selective media exposure and partisan differences about sarah palin's candidacy

27
Selective Media Exposure and Partisan Differences about Sarah Palin’s Candidacy DAVID A. JONES James Madison University KATHLEEN FERRAIOLO James Madison University JENNIFER BYRNE James Madison University In today’s fragmented media environment, citizens have an enhanced ability to select information outlets that match their preexisting political beliefs and avoid information sources that clash. This modern version of “selective exposure” means that citizens can more easily isolate themselves from perspectives unlike their own. In this article, we employed a quasi-experimental design to test (1) the extent to which subjects selectively exposed themselves to opinion articles that were congenial to their presumed predispositions about Sarah Palin and (2) whether selective exposure reinforced these predispositions, thereby magnifying differences between Republicans and Democrats. We found that subjects did indeed tend to engage in selective exposure. And among those who did, there was evidence of reinforcement. Interestingly, there was evidence of persuasion among the relatively few subjects who exposed themselves to sources that ran counter to their predispositions. Keywords: Political Communications, New Media, Media Effects, Selective Exposure. En el ambiente fragmentado de los medios de comunicación actuales los ciudadanos tienen una mayor habilidad de seleccionar medios de comunicación que sean compatibles con sus creencias políticas preexistentes y evitar fuentes de información incompatibles. Esta versión moderna de la “exposición selectiva” permite que los ciudadanos puedan aislarse más fácilmente de perspectivas contrarias a la propia. En este artículo, empleamos un diseño cuasi-experimental para Acknowledgements: We would like to thank our colleague Charles H. Blake and the four anonymous reviewers for their important suggestions and constructive criticism. Politics & Policy, Volume 39, No. 2 (2011): 195-221. Published by Wiley Periodicals, Inc. © The Policy Studies Organization. All rights reserved.

Upload: david-a-jones

Post on 21-Jul-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

polp_288 195..222

Selective Media Exposure and PartisanDifferences about Sarah Palin’s Candidacy

DAVID A. JONESJames Madison University

KATHLEEN FERRAIOLOJames Madison University

JENNIFER BYRNEJames Madison University

In today’s fragmented media environment, citizens have an enhancedability to select information outlets that match their preexistingpolitical beliefs and avoid information sources that clash. This modernversion of “selective exposure” means that citizens can more easilyisolate themselves from perspectives unlike their own. In this article, weemployed a quasi-experimental design to test (1) the extent to whichsubjects selectively exposed themselves to opinion articles that werecongenial to their presumed predispositions about Sarah Palin and (2)whether selective exposure reinforced these predispositions, therebymagnifying differences between Republicans and Democrats. We foundthat subjects did indeed tend to engage in selective exposure. And amongthose who did, there was evidence of reinforcement. Interestingly, therewas evidence of persuasion among the relatively few subjects whoexposed themselves to sources that ran counter to their predispositions.

Keywords: Political Communications, New Media, Media Effects,Selective Exposure.

En el ambiente fragmentado de los medios de comunicación actualeslos ciudadanos tienen una mayor habilidad de seleccionar medios decomunicación que sean compatibles con sus creencias políticaspreexistentes y evitar fuentes de información incompatibles. Estaversión moderna de la “exposición selectiva” permite que los ciudadanospuedan aislarse más fácilmente de perspectivas contrarias a la propia.En este artículo, empleamos un diseño cuasi-experimental para

Acknowledgements: We would like to thank our colleague Charles H. Blake and the fouranonymous reviewers for their important suggestions and constructive criticism.

Politics & Policy, Volume 39, No. 2 (2011): 195-221. Published by Wiley Periodicals, Inc.© The Policy Studies Organization. All rights reserved.

examinar: (1) la medida en la que los sujetos selectivamente se exponena artículos de opinión que congenien con sus predisposiciones acerca deSarah Palin y (2) si la exposición selectiva refuerza esta opinión, por lotanto magnificando las diferencias entre Republicanos y Demócratas.Encontramos que los sujetos efectivamente tienden a emplear unaexposición selectiva. Y entre aquellos que lo hacen, hay evidencia delreforzamiento de dicha opinión. Curiosamente, hubo evidencia depersuasión entre los sujetos que se expusieron a fuentes contrarias a suideología.

Media fragmentation brings good news and bad news. The good news isthat the media diet for most Americans is more varied and interesting thanit used to be. Americans now have access to a wide array of politicalcommunication outlets offering a comprehensive menu of perspectives andviewpoints. But as is often the case, the bad news drowns out the good. Bysome accounts, the expansion of “media choice” poses challenges to ademocracy that thrives when its citizens are well-informed about and engagedin public affairs (Prior 2007). Among the concerns is the enhanced ability ofcitizens to select media outlets that match their preexisting political beliefs andavoid information sources that clash. This partisan version of “selectiveexposure” means that citizens can more easily isolate themselves fromperspectives unlike their own. Many individuals might enclose themselves in asort of partisan media echo chamber, perhaps resulting in an enhancement oftheir opinions over time. If this is the case, the current fragmented mediaenvironment might well be contributing to opinion polarization in the U.S.electorate (Sunstein 2007).

In this study, we examined partisan selective exposure and its effects.Using a quasi-experimental research design, we analyzed the extent to whichsubjects selectively exposed themselves to information that supported theirpreexisting attitudes about Sarah Palin, the Republican vice presidentialnominee in the 2008 election. We then assessed the extent to which theopinions of subjects who chose a congenial source underwent some sort ofsystematic change. In short, we found that many subjects did engage inselective exposure when informing themselves about Sarah Palin. And whenthey did, their predispositions tended to be reinforced. When they did not—that is, when they sought out a perspective that ran counter to their partisantendencies—there was evidence of persuasion. We build on the existingliterature by examining whether selective exposure is occurring when a newissue or topic is introduced to the political scene. Moreover, we do this byusing real candidates, discussed in real articles, in real time to draw ourconclusions rather than a hypothetical scenario. Finally, we offer evidencethat selective exposure may be more likely to occur when an issue is new andfresh, and argue that timing is an important issue to consider in future workregarding selective exposure effects.

196 | POLITICS & POLICY / April 2011

Selective Exposure

Selective exposure1 is hardly a new concept in political communicationresearch. It is a key element of cognitive dissonance, whereby people seek meansto rationalize their beliefs by avoiding information that clashes with theirexisting views (Festinger 1957). The concept is also related to the “confirmatorybiases” that people exhibit when they misinterpret ambiguous information inways that affirm their expectations (Rabin and Schrag 1999; Taber and Lodge2006). For researchers in the “minimal effects” tradition, selective exposurehelps explain why mass media outlets once seemed to do little more thanreinforce people’s preexisting beliefs (Klapper 1960). At the time when much ofthis early communication research was conducted—the 1940s and 1950s—television was in its infancy, and voters still relied more heavily on interpersonalcommunication and party-supplied campaign materials for information aboutcandidates running for office. To the extent that people turned to the media(radio and newspapers) for information, their “predispositions [came] intoplay,” making them “more likely to tune in some programs than others”(Lazarsfeld, Berelson, and Gaudet 1968, 76). This meant that “exposure wasconsistently partisan, and such partisan exposure resulted in reinforcement.”During election campaigns, voters seemed to gravitate toward appeals from thecandidate or party they preferred (Lazarsfeld, Berelson, and Gaudet 1968, 89).However, subsequent research on media effects offered mixed results, with somestudies finding evidence that people selectively expose themselves to informationabout candidates they support (see, e.g., Chaffee and McLeod 1973); othersfinding mixed results (Chaffee and Miyo 1983); and others casting doubt on theextent to which people avoid dissonant messages when acquiring politicalinformation (Frey 1986; Sears and Freedman 1967). Still others found norelationship between political predispositions and their information sources(Feather 1962; Freedman 1965). In a critical essay on the subject, Sears andFreedman (1967) concluded that the evidence for selective exposure was weak.

Yet, recent changes in the media environment have rekindled interest in theconcept. Until recently, it was more difficult for people to limit themselves tocongenial media outlets.

Fifty years ago, U.S. voters depended primarily on the evening newscastsbroadcast by ABC, CBS, and NBC to keep abreast of the world of publicaffairs. The norms of journalism meant that no matter which networkvoters tuned in to, they encountered the same set of news reports,

1 Other researchers prefer the term “selective attention,” but they seem to be describing the sameconcept. Graf and Aday (2008, 87), for example, define selective attention as “the preference forinformation that is consistent with previously held beliefs together with the avoidance ofinformation counter to those beliefs.” This echoes our understanding of selective exposure, a termin wider use even today (see Bennett and Iyengar 2008; Iyengar and Hahn 2009; Iyengar et al.2008; Stroud 2006).

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 197

according balanced attention to parties, candidates, or points of view. Inthe era of ‘old media’ accordingly, it made little difference where votersgot their news. The offerings of all news organizations were sufficientlyhomogeneous and standardized to represent an ‘information commons.’Americans of all walks of life and political inclination were exposed to thesame information. (Bennett and Iyengar 2008, 717)

Today, however, people may choose from a wide array of political mediaoutlets, some of them outwardly partisan. Talk radio programs have attracteda predominately conservative audience since at least the 1980s (Barker 2002).Left-leaning perspectives seem to prevail in the blogosphere (Sunstein 2007).Although the Fox News Channel claims to be “Fair and Balanced,” it “hasestablished an unmistakable identity: it is opinionated and conservative, and itsnews is delivered by people who themselves are often unabashedly opinionatedand conservative” (Auletta 2003, 60). Not surprisingly, regular Fox NewsChannel viewers tend to lean to the right and vote accordingly (DellaVigna andKaplan 2007; Jones 2006). According to one study drawing upon data from themid- to late-1990s, the entry of Fox News Channel into a cable market resultedin an increased tendency for voters in that market to support Republicancandidates (DellaVigna and Kaplan 2007). Meanwhile, many liberals andDemocrats seem to be gravitating toward MSNBC, which is using its prime-time talk shows to cater to left-leaning audiences (Spruiell 2008). These shiftsmay represent a deliberate effort by news outlets to carve out definable audienceniches in a fiercely competitive media market (Gentzkow and Shapiro 2006;Mullainathan and Shleifer 2005).

This new fragmented media environment raises a number of importantconcerns. For one, it is easier for some citizens to tune out politics altogether.On broadcast outlets, large chunks of the programming day—particularly theearly evening—are dedicated to news. This meant that in the 1960s and 1970s,before cable was king, most Americans who wanted to watch television in theearly evening had almost no choice but to watch the news. The nightly news wasthe only show on the major networks during that time; “choice” constitutedeither selecting between one of the three network news programs, syndicatedprogramming on a local independent channel, or turning off the television. Evenunmotivated viewers gained substantive political information via television—even if it was inadvertent. Today, however, citizens who are less engaged inpublic affairs are more likely to opt for entertainment programming (Prior2007). Whereas less engaged citizens were once compelled to experience “politicsby default” (Neuman 1996), they now may completely avoid news about publicaffairs. This is what some call the “demise of the inadvertent audience,” whichmeans that “people uninterested in politics can avoid news programmingaltogether by tuning into ESPN or the Food Network” (Bennett and Iyengar2008, 717). As a result, there are wider gaps in knowledge and participationbetween motivated and nonmotivated citizens (Prior 2007).

198 | POLITICS & POLICY / April 2011

The concern of this article is that media fragmentation also makes it easierfor people to engage in selective exposure when informing themselves aboutpublic affairs (Iyengar et al. 2008; Stroud 2007). Although at least one studysuggests that the phenomenon may be overstated (Gentzkow and Shapiro2010a), others find evidence of selective exposure in the new media era. In astudy of Fahrenheit 9/11, the Michael Moore documentary that satirized theBush Administration’s War on Terror, people’s ideology and their attitudestoward Bush predicted their intent to watch the movie (Stroud 2007). Researchon website use during campaigns suggests that political views helped explainwhich candidate websites people accessed online (Bimber and Davis 2003). Evenwith traditional media outlets, such as daily newspapers, readers seem togravitate toward newspapers slanted toward their own ideology (Gentzkow andShapiro 2010b), although the biases in news content may be more related to thepartisanship of editorial-page endorsements than the orientations of readers(Larcinese, Puglisi, and Snyder 2007; Puglisi and Snyder 2008). A study of the2000 election found that Republicans seemed to prefer information providedby George W. Bush at the expense of information coming from Vice PresidentGore (Iyengar et al. 2008). A related experiment showed that Republicans choseFox News over CNN, the BBC, and National Public Radio when informingthemselves about political issues (Iyengar and Hahn 2009). These findings maystem from the fact that changes in the media environment make selectiveexposure easier than it used to be.

Given the intensification of partisan animus, it is not surprising that mediachoices increasingly reflect partisan considerations. People who feelstrongly about the correctness of their cause or policy preferences are morelikely to seek out information they believe is consistent with theirpreferences. But while as recently as 25 years ago, these partisans wouldhave been hard pressed to find overtly partisan sources of information,today the task is relatively simple. In the case of Republicans, all they needto do is tune in to Fox News or the O’Reilly Factor. (Bennett and Iyengar2008, 720)

It seems, then, that many people are taking advantage of increasedopportunities to selectively expose themselves to compatible informationsources. This raises serious questions about the media’s ability to foster ahealthy democracy. Apparently, Americans depend almost entirely on the massmedia for exposure to political views unlike their own. And a case could bemade that “exposure to conflicting views” is perhaps “the sine qua non” of “thekind of political dialogue needed to maintain a democratic citizenry” (Mutz andMartin 2001, 97). Encountering “divergent perspectives and viewpoints . . . liesat the core of collective deliberation among citizens in democratic politics”(Huckfeldt, Mendez, and Osborn 2004, 65). When individuals hear and readabout opinions unlike their own, they learn to appreciate the perspectives ofothers; gain an appreciation for their own position; consider changing their

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 199

mind; and grant legitimacy even when they disagree with the outcome (Mutzand Martin 2001). They also are more capable of providing reasons to supporttheir opinions (Price, Cappella, and Nir 2002). Interestingly, media outlets arebetter sources of information about opposing viewpoints than interpersonalnetworks (Mutz and Martin 2001). With interpersonal networks, diversity ishindered by the trend toward residential balkanization—with Republicansmore likely to live next door to fellow Republicans; ditto for Democrats—andthe corresponding tendency of people to seek out individuals like themselves(Huckfeldt et al. 1995). All of this results in “a disappointing tendency towardhomogeneity” (Mutz and Martin 2001, 110). Media outlets are, by contrast,“hotbeds of diversity, not because the media are doing such an exemplary jobpursuing diversity, but because individuals are doing such a poor one” (110).Hence a potential problem with a media environment that is more conduciveto selective exposure: diminished likelihood that certain people are exposed todiverse viewpoints.

Polarization Effect

Stemming from the issue of increased selective exposure is the potential forpolarization. When people seek out and expose themselves to information thatis compatible with their existing views, they presumably become more confidentin their opinions. Although their opinions may not change from one side to theother, their views could become more extreme (Sunstein 2007). By this logic,today’s political media environment—which allows attentive citizens to pickand choose among a wide array of sources, many of them partisan—could becontributing to opinion polarization in the U.S. electorate (Sunstein 2007).Conservatives who watch Fox News or listen to Rush Limbaugh would shiftfurther to the right over time; liberals who watch MSNBC or read progressiveblogs would shift further to the left.

Group polarization research sheds light on how this might happen (seeStroud 2007). Under this theory, discussion or other forms of deliberativecommunication prompt group members to move further in the direction wherethey are predisposed to lean. As a result, “groups of like-minded people,engaged in discussion with one another, end up thinking the same thing thatthey thought before—but in more extreme form” (Sunstein 2007, 60-1).Whereas diverse discussion networks lead to greater awareness of multipleperspectives, attitudes become polarized when people limit their deliberationto like-minded associates (Huckfeldt, Mendez, and Osborn 2004). Polarizationcan occur for several reasons. For one, when a group of like-mindedpeople discusses politics, congenial arguments will tend to outnumbercounterarguments. Second, group members who are out of sync with the groupwill either adjust their views to more closely fit the prevailing viewpoint orabandon the group and seek out other networks more closely aligned with theirpredispositions. Finally, members who are uncertain about their views on

200 | POLITICS & POLICY / April 2011

specific topics gain confidence in their predispositions after hearing or readingthe opinions expressed by people like them (Sunstein 2007).

The logic of group polarization may be applied to the new mediaenvironment (Stroud 2007). Today, like-minded people can easily congregate onthe air and online. Talk radio and blogs foster intragroup discussion amongthose who tend to agree with each other. Even with nondeliberative media, suchas television talk shows, group dynamics come into play because people cantune in to find out what others like them think. When the “discussion” ends,viewers are left with greater familiarity with the arguments that support theirposition (if not the counterarguments) and more confidence that they are right.Sometimes the result is more extreme views. Therefore, it was not surprisingthat people who intended to watch Fahrenheit 9/11, the Michael Mooredocumentary that mocked George W. Bush’s antiterrorism policy, alreadyharbored negative opinions of the president. What is remarkable is thatwatching the film was associated with even less favorable views of the BushAdministration (Stroud 2007). Similarly, the anti-Clinton predispositions ofRepublicans in the 1990s were exacerbated among those who listened toconservative talk radio (Jones 2001). During the 1996 presidential campaign,advertising seemed remarkably effective at reinforcing predispositions about thecandidates (Kaid 1997). In Canada, a fragmented media environment seemed tomagnify preexisting differences between French Quebeckers and citizens of therest of the country (Mendelsohn and Nadeau 1996).

In this study, we employed a quasi-experiment of undergraduate studentsto examine whether such reinforcement effects occurred during the 2008presidential election. Specifically, we analyzed both (1) the extent to which oursubjects selectively exposed themselves to congenial persuasive arguments aboutRepublican vice presidential nominee Sarah Palin, and (2) whether such choicesstrengthened subjects’ preexisting views of Palin.

Research Design

We conducted our quasi-experiment during October 2008. Participants wererecruited from a large introductory General Education U.S. government course,comprising predominantly non-political science majors, at a comprehensiveMasters-level Southern public university. Students who agreed to participate inthe study were offered extra credit on their course grades. Of the 697 studentswho completed the online questionnaire, 30 percent were men and 70 percentwere women. The higher percentage of female respondents reflects the studentbody at the university where this research was conducted (and at many collegesand universities today, where women outnumber men), which has a male-to-female ratio of 40 percent to 60 percent. When “leaners” were includedin party categories, we found that 41 percent of respondents self-identifiedas Republicans, 7 percent as pure independents, and 52 percent as Democrats.

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 201

Forty-eight percent of subjects were liberal, 19 percent were moderate, and 33percent were conservative.

The treatment to which students were exposed consisted of online newsarticles or opinion pieces about Republican vice presidential candidate SarahPalin. Since a pretest designed to gauge students’ initial views about Palin wouldhave taken place such a short time prior to posttest questions and thus wouldhave likely significantly affected posttest scores, we employed a simple posttest-only design. The questionnaire was created and conducted using the onlinesurvey tool Qualtrics. Participants were eligible to complete the survey fromTuesday, October 21 until Sunday, October 26.

Study participants were assigned to one of three groups based on courseinstructor.2 The first group (Group A) consisted of a class section whosemembers read no news or opinion article but who were presented with the samesurvey questions as all other participants. A total of 141 students made up thiscontrol group. The second group (Group B) consisted of participants who werepresented with a choice of two “straight news” articles and asked to select oneof the two. A total of 95 students comprised this group. The remaining 461participants (Group C) were offered a choice of four opinion articles, two ofwhich were critical of Palin and two of which were supportive.3 Participants inGroup B and Group C made their selection based on the article title and authorinformation the survey provided. Of all the subjects in Group C, 238 chose toread an anti-Palin article and 223 chose to read a pro-Palin article.

As indicated above, our primary research questions concerned the extent towhich people choose information sources that support their predispositions andthe extent to which those choices result in the reinforcement of preexistingattitudes. To address these questions, our basic strategy was to compare feelingsand attitudes about Sarah Palin among respondents in all three groups. Wefocused primarily on comparing the views of subjects who read pro-Palinopinion articles with those who read anti-Palin opinion articles.

We spent several hours during September and October 2008 scouring mediawebsites to locate opinion pieces we believed were representative of the pro-

2 None of the instructors (all were consulted) had covered Sarah Palin or the presidential race inany depth at this point in the semester, so there was no concern about nonrandom bias based oninstructor in separating out the groups in this manner. Furthermore, students were not given anyinformation about the research project or intent of the survey, and it was administered as one ofseveral extra credit opportunities available to students throughout the semester. Thus, thepotential for Hawthorne effects is minimized.3 Students in this introductory general education course choose their classes largely on the basis ofschedule and not instructor, and most students have little awareness of political sciencedepartment faculty; consequently, it is extremely unlikely that students would have selected aninstructor who they believed supported their previous partisan leanings. Although all instructorsmust teach the course in accordance with university-established learning objectives, as is the casewith any such course, all instructors teach it with slight variations. As a result, while there may besome nonrandom bias, we propose that this bias is likely to have a small effect on studentattitudes.

202 | POLITICS & POLICY / April 2011

Palin and anti-Palin arguments circulating at the time of our study. In addition,we searched far and wide for news articles that did not carry any discernible biasin favor of or against Sarah Palin. The articles selected were as follows (Table 1).

Subjects in all three groups were asked questions tapping attention topolitics, intention to vote in 2008, and candidate preference (see Appendix forcomplete question wording). The survey began by measuring subjects’ attentionto government and public affairs and the degree to which they had opinionsabout political issues. Next, we assessed subjects’ ideology (using the standardseven-point scale) and party identification (for subjects who responded “don’tknow,” we included questions designed to assess whether they leaned towardone ideology or party), as well as the strength of their party identification.

We then asked subjects about their intention to vote in the 2008 presidentialelection and for whom they planned to vote, again attempting to encouragethose with weak preferences to make a selection. We also asked participantswhether their candidate preference was strong or weak. Before posing a series ofquestions about Sarah Palin, we inquired as to which media outlet subjectswould normally turn to first if seeking out information about the candidate.Subjects were then directed to select their course instructor and, based on theirresponse, were divided into three groups and provided with no article, a choiceof news articles, or a choice of opinion articles.

Subjects in Group B selected one of the two news articles and were providedwith a link to the article. Similarly, subjects in Group C chose one of fouropinion articles and were instructed to click on the link to the article URL.Participants were then asked to verify that they read the article before movingon.

The approach described above measures participants’ reactions to a singlearticle rather than to a series of stories or a television network, newspaper,

Table 1. Treatment and Control Group Articles

Author Title Date Source

Straight-newsarticles

AssociatedPress

“Palin talks economicworries in small townOhio”

October 12, 2008 USA Today

Ken Dilanian “Palin pick shakes upAlaska politics”

October 20, 2008 USA Today

Pro Palin J. C. Watts “Say what you want,Palin is a ‘realperson’ ”

October 12, 2008 Las VegasReviewJournal

Hugh Hewitt “An army of Sarahs” September 12, 2008 Townhall.comAnti Palin Thomas

Friedman“Palin’s kind of

patriotism is reckless”October 8, 2008 The New York

TimesPaul Daly “Whatever Sarah

Palin says, sheexaggerates”

October 14, 2008 New YorkDaily News

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 203

website, or radio personality. As discussed earlier, today’s fragmented mediaenvironment affords citizens the opportunity to obtain political news from awide array of sources, including not only newspaper articles and op-eds but alsocable news shows, talk radio programs, and blogs. A limitation of our researchdesign is that subjects were exposed only to more traditional sources of politicalopinions and information. However, the kinds of articles we selected are oftendisseminated among, and discussed within, other media outlets and take on alife of their own on television and in the blogosphere long after they appear inprint.

Another limitation of our design is that we only examine views at one pointin time rather than having students read a series of articles over the course of anentire semester. This can be viewed as a limitation because we cannot evaluatea change of intensity of the occurrence of selective exposure but can onlyestablish its presence in the selection and reaction to this particular set ofarticles. However, one advantage of this approach is that at the time the surveyswere administered, Sarah Palin was a relatively new political face, even amongRepublicans. She was not completely unknown, as the public had about sixweeks to absorb information that was disseminated during this time period.However, she was still new in comparison with the rest of the key players in thepresidential race. Consequently, it can be argued that the timing makes thisstudy more compelling, and we would have a different set of findings altogetherif we were to examine views over time as Palin became better known. We canalso argue that selective exposure effects should be at their strongest in asituation like this, where partisans more than ever are in need of informationto confirm their existing views on the selection of Palin as the Republicanvice presidential candidate, whether they be positive or negative. In short, ifDemocrats were looking for a reason to dislike Palin, and if Republicans werelooking for a reason to like her, and with many opinions on both sides not beingfully formed at this point, this is the time period during which we should seesuch effects occurring. Our study is effectively able to tap whether there wasselective exposure occurring and whether or not the type of article that subjectsread had any effect on their views toward Palin.

A final limitation of our design is the use of a nonrandom sample. It isoften difficult to justify a convenience sample, but in the interest of testingattitudes in real time, before Palin was well known as a political candidate, wechose to utilize such a sampling method. As mentioned, the students were notpolitical science majors and thus could not have been expected to havesubstantially more knowledge of the candidates or the issues than the averageperson. Finally, nonrandom sampling can have its advantages when time andresources are limited, and here we used it because of time limitations to ourperiod of interest in our study. We contend that while external validity may bemore limited than in previous research that utilizes random samplingtechniques, we are able to test for selective exposure effects in a new andinnovative manner.

204 | POLITICS & POLICY / April 2011

The next portion of the questionnaire included several questions designed tomeasure respondents’ attitudes about Sarah Palin and other prominent politicalfigures and groups. Respondents were asked to rate their feelings about SarahPalin as well as George Bush, John McCain, Barack Obama, Joe Biden, HillaryClinton, Republicans, Democrats, liberals, and conservatives on a standardfeeling thermometer scale of 0-100. The ratio-level Palin feeling thermometerwas one of the key dependent variables in our study. Respondents were alsopresented with Likert scales designed to assess the favorability of theirimpressions of Obama, McCain, Biden, and Palin and their feelings aboutBiden’s and Palin’s qualifications to be president if necessary. These ordinal-level measures tapping respondents’ overall impressions of Palin and herqualifications served as a second set of dependent variables in our study.Finally, respondents were asked to rate Obama’s decision in selecting Biden ashis running mate, and McCain’s decision in selecting Palin as his running mate,on a scale from 0 to 10. The latter served as our third dependent variable. Thefinal page of the questionnaire included a series of demographic questionsconcerning participants’ age, class year, gender, race or ethnicity, andhousehold income.

Findings

Based on the research on selective exposure and related concepts, weexpected that subjects in Group C would choose an opinion article that wascongenial to their predispositions. That indeed proved to be the case. And thepatterns were remarkably consistent. About 76 percent of respondents wholabeled themselves either conservative or extremely conservative chose one ofthe articles defending John McCain’s selection of Sarah Palin as his runningmate (Figure 1). About 73 percent of liberal or extremely liberal respondentsopted for one of the negative articles about Palin. Similarly, an anti-Palin articlewas the choice for 70 percent of Obama supporters and 69 percent of self-described Democrats (Figure 2). Sixty-nine percent of Republicans and 73percent of McCain supporters opted for one of the pro-Palin articles (Figure 1).In short, more than two-thirds of subjects engaged in selective exposure wheninforming themselves about Sarah Palin late in the electoral process.

Did this high level of selective exposure have a polarizing effect? Weexpected that it would. And at the aggregate level, exposure to reinforcingmessages did seem to magnify differences between Republicans and Democrats.When comparing feeling thermometer ratings, the control group results suggestthat the natural difference between Republicans’ and Democrats’ views aboutSarah Palin was about 36 points (see Figure 3), with Republicans (N = 35)averaging a 61.5 rating and Democrats (N = 40) a 25.4. As expected, readingreinforcing messages widened that gap—to 47 points. That is, the 109Republicans who chose to read one of the articles that expressed support forPalin tended to rate her extraordinarily high (mean = 70.0) whereas the 127

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 205

Democrats who read the skeptical article rated her extraordinarily low(mean = 22.9). Similar patterns emerged when comparing ratings of JohnMcCain’s decision to pick Sarah Palin as his running mate on a scale of 1-10(Figure 4). Whereas for the control group the difference between Republicansand Democrats was 3.2, that gap widened to 4.9 when comparing partisans who

Figure 1.Percent Who Chose Pro-Palin Article

0

10

20

30

40

50

60

70

80

Liberal orextremely

liberal

Slightly ormoderately

liberal

Slightly ormoderatelyconservative

Conservativeor extremelyconservative

Republicanssupporter

Obamasupporter

McCainDemocrats

Figure 2.Percent Who Chose Anti-Palin Article

0

10

20

30

40

50

60

70

80

Liberal orextremely

liberal

Slightly ormoderately

liberal

Slightly ormoderatelyconservative

Conservativeor extremelyconservative

Republicanssupporter

Obamasupporter

McCainDemocrats

206 | POLITICS & POLICY / April 2011

chose to read articles that reinforced their presumed predispositions towardPalin. Comparable evidence of reinforcement also surfaced when comparingRepublicans and Democrats on their overall impression of Sarah Palin(Figure 5) and their assessment of her qualifications for becoming president ifnecessary (Figure 6). For example, only 54.3 percent of Republicans who readno article had either a “very favorable” or “favorable” overall impression ofPalin. That compares with 77.1 percent of Republicans who read a reinforcingarticle, resulting in a gap of 76.3 percentage points between Republicans andDemocrats who read reinforcing articles (only one Democrat in this group ratedPalin favorably)—about 26 percentage points wider than their counterparts inthe control group (Figure 5).

Such differences suggest a polarizing effect of reinforcing messages.Interestingly, the opposite occurred when subjects read articles that ran counterto their partisan predispositions. This was a much smaller group; only about 30percent of Republicans (N = 46) read articles that expressed skepticism aboutPalin and only 30 percent of Democrats (N = 52) read opinion articles thatsupported Palin. But for these subjects who sought out opposing messages, thedifferences between Republicans and Democrats narrowed to 7.1 on the feeling

Figure 3.Palin Feeling Thermometer/Differences between Republicans and Democrats

0

10

20

30

40

50

Control (no article) 36.1

Reinforcing article 47.1

Opposing article 7.1

News article 44.7

Notes: Control, respondent who read no article; reinforcing, either Republican whoread pro-Palin article or Democrat who read anti-Palin article; opposing, eitherRepublican who read anti-Palin article or Democrat who read pro-Palin article; newsarticle, respondent who read news article.

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 207

thermometer, 1.1 on the rating of McCain’s decision to select Palin as hisrunning mate, 6.1 on the overall impression, and 3.5 on the assessment ofPalin’s qualifications. Rather than reinforcing their predispositions, thesesubjects may have been persuaded by messages that ran counter to their partisanleanings.

Even more surprising was the apparent effect of reading a less opinionatednews article on Sarah Palin. As with subjects who read opinion articles thatreinforced their partisan predispositions, reading a standard news articleabout Palin also seemed to magnify the differences between Republicans andDemocrats. This happened primarily because the Republicans who read anews article reported more positive feelings about Palin than Republicans inthe control group. For example, the mean thermometer score for Republicanswho read a news article was 70, about 8.5 points higher than Republicanswho did not read an article. This finding suggests that some partisans—Republicans in this case—can find reinforcing information in seeminglyinnocuous sources. Democrats who read a news article about Palin rated her

Figure 4.Rate Palin Choice (1-10)/Difference between Republicans and Democrats

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

No article (control)

Reinforcing article

Opposing article

News article

3.2

4.9

1.1

4.4

Notes: Control, respondent who read no article; reinforcing, either Republican whoread pro-Palin article or Democrat who read anti-Palin article; opposing, eitherRepublican who read anti-Palin article or Democrat who read pro-Palin article; newsarticle, respondent who read news article.

208 | POLITICS & POLICY / April 2011

25.4 on average, no different than the average rating for Democrats in thecontrol group.4

At this basic level of analysis, however, we cannot rule out the possibilitythat the differences between treatment groups stem from their existing attitudesrather than as an effect of reading (or not reading) a particular article aboutPalin. Indeed, Republicans who read the Palin article also rated Joe Bidenslightly (but significantly) lower than Republicans in the control group. Andthey rated Barack Obama slightly lower and John McCain slightly higher. Thisreminds us of the implications of Figures 1 and 2: the more partisan a person is,the more likely he or she was to seek out information sources that are congenialto his or her predispositions. A person who chose the pro-Palin articles was thekind of person who already liked Sarah Palin. To tease out whether reading the

4 Similar patterns emerged when comparing Obama versus McCain supporters and conservativesversus liberals. For example, the mean Palin thermometer score among conservatives who read anews article was 71, nearly six points higher than the score among conservatives who read noarticle; moderates who read a news article also ranked Palin over eight points higher thanmoderates who did not read any article.

Figure 5.Percent Who Rated Palin “Favorable” or “Very Favorable”/Differences between

Republicans and Democrats

0

10

20

30

40

50

60

70

80

No article (control)

Reinforcing article

Opposing article

News article

50.3

76.3

6.1

65.3

Notes: Control, respondent who read no article; reinforcing, either Republican whoread pro-Palin article or Democrat who read anti-Palin article; opposing, eitherRepublican who read anti-Palin article or Democrat who read pro-Palin article; newsarticle, respondent who read news article.

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 209

articles was associated with distinctly positive feelings toward Palin, wecontrolled for respondents’ core ideological leanings and partisanship, alongwith other key variables. Specifically, we regressed indicators of attitudestoward Palin on dummy variables for all three article types—pro-Palin, anti-Palin, and news article (leaving out the control group who did not readanything). Control variables included ideology, a dummy variable for female, adummy variable for strong partisans,5 indicators for year in school (rangingfrom sophomore to senior), and income. For the sake of simplicity, we ranseparate regressions for Republicans and Democrats (Table 2) so that we couldisolate the role of partisanship and examine the effects of article choice onattitudes about Palin separately for each partisan group. Because disaffectedHillary Clinton supporters might be expected to be more favorably disposedtoward Palin, we also controlled for respondents’ feeling thermometer rating ofthe former First Lady.

5 Strong partisans were the respondents who rated themselves as either “Strong Republican” or“Strong Democrat.”

Figure 6.Percent Who Rated Palin “Qualified” or “Very Qualified”/Differences between

Republicans and Democrats

0

10

20

30

40

50

60

No article (control)

Reinforcing article

Opposing article

News article

33.9

51.9

3.5

44.6

Notes: Control, respondent who read no article; reinforcing, either Republican whoread pro-Palin article or Democrat who read anti-Palin article; opposing, eitherRepublican who read anti-Palin article or Democrat who read pro-Palin article; newsarticle, respondent who read news article.

210 | POLITICS & POLICY / April 2011

For the most part, the results fit the pattern described above. Even with thecontrols, Republicans who read one of the pro-Palin opinion articles rated hersignificantly higher than Republicans who did not read anything. The same wastrue for Republicans who read a news article about Palin. Republicans who readone of the negative opinion articles rated her significantly lower on the feelingthermometer, although they were no more critical of McCain’s decision tochoose Palin as his running mate than Republicans who did not read an article.

Fewer differences emerged among Democrats. With the controls in place,Democrats who read opinion articles that criticized Palin rated her no lowerthan Democrats who did not read an article about Palin. Reading a news articleabout her also had no effect. By contrast, the relatively small number ofDemocrats who read a positive opinion article about Palin reported strikinglymore positive feelings toward the vice presidential nominee, even with thecontrols in place.

We also examined the relationship between article readership and attitudesabout Sarah Palin using two ordinal-level variables and identical controls forideology, strength of partisanship, gender, the Hillary Clinton feelingthermometer, year in school, and income. The first ordinal variable tapssubjects’ overall impression of Palin, and the second measures whetherparticipants believe that Palin is qualified to become president if necessary.Five-point response options for both variables ranged from “very favorable” to“very unfavorable” and “very qualified” to “very unqualified,” including amiddle “neutral” option. The results from the ordered probit analysis arepresented in Table 3. Key findings for both Republicans and Democrats were

Table 2. Article Choices as Predictors of Attitudes toward Sarah Palin

Republicans Democrats

DV = FeelingThermometer

DV = Rate PalinChoice (1-10)

DV = FeelingThermometer

DV = Rate PalinChoice (1-10)

Read pro-Palin article .229*** .295*** .391*** .242***Read anti-Palin article -.138* -.014 .052 -.009Read news article .160** .216*** .047 .079Ideology (7 = extremely

conservative).372*** .291*** .211*** .105*

Female .052 .046 .030 .096Hillary Clinton feeling

thermometer.152** .119* .058 .069

Strong partisan .135* .117 -.282*** -.343***Year in school -.034 -.036 -.088 .014Income .037 .016 .034 .082

N = 205 N = 207 N = 254 N = 256Adj R2 = .329 Adj R2 = .209 Adj R2 = .306 Adj R2 = .210

Notes: DV, dependent variable.*** p < .01; ** p < .05; * p < .10.

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 211

similar to the regression model results. Republicans who read a pro-Palin articlewere significantly more likely than Republicans who read nothing to have afavorable impression of Palin and to believe she was qualified to be president.The same was true for Democrats. Republicans who read a news article,however, were no more likely to have more positive views of Palin thanRepublicans who read no article. Similarly, unlike in the regression models,where we found a relationship between reading a negative opinion article aboutPalin and lower feeling thermometer ratings among Republicans, in the probitmodels we found no statistically significant relationship between these variablesfor either Republicans or Democrats. We found similar results as thosepresented above for the ideology control, with conservative Republicansexpressing more favorable views about the Republican vice presidentialnominee’s candidacy and qualifications. Moreover, Republicans who classifiedthemselves as strong partisans were more likely to agree that Palin possesses thequalifications to be president; the opposite was true for strong Democrats.

In summary, reading about Palin did seem to magnify differences betweenDemocrats and Republicans, but not always in the ways we expected. First,Republicans shifted more than Democrats. When given a choice, mostRepublicans selected one of the opinion articles that expressed support forPalin. And when they did, their pro-Palin predispositions were reinforced andstrengthened. To our surprise, there was at least some evidence that reading anews article had a similar effect. For Democrats, however, the only noticeable

Table 3. Article Choices as Predictors of Attitudes toward Sarah Palin, OrderedProbit Results

Republicans Democrats

DV = Impressionof Palin

DV = WhetherPalin isQualified

DV = Impressionof Palin

DV = WhetherPalin isQualified

Read pro-Palin article .782*** .605*** .968*** .567**Read anti-Palin article -.222. -.007 -.011 -.184Read news article .719-. .493* .103 .142Ideology (7 = extremely

conservative).418***. .337*** .234** .193**

Female .330* .367** .060 .315*Strong partisan .494** .490** -.734**** -.791***Hillary Clinton feeling

thermometer.006 .003 -.003 -.000

Year in school -.092 -.065 -.048 -.115Income .001 .017 -.013 -.011

N = 207 N = 199 N = 257 N = 249Adj R2 = .148 Adj R2 = .099 Adj R2 = .119 Adj R2 = .105

Notes: DV, dependent variable.*** p < .01; ** p < .05; * p < .10.

212 | POLITICS & POLICY / April 2011

shift was among those who read the pro-Palin opinion articles. When given achoice, Democrats tended to selectively expose themselves to opinion articlesthat were critical of Palin. But for the most part, their views of Palin were notsignificantly harsher than Democrats who read nothing. To the extent thatmedia choices widened the gap between Republicans and Democrats, much ofthe change can be explained by shifts among the respondents who identifiedwith the GOP.

Discussion

Consistent with some of the relevant literature, we found that many citizensdo engage in selective exposure and that such exposure to compatibleinformation sources magnifies partisan differences. Conservatives, Republicans,and McCain supporters were much more likely to select articles that supportedSarah Palin’s vice presidential candidacy, while liberals, Democrats, andObama supporters were much more likely to choose articles criticizing hercandidacy. Moreover, exposure to these opinion-based information sourcesstrengthened some subjects’ preexisting views about Palin: compared withpartisan members of a control group who read no article, Republicans who reada supportive opinion article were more supportive of Palin. Counter toexpectations, however, we found some evidence of persuasion among partisanswho selected an article that ran contrary to their predispositions, as within thisgroup the gap between Republicans’ and Democrats’ assessments of Palinnarrowed. Finally, we discovered that reading a news article also magnifiedpartisan differences, perhaps because Republicans were inclined to seek outreinforcing information even in relatively objective information sources.

Although we found evidence of selective exposure and reinforcement onboth sides of the political spectrum, Republicans who selected pro-Palin articlesand even news articles were most likely to have their views about Palinreinforced. In short, Republicans were more susceptible to reinforcement thanDemocrats. Regardless of what they read or did not read, Democrats’ viewsabout Palin were critical, although reading a pro-Palin article did softenattitudes compared with Democrats who read either a negative article or noarticle. Among Republicans, the effect of reading either a pro-Palin article or anews story was more pronounced, with subjects in both groups expressing muchmore supportive attitudes.

The particulars of the Palin phenomenon may partly explain the differencesbetween Republicans and Democrats. Our experiment was conducted six weeksafter Palin was introduced as McCain’s vice presidential candidate at theRepublican National Convention. By that time, Democrats’ antipathy towardPalin was firmly entrenched and therefore less susceptible to influence. ManyRepublicans, by contrast, were eager for supportive commentary that easedtheir emerging anxieties about their nominee’s running mate, who had by thensuffered weeks of harsh media scrutiny and late-night comedy parody. Although

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 213

Republicans were inclined to support their candidate, doubts had emerged bymid-October. Perhaps reading the articles reassured them, providing a reminderof why they liked Palin when she was first introduced. Even the news articlesprovided a boost simply because they were not critical—unlike the prevailingtone of Palin-related messages elsewhere.

Future research might explore the possibility that some partisans are morelikely to engage in selective exposure than others. Are Republicans moresusceptible to the reinforcing effects of selective exposure? We think not. In apolitical context where Democrats are experiencing high levels of uncertaintyabout their opinions, they might also turn to reinforcing messages to help allaytheir misgivings. Late in the 2008 presidential race, Democrats were relativelyhappy with their candidate and his running mate. They were less likely to “need”messages that reminded them why they favored one candidate over the other.But in a more ambiguous political setting—for example, the health-care debatein late summer and fall of 2009—we would expect Democrats to behave in muchthe same way as Republicans did in fall 2008. Future research could furtherexplore whether selective exposure is more likely to occur in certaincircumstances or across a particular set of issues.

In today’s choice-rich media environment, it is easy for partisans on the leftand right to seek out and find supporting viewpoints in the media. Even whentheir choice is limited to a handful of sources—as was the case with ourexperiment—they gravitate toward perspectives that are compatible with theirown. To some extent, the result is a magnification of differences.

About the Authors

David A. Jones is an associate professor of political science at JamesMadison University. His research interests include media and politics, politicalpolarization, and public opinion. In addition to coordinating the university’sWashington Semester program, he teaches courses on political media, electioncampaigns, and U.S. government. His research has been published in PoliticalCommunication, the Harvard International Journal of Press/Politics, Politics &Policy, the American Review of Politics, and the International Journal of PublicOpinion Research.

Kathleen Ferraiolo is an assistant professor of political science at JamesMadison University in Harrisonburg, Virginia. Her research agenda focusesprimarily on direct democracy as a policy-making institution in the UnitedStates. Recent and ongoing projects examine state legislators’ response tosuccessful initiatives, the federalism implications of direct democracy, andvariations in state and federal approaches to morality policies including drugcontrol and gambling. Her work has been published in the American Review ofPolitics, Politics & Policy, Publius, Polity, the Journal of Policy History, PolicyStudies Journal, and Teachers College Record.

214 | POLITICS & POLICY / April 2011

Jennifer Byrne is an assistant professor at James Madison University.Jennifer’s research interests focus on U.S. politics, in particular the relationshipbetween identity politics and political attitudes. Her current projects focuson how differing conceptions of American national identity influenceattitudes toward immigrants, and how conceptions of national identity amongrefugees in protracted situations shapes attitudes toward possible durablesolutions.

Appendix

Selected Questions from Online Survey

1. How often do you discuss politics with your family or friends?• Several times a day• Every day• Once or twice a week• Less than weekly• Never

2. Some people seem to follow what’s going on in government and public affairsmost of the time, whether there’s an election going on or not. Others aren’t thatinterested. Would you say you follow what’s going on in government andpublic affairs MOST OF THE TIME, SOME OF THE TIME, ONLY NOWAND THEN, or HARDLY AT ALL?• Most of the time• Some of the time• Only now and then• Hardly at all

3. Some people have opinions about almost everything; other people haveopinions about just some things; and still other people have very few opinions.What about you? Would you say you have opinions about ALMOSTEVERYTHING, about MANY THINGS, about SOME THINGS, or aboutVERY FEW THINGS?• Almost everything• Many things• Some things• Very few things

4. We hear a lot of talk these days about liberals and conservatives. Here is aseven-point scale on which the political views that people might hold arearranged from extremely liberal to extremely conservative. Where wouldyou place YOURSELF on this scale, or haven’t you thought much aboutthis?• Extremely liberal• Liberal• Slightly liberal• Moderate; middle of the road• Slightly conservative

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 215

• Conservative• Extremely conservative• Don’t know

5. [If respondent answered “Moderate/middle of the road” or “Don’t know” toprevious question]. If you had to choose, would you consider yourself aLIBERAL or a CONSERVATIVE?• Liberal• Conservative• Don’t know

6. Generally speaking, do you usually think of yourself as a REPUBLICAN, aDEMOCRAT, an INDEPENDENT, or what?• Republican• Independent• Democrat• Other party• Don’t know• No preference

7. [If respondent answered “Republican” or “Democrat” to previous question].Would you call yourself a STRONG [Democrat/Republican] or a NOT VERYSTRONG [Democrat/Republican]?• Strong• Not very strong• Don’t know

8. [If respondent did not answer “Republican” or “Democrat”]. Do you thinkof yourself as CLOSER to the Republican Party or to the DemocraticParty?• Closer to Democratic• Closer to Republican• Don’t know• Neither

9. So far as you know now, do you expect to vote in the national elections thiscoming November or not?• Yes• No• Not sure

10. [If respondent answered “Yes” to previous question]. Who do you think youwill vote for in the election for President?• John McCain• Barack Obama• Other• Don’t know—haven’t decided

11. [If respondent answered “No” or “Not sure”]. If you were going to vote, whodo you think you would vote for in the election for President?• John McCain• Barack Obama• Other• Don’t know—haven’t decided

216 | POLITICS & POLICY / April 2011

12. Would you say that your preference for this presidential candidate is strong ornot strong?• Strong• Not strong

13. [If respondent was assigned to Treatment Group B (news articles)]. Now wewould like to give you an opportunity to read an article about Sarah Palin.Choose the item that you would most like to read, then click on the double-arrow.• “Palin Talks Economic Worries in Small Town Ohio”• “Palin Pick Shakes Up Alaska Politics”

14. [If respondent was assigned to Treatment Group C (opinion articles)]. Nowwe would like to give you an opportunity to read an article about SarahPalin. Choose the item that you would most like to read, then click on thedouble-arrow.• “Say What You Want, Palin is a ‘Real Person’” by J. C. Watts• “Whatever Palin Says, She Exaggerates” by Paul Daly• “An Army of Sarahs” by Hugh Hewitt• “Palin’s Kind of Patriotism is Reckless” by Thomas Friedman

15. Now we would like to get your feelings toward some of our political leadersand other people who are in the news these days. As you read each name,please rate that person using something we call the feeling thermometer.Ratings between 50 degrees and 100 degrees mean that you feelFAVORABLE and WARM toward the person. Ratings between 0 degreesand 50 degrees mean that you DON’T feel favorable toward the person andthat you don’t care too much for that person. You would rate the person atthe 50 degree mark if you don’t feel particularly warm or cold toward theperson. If we come to a person whose name you don’t recognize, you don’tneed to rate that person. Just move on to the next one. On a scale of 0 to 100,how would you rate . . . ?• George Bush• Barack Obama• John McCain• Joe Biden• Sarah Palin• Hillary Clinton• Bill Clinton• Republicans• Democrats• Liberals• Conservatives

16. Overall, what is your impression of [Very favorable/Favorable/Neutral/Unfavorable/Very unfavorable] . . .• Barack Obama• John McCain• Sarah Palin• Joe Biden

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 217

17. Based on what you know, how qualified is JOE BIDEN to become president ifit becomes necessary?• Very qualified• Somewhat qualified• Neutral• Somewhat unqualified• Very unqualified• Don’t know

18. Based on what you know, how qualified is SARAH PALIN to becomepresident if it becomes necessary?• Very qualified• Somewhat qualified• Neutral• Somewhat unqualified• Very unqualified• Don’t know

19. On a scale of 0 to 10, how would you rate Barack Obama’s decision to select JoeBiden as his running mate?

20. On a scale of 0 to 10, how would you rate John McCain’s decision to selectSarah Palin as his running mate?

21. What is your age?22. At school, are you a . . .

• Freshman• Sophomore• Junior• Senior

23. What is your gender?• Male• Female

24. What racial or ethnic group best describes you?• Caucasian• African-American• Hispanic or Latino• Asian• Other• Prefer not to say

25. What is your best estimate of your household’s total combined income for theyear, before taxes?• Under $25,000• $25,000-$29,999• $30,000-$39,999• $40,000-$49,999• $50,000-$59,999• $60,000-$74,999• $75,000-$99,999• $100,000-$124,999• $125,000-$149,999

218 | POLITICS & POLICY / April 2011

• $150,000-$174,999• $175,000-$199,999• $200,000 or more• Prefer not to say

References

Auletta, Ken. 2003. “Vox Fox.” The New Yorker. May 26, 58-73.

Barker, David C. 2002. Rushed to Judgment: Talk Radio, Persuasion, andAmerican Political Behavior. New York: Columbia University Press.

Bennett, W. Lance, and Shanto Iyengar. 2008. “A New Era of MinimalEffects? The Changing Foundations of Political Communication.” Journal ofCommunication 58 (4): 707-731.

Bimber, Bruce A., and Richard Davis. 2003. Campaigning Online: TheInternet in U.S. Elections. Oxford: Oxford University Press.

Chaffee, Steven H., and Jack M. McLeod. 1973. “Individual V. SocialPredictors of Information Seeking.” Journalism Quarterly 50 (2): 237-245.

Chaffee, Steven H., and Yuko Miyo. 1983. “Selective Exposure and theReinforcement Hypothesis: An Intergenerational Panel Study of the 1980Presidential Campaign.” Communication Research 10 (1): 3-36.

DellaVigna, Stefano, and Ethan Kaplan. 2007. “The Fox News Effect:Media Bias and Voting.” Quarterly Journal of Economics 122 (3): 1187-1234.

Feather, Norman T. 1962. “Cigarette Smoking and Lung Cancer: A Study ofCognitive Dissonance.” Australian Journal of Psychology 14 (1): 55-64.

Festinger, Leon. 1957. A Theory of Cognitive Dissonance. Stanford, CA:Stanford University Press.

Freedman, Jonathan L. 1965. “Preference for Dissonant Information.”Journal of Personality and Social Psychology 2 (2): 287-289.

Frey, Dieter. 1986. “Recent Research on Selective Exposure to Information.”In Advances in Experimental Social Psychology, vol. 19, edited by LeonardBerkowitz. Orlando, FL: Academic Press. 41-80.

Gentzkow, Matthew, and Jesse M. Shapiro. 2006. “Media Bias andReputation.” Journal of Political Economy 114 (2): 280-316.

___. 2010a. “Ideological Segregation Online and Offline.” (September 21).Accessed on January 5, 2011. Available online at http://faculty.chicagobooth.edu/jesse.shapiro/research/echo_chambers.pdf

___. 2010b. “What Drives Media Slant? Evidence from U.S. DailyNewspapers.” Econometrica 78 (1): 35-71.

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 219

Graf, Joseph, and Sean Aday. 2008. “Selective Attention to Online PoliticalInformation.” Journal of Broadcasting & Electronic Media 52 (1): 86-100.

Huckfeldt, Robert, Paul Allen Beck, Russell J. Dalton, and JeffreyLevine. 1995. “Political Environments, Cohesive Social Groups, and theCommunication of Public Opinion.” American Journal of Political Science 30(4): 1025-1054.

Huckfeldt, Robert, Jeanette Morehouse Mendez, and Tracy Osborn.2004. “Disagreement, Ambivalence, and Engagement: The PoliticalConsequences of Heterogeneous Networks.” Political Psychology 25 (1): 65-95.

Iyengar, Shanto, and Kyu S. Hahn. 2009. “Red Media, Blue Media: Evidenceof Ideological Selectivity in Media Use.” Journal of Communication 59 (1):19-39.

Iyengar, Shanto, Kyu S. Hahn, Jon A. Krosnick, and John Walker. 2008.“Selective Exposure to Campaign Communications: The Role of AnticipatedAgreement and Issue Public Membership.” Journal of Politics 70 (1): 186-200.

Jones, David A. 2001. “The Polarizing Effect of New Media Messages.”International Journal of Public Opinion Research 14 (2): 158-174.

___. 2006. “Partisan New Media: The Case of Fox News.” Presented to theAnnual Meeting of the Northeastern Political Science Association, Boston,MA. November 9-11.

Kaid, Lynda Lee. 1997. “Effects of Television Spots on Images of Dole andClinton.” The American Behavioral Scientist 40 (8): 1085-1094.

Klapper, Joseph T. 1960. The Effects of Mass Communication. New York: FreePress.

Larcinese, Valentino, Riccardo Puglisi, and James M. Snyder, Jr. 2007.“Partisan Bias in Economic News: Evidence on the Agenda-Setting Behavior ofU.S. Newspapers.” National Bureau of Economic Research Working Paper13378 (September). Accessed on January 5, 2011. Available online at http://www.nber.org/papers/w13378

Lazarsfeld, Paul, Bernard Berelson, and Hazel Gaudet. 1968. ThePeople’s Choice: How the Voter Makes up His Mind in a Presidential Campaign.3rd edition. New York: Columbia University Press.

Mendelsohn, Matthew, and Richard Nadeau. 1996. “The Magnificationand Minimization of Social Cleavages by the Broadcast and Narrowcast NewsMedia.” International Journal of Public Opinion Research 8 (4): 374-389.

220 | POLITICS & POLICY / April 2011

Mullainathan, Sendhil, and Andrei Shleifer. 2005. “The Market forNews.” The American Economic Review 95 (October): 1031-1053. Accessed onJanuary 5, 2011. Available online at http://www.economics.harvard.edu/faculty/mullainathan/files/marketfornews_102304.pdf

Mutz, Diana, and Paul S. Martin. 2001. “Facilitating Communication acrossLines of Political Difference: The Role of Mass Media.” American PoliticalScience Review 95 (1): 97-114.

Neuman, W. Russell. 1996. “Political Communication Infrastructure.” TheAnnals of the American Academy of Political and Social Science 546 (July): 9-21.

Price, Vincent, Joseph N. Cappella, and Lilach Nir. 2002. “DoesDisagreement Contribute to More Deliberative Opinion?” PoliticalCommunication 19 (1): 95-112.

Prior, Markus. 2007. Post-Broadcast Democracy: How Media Choice IncreasesInequality in Political Involvement and Polarizes Elections. New York:Cambridge University Press.

Puglisi, Riccardo, and James M. Snyder, Jr. 2008. “Media Coverage ofPolitical Scandals.” National Bureau of Economic Research Working Paper14598 (December). Accessed on January 5, 2011. Available online at http://www.nber.org/papers/w14598

Rabin, Matthew, and Joel L. Schrag. 1999. “First Impressions Matter: AModel of Confirmatory Bias.” Quarterly Journal of Economics 114 (1): 37-82.

Sears, David O., and Jonathan L. Freedman. 1967. “Selective Exposure toInformation: A Critical Review.” Public Opinion Quarterly 31 (2): 194-213.

Spruiell, Stephen. 2008. “Olbermann Broadcast Corporation.” NationalReview 60 (17): 36-44.

Stroud, Natalie Jomini. 2006. Selective Exposure to Partisan Information.Unpublished dissertation in partial fulfillment of the requirements for the PhDdegree, University of Pennsylvania.

___. 2007. “Media Effects, Selective Exposure, and Fahrenheit 9/11.” PoliticalCommunication 24 (4): 415-432.

Sunstein, Cass. 2007. Republic.com 2.0. Princeton, NJ: Princeton UniversityPress.

Taber, Charles S., and Milton Lodge. 2006. “Motivated Skepticism in theEvaluation of Political Beliefs.” American Journal of Political Science 50 (3):755-769.

Jones et al. / SELECTIVE EXPOSURE AND SARAH PALIN | 221