutch cabinet. In1991, the Dutch government proposed a drastic cut inthe level of disability benefits, resulting in a period oflarge-scale protest. The authors’ analysis of the massmedia reporting about the conflict led them to con-clude that “It is clear who the actors are … it is muchless clear what exactly the controversy is about” (p.322). Put differently, in a naturalistic setting messagearguments again appeared to be less salient than cue in-formation.

Thus, in political news, cue information prevails.This is an intriguing finding because this is exactly thecondition in which the dual-route models make predic-tions that are different from the predictions made bythe unimodel. Within the dual-route models it is com-monly assumed that the politically aware judge a poli-tician according to his or her issue positions, whereasthe politically unaware use all sorts of judgmentalheuristics (e.g., Snidermann, Brody, & Tetlock, 1991).The question is whether this hypothesis holds with theunimodel in mind. Are the politically aware indeedmainly persuaded by the strength of arguments? Thedual-route models would predict that this is the casegiven the higher motivation and ability of the politi-cally aware. The unimodel, however, predicts that thiseffect depends on variations in the information of cuesand message arguments: If cue information is domi-nant, then the politically aware’s acceptance or rejec-tion of the political message would be the result of theextensive processing of cue information. All this de-

pends on the structure of the information that is pro-vided. Whereas dual-route models attribute effects ofpersuasion to different processing modes (central orperipheral), the unimodel attributes the effects of per-suasion to information characteristics (e.g., length orcomplexity).

Elite Influence on Public Opinion:An Example From the Gulf War

One of the more consistent findings in political per-suasion research is that when the political elite (e.g.,politicians, scientists, journalists) consistently, andover a long period of time, stresses one point of viewover another, the politically aware tend to follow theelites’ point of view (the mainstream effect; Zaller,1987, 1992). When, for example, both Democrat andRepublican U.S. politicians favored sending troops tothe Persian Gulf in the period just after the Iraqi inva-sion of Kuwait, the politically aware, both Democratsand Republicans, were the first to accept this point ofview (Zaller, 1994). The question arises whether thepolitically aware came to support the elite point ofview because of the presented message arguments orbecause of the partisan cues that were provided.Whereas dual-route models would predict arguments,whether or not processed in a biased way, to be the vi-tal factor here, the Unimodel would predict that politi-cal involvement leads people to process the cues moreextensively. Cues in such a situation are powerful: Al-most everybody that matters in Washington agreedthat troops should be sent.

Departing from dual-route models, Zaller (1992)explained the mainstream effect by stating that all citi-zens should be regarded as the low-involvement par-ticipants in the ELM or HSM experiments. He thusquestioned the generalizability of the ELM and HSMfindings concerning the highly involved to the highlypolitically aware. In his view, the only difference be-tween the politically aware and the politically unawareis the amount of exposure to political information andthe ease with which the politically aware can put politi-cal (mostly partisan) cues to use. The unimodel wouldstate that whether the politically aware are persuadedmore by arguments than by cues depends on the struc-ture of the information that is provided. Both argu-ments and cues can be processed extensively and bothcan lead the politically aware to change their opinion.

The Gulf War example clearly shows the value ofthe unimodel in explaining political persuasion. Theunimodel can explain the mainstream effect withoutmaking the additional assumption, like Zaller does,that everybody is relatively uninvolved. Similarly, itcan explain the mainstream effect without assumingthat the politically aware are convinced by message

139

COMMENTARIES

Copyright © 2000 All Rights Reserved

arguments mainly, like the dual-route models wouldargue.

Concluding Remarks

Political persuasion research is often conducted innaturalistic settings (e.g., mass surveys). These set-tings make it difficult to control for the type of infor-mation that people have encountered in the way that istypical for Kruglanski and Thompson’s studies. Re-search in naturalistic settings often has to rely on as-sumptions about the information that people encounterand about the way they process this information. Theseassumptions are often based on dual-route models likethe ELM and HSM. The implicit assumption in theELM is that cue information is short and simple,whereas arguments are lengthy and complex. Al-though one would wish this to be the case in politicalnews, the practice of newsgathering makes such astructure of news very unlikely. Extensive argumenta-tion about political issues is usually confined to thosepages that very few people read.

Kruglanski and Thompson’s unimodel should leadscholars of political persuasion to take another lookat the assumptions they hold concerning the structureof political information and the way the politicallyaware and the politically unaware use this informa-tion to come to a judgment. This could lead to hy-potheses that are the opposite of the hypotheses de-rived from dual-route models. For example, undercertain conditions, message arguments could lead thepolitically aware to change their opinions, whereasthe politically aware are influenced by extensive cueinformation.

It remains to be seen whether the unimodel can ac-count for several well-established differences betweenthe attitudes of the politically aware and unaware. Thestudies conducted by Kruglanski and Thompson didnot consider the consequences of processing cues ver-sus message arguments under conditions of high in-volvement. For example, does cue processing underhigh involvement lead to the stable attitudes that arecharacteristic of highly politically aware citizens? Arethese attitudes as consequential for behavior as atti-tudes resulting from extensive arguments processing?These are some questions still to be answered by theunimodel.

Note

Peter Kerkhof, Department of Social Psychology,Vrije Universiteit Amsterdam, Van derBoechorststraat 1, 1081 BT Amsterdam, The Nether-lands. E-mail: P.Kerkhof@ psy.vu.nl

References

Chaiken, S. (1980). Heuristic versus systematic information process-ing and the use of source versus message cues in persuasion.Journal of Personality and Social Psychology, 39,752–766.

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: Free Press.

Fiske, S. T., Kinder, D. R., & Larter, W. M. (1983). The novice andthe expert: Knowledge-based strategies in political cognition.Journal of Experimental Social Psychology, 19,381–400.

Judd, C. M., Krosnick, J. A., & Milburn, M. (1981). Political involve-ment, and attitude structure in the general public.American So-ciological Review, 46,660–669.

Klandermans, B., & Goslinga, S. (1996). Media discourse, move-ment publicity and the generation of collective action frames:Theoretical and empirical exercises in meaning construction. InD. McAdam, J. McCarthy, & M. Zald (Eds.),Opportunities,mobilizing structures and frames: Comparative applications ofcontemporary movement theory(pp. 312–337). Cambridge,England: Cambridge University Press.

Kleinnijenhuis, J., Oegema, D., De Ridder, J., & Bos, H. (1995).Dedemocratie op drift: Een evaluatie van de verkiezingscampagnevan 1994[Drifting democracy: An evaluation of the election cam-paign of 1994]. Amsterdam, The Netherlands: VU Uitgeverij.

McGraw, K. M., & Hubbard, C. (1996). Some of the people some ofthe time: Individual differences in acceptance of political ac-counts. In D. C. Mutz, P. M. Sniderman, & R. A. Brody (Eds.),Political persuasion and attitude change(pp. 145–170). AnnArbor: University of Michigan Press.

Page, B., & Shapiro, R. (1992).The rational public: Fifty years oftrends in American’s policy preferences. Chicago: University ofChicago Press.

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood modelof persuasion. In L. Berkowitz (Ed.),Advances in experimental so-cial psychology(Vol. 19, pp. 123–205). New York: Academic.

Phelan, J., Link, B. G., Stueve, A., & Moore, R. E. (1995). Education,social liberalism, and economic conservatism: Attitudes towardhomeless people.American Sociological Review, 60,126–140.

Sniderman, P. M., Brody, R. A., & Tetlock, P. E. (1991).Reasoningand choice: Explorations in political psychology.Cambridge,England: Cambridge University Press.

Zaller, J. (1987). The diffusion of political attitudes.Journal of Per-sonality and Social Psychology, 53,821–833.

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

Zaller, J. (1994). Elite leadership of mass opinion. New evidence fromthe Gulf War. In W. L. Bennett & D. L. Paletz (Eds.),Taken bystorm. The media, public opinion, and U.S. foreign policy in theGulf War(pp. 186–209). Chicago: University of Chicago Press.

140

COMMENTARIES

Copyright © 2000 All Rights Reserved