14
Please cite this article in press as: Corrigall-Brown, C., & Wilkes, R. Media exposure and the engaged citizen: How the media shape political participation. The Social Science Journal (2014), http://dx.doi.org/10.1016/j.soscij.2014.03.009 The Social Science Journal xxx (2014) xxx–xxx Contents lists available at ScienceDirect The Social Science Journal journa l h om epa ge: www.elsevier.com/locate/soscij Media exposure and the engaged citizen: How the media shape political participation Catherine Corrigall-Brown a,, Rima Wilkes b a Department of Sociology, University of Western Ontario, Canada b Department of Sociology, University of British Columbia, Canada a r t i c l e i n f o Article history: Received 26 January 2013 Received in revised form 13 February 2014 Accepted 16 March 2014 Keywords: Media exposure Political participation Voting Protest Political knowledge Political efficacy Political trust a b s t r a c t Media exposure is widely known to increase institutional forms of political participation such as voting. Less well understood is whether media exposure also affects protest, a less institutional form of engagement. This paper examines the mechanics through which this relationship operates by considering the media’s direct and indirect effect on voting and protesting, via political trust, efficacy, and knowledge. We make these comparisons by ana- lyzing the unique Jennings panel dataset that collects information on respondents at three separate points. The results show that media exposure affects voting more than protesting and that these relationships operate through different mechanisms. While media exposure leads to voting because it increases political knowledge, it is associated with protest via external political efficacy. Furthermore, while this relationship is causal for voting it is only correlational for protest. The results illustrate the importance of disentangling forms of political engagement when considering media effects. Crown Copyright © 2014 Published by Elsevier Inc. on behalf of Western Social Science Association. All rights reserved. 1. Introduction The media are a vital institution of democracy, operat- ing as the central mechanism through which citizens learn about and engage with the democratic process (Boulianne, 2011; Eveland, Hayes, Shah, & Kwak, 2005; Klandermans, van der Toorn, & van Stekelenburg, 2008; Livingstone & Markham, 2008; Shah, Kwak, & Holbert, 2001; Shah, McLeod, & Yoon, 2001; Tolbert & McNeal, 2003). Theo- retical explanations of this connection highlight both the media’s direct role in providing citizens with information about the political process as well as its indirect role in fos- tering attitudes and resources, such as trust, feelings of effi- cacy, and political knowledge, that encourage engagement. Corresponding author. E-mail address: [email protected] (C. Corrigall-Brown). To date, this research has focused on the relationship between media and institutional forms of political action, particularly voting (Gastil & Xenos, 2010; Livingstone & Markham, 2008; McLeod, Scheufele, & Moy, 1999). How- ever, the case has been made that it is also important to consider the media’s connection to non-institutional forms of participation, such as protesting (Boyle & Schmierbach, 2009). Since media coverage of institutional politics and electoral activity is both more frequent and more positive than the media attention accorded to protest campaigns (Boyle, Mcluskey, Devanathan, Stein, & McLeod, 2004; McLeod, 2000; Smith, McCarthy, McPhail, & Augustyn, 2001), it cannot be assumed that the effect of media expo- sure on voting and protesting will be the same. This study seeks to develop a better understanding of these relation- ships by comparing the effect of exposure to the media on voting and protesting. We do so by considering the mechanics through which this exposure may be directly or http://dx.doi.org/10.1016/j.soscij.2014.03.009 0362-3319/Crown Copyright © 2014 Published by Elsevier Inc. on behalf of Western Social Science Association. All rights reserved.

Media exposure and the engaged citizen: How the media shape political participation

  • Upload
    rima

  • View
    245

  • Download
    7

Embed Size (px)

Citation preview

Page 1: Media exposure and the engaged citizen: How the media shape political participation

Ms

Ca

b

a

ARRA

KMPVPPPP

1

ia2v&Mrmatc

0

The Social Science Journal xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

The Social Science Journal

journa l h om epa ge: www.elsev ier .com/ locate /sosc i j

edia exposure and the engaged citizen: How the mediahape political participation

atherine Corrigall-Browna,∗, Rima Wilkesb

Department of Sociology, University of Western Ontario, CanadaDepartment of Sociology, University of British Columbia, Canada

r t i c l e i n f o

rticle history:eceived 26 January 2013eceived in revised form 13 February 2014ccepted 16 March 2014

eywords:edia exposure

olitical participationoting

a b s t r a c t

Media exposure is widely known to increase institutional forms of political participationsuch as voting. Less well understood is whether media exposure also affects protest, a lessinstitutional form of engagement. This paper examines the mechanics through which thisrelationship operates by considering the media’s direct and indirect effect on voting andprotesting, via political trust, efficacy, and knowledge. We make these comparisons by ana-lyzing the unique Jennings panel dataset that collects information on respondents at threeseparate points. The results show that media exposure affects voting more than protestingand that these relationships operate through different mechanisms. While media exposureleads to voting because it increases political knowledge, it is associated with protest via

rotestolitical knowledgeolitical efficacyolitical trust

external political efficacy. Furthermore, while this relationship is causal for voting it is onlycorrelational for protest. The results illustrate the importance of disentangling forms ofpolitical engagement when considering media effects.

Crown Copyright © 2014 Published by Elsevier Inc. on behalf of Western Social ScienceAssociation. All rights reserved.

. Introduction

The media are a vital institution of democracy, operat-ng as the central mechanism through which citizens learnbout and engage with the democratic process (Boulianne,011; Eveland, Hayes, Shah, & Kwak, 2005; Klandermans,an der Toorn, & van Stekelenburg, 2008; Livingstone

Markham, 2008; Shah, Kwak, & Holbert, 2001; Shah,cLeod, & Yoon, 2001; Tolbert & McNeal, 2003). Theo-

etical explanations of this connection highlight both theedia’s direct role in providing citizens with information

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

bout the political process as well as its indirect role in fos-ering attitudes and resources, such as trust, feelings of effi-acy, and political knowledge, that encourage engagement.

∗ Corresponding author.E-mail address: [email protected] (C. Corrigall-Brown).

http://dx.doi.org/10.1016/j.soscij.2014.03.009362-3319/Crown Copyright © 2014 Published by Elsevier Inc. on behalf of West

To date, this research has focused on the relationshipbetween media and institutional forms of political action,particularly voting (Gastil & Xenos, 2010; Livingstone &Markham, 2008; McLeod, Scheufele, & Moy, 1999). How-ever, the case has been made that it is also important toconsider the media’s connection to non-institutional formsof participation, such as protesting (Boyle & Schmierbach,2009). Since media coverage of institutional politics andelectoral activity is both more frequent and more positivethan the media attention accorded to protest campaigns(Boyle, Mcluskey, Devanathan, Stein, & McLeod, 2004;McLeod, 2000; Smith, McCarthy, McPhail, & Augustyn,2001), it cannot be assumed that the effect of media expo-sure on voting and protesting will be the same. This study

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

seeks to develop a better understanding of these relation-ships by comparing the effect of exposure to the mediaon voting and protesting. We do so by considering themechanics through which this exposure may be directly or

ern Social Science Association. All rights reserved.

Page 2: Media exposure and the engaged citizen: How the media shape political participation

Social S

2 C. Corrigall-Brown, R. Wilkes / The

indirectly related to these two modes of political engage-ment.

First, media exposure could have a direct effect onengagement. If media exposure is higher among the kindsof people already disposed to participate in the politicalprocess, then this relationship is conditional. If increasesin media usage also lead to further political activity thenthis relationship is causal. Newspapers may, for exam-ple, remind readers of upcoming elections, the openinghours of polling stations, and/or the location of protestevents.

Second, the effect of media exposure may beindirect if it is associated with, or increases, otherpredictors of political engagement. Here we con-sider three such predictors – trust, external politicalefficacy, and knowledge – that have been widely consid-ered important in previous work on the media–politicsrelationship (Appendix A). If the kinds of people who usethe media more are also the kinds of people who havehigher levels of trust, external efficacy, and knowledgethen this relationship is correlational. If media usageincreases the amount of trust, external efficacy, and/orknowledge people have and this, in turn, increases theirengagement, then this relationship is causal. That is,citizens who engage with the media not only becomemore socially and politically aware but also gain a sensethat they can and should affect social change.

In order to consider these multiple scenarios, we takeadvantage of the unique Jennings and Stoker (2004) paneldataset. This dataset, originally designed to study youthpolitical socialization, asks the same respondents abouttheir institutional and non-institutional political participa-tion as well as their media exposure,1 political knowledge,trust, and external efficacy at three separate time points.The fact that this dataset contains multiple measures for thesame individuals allows us to ascertain the extent to whichthere are direct and indirect effects of media usage on par-ticipation and to also ascertain whether these effects arecorrelational or causal. Our analyses show that media expo-sure has both a direct and indirect association with votingand protesting but that the nature of these connections isdifferent.

The direct effects of the media are not only stronger inthe case of voting, but they are also more likely to be causal.When individuals increase their media exposure over time,they also increase their propensity to vote. However, therelationship between media exposure and protesting issimply correlational. The type of individual who uses moremedia is more likely to protest but increasing an individ-ual’s media consumption over time does not increase theirpropensity to protest. In addition, the relationship betweenmedia exposure and voting is mediated by political knowl-

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

edge, whereas, the relationship between media exposureand protesting is mediated by external political efficacy.

1 The Internet is not assessed in this analysis because the data forthe study were collected between 1965 and 1997, a period duringwhich Internet use was either non-existent or marginal. A discussionof the implications of the Internet for participation is presented in theconclusion.

cience Journal xxx (2014) xxx–xxx

2. The media and political engagement

2.1. Mass media and institutional and non-institutionalpolitical engagement

The importance of mass media in contemporary societyhas sparked a wide array of research on the individual-levelconsequences of media exposure. In particular, an expan-sive body of work examines the role of media exposurein shaping institutional political participation, particu-larly voting (Chaffee & Kanihan, 1997; Lawson & McCann,2005; McLeod et al., 1999; Putnam, 2000). Livingstone andMarkham (2008), for example, show that exposure to themass media and, in particular, listening to political radioand engaging with news is associated with a significantlyhigher propensity to vote. Other work finds that readingthe newspaper is associated with an increased likelihoodof both voting and participating in political campaigns(Eveland & Scheufele, 2000). While this work shows thatmedia use is related to voting, the causal direction throughwhich this relationship operates is unclear. Is it that peo-ple who are heavy users of media are more likely to voteor is it that those who vote then seek out media coverage(Livingstone & Markham, 2008)? Furthermore, few studiesexamine the relationship between media use and protestbehavior (except Boyle & Schmierbach, 2009; Livingstone& Markham, 2008).

This is significant because there are several reasons toexpect that the effect of media exposure differs for votingand protesting. Mainstream media coverage of protest andelectoral politics differ in both the amount and tone of thecoverage each of these forms of engagement receive andthis could differentially impact the mobilizing potential ofthe media across these modes of participation. Researchconsistently shows that activities that are either outsidethe normal range of routine politics or that threaten thestatus quo receive either little coverage or no coverage atall (Boyle et al., 2004; McCarthy, McPhail, & Smith, 1996;Smith et al., 2001) although there is evidence that this isshifting with recent changes in the media environment(Weaver & Scacco, 2013). A principle means through whichthe media fosters voting is by providing mobilizing infor-mation, informing viewers as to where, when, and how toparticipate in elections (Schussman & Soule, 2006; Verba,Schlozman, & Brady, 2002). Even if the media does notcover a local election race, the media covers the electionat the state or national level, providing information on thedates and locations of polling. Because protest events getless coverage than institutional politics, such as voting, themedia can provide less mobilizing information to potentialparticipants.

In addition to the low levels of protest coverage, protestevents that are covered do not necessarily benefit from thepositive framing that electoral politics often receive in themedia. It has for some time been clear that media coverageoften undermines movement goals, members, and issues(Giltin, 1980). As Smith et al. (2001, p. 1397) note, “even

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

when movements succeed at obtaining the attention ofmass media outlets, media reports portray protests in waysthat may undermine social movement agendas”. Empiricalwork supports this assertion. Media protest coverage tends

Page 3: Media exposure and the engaged citizen: How the media shape political participation

Social S

t1iftpm22

ocecnicaimg

pgepsTaert&w1irtinipottidpph

ttiocipabfi

with a lower likelihood of participating in any politicalaction, be it voting, political campaigning, or engagementwith political parties (Bélanger & Nadeau, 2005), a relation-ship that holds in cross-national studies (Putnam, 2000).2

C. Corrigall-Brown, R. Wilkes / The

o show protesters in a negative light (McLeod & Detenber,999), include critical headlines (Boyle et al., 2004), and

nvoke marginalizing frames and labels (McLeod, 2000). Inact, content analyses of protest coverage shows a consis-ent bias against protesters and the events they sponsor,articularly when protesters are challenging larger ele-ents of the political system (Corrigall-Brown & Wilkes,

012; McLeod & Hertog, 1992; Wilkes & Corrigall-Brown,010).

The lower levels of coverage and the negative framingf protest events in the media can have serious impli-ations for how media exposure differentially promotesngagement in voting and protesting. A lack of protestoverage can make it difficult for those interested in alter-ative participation to find appropriate models of political

nfluence (Boyle & Schmierbach, 2009). Negative protestoverage can also serve as reinforcement against protest,s compared to the media’s generally positive depiction ofnstitutional engagement. As a result, mainstream media

ay discourage participation in protest activity even as itenerally encourages voting.

It is also important to note that different media coverrotest in different ways. For example, print media isenerally less extreme in its criticism of protest (Chot al., 2003) and, given the greater amount of space inrint media, is more likely to cover protest events andocial movement causes (Entman, 2005; Kaniss, 1991).he larger amount of coverage space in print media alsollows for greater depth of coverage and contextual cov-rage of protest issues (McVeigh & Smith, 1999). For theseeasons, newspaper and radio have the biggest role in fos-ering the acquisition of political knowledge (Livingstone

Markham, 2008). Television exposure, however, has aeaker or negative effect on engagement (McLeod et al.,

999; Patterson, 2001; Scheufele & Nisbet, 2002). This is,n part, because television news uses a debate style ofeporting and, as a result, is less likely to provide con-extual information (Boyle & Schmierbach, 2009). Theres also more focus on tempo and emotions in televisionews (Altheide & Snow, 1991; Meyrowitz, 1985). This more

ntense quality may heighten negative representations ofrotesters through, for example, images of angry activistsr of confrontations with police. Therefore, we hypothesizehat H1a: media exposure is positively related to voting andhat H1b: media exposure is negatively related to protest-ng. We also expect that the strength of these relationshipsepends on the type of medium examined with newspa-er being the most important predictors of both voting androtesting. We predict that exposure to political televisionas a negative or neutral effect on engagement in politics.

The hypotheses thus far denote that the media’s rela-ionship with political engagement is direct. Howeverhis relationship may also be indirect if media exposurencreases political participation through its relationship tother individual features. In the discussion that follows, weonsider three factors that are given extensive attentionn the media–politics literature: political trust, external

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

olitical efficacy, and political knowledge. Appendix A lists selection of key studies that examine the relationshipetween media exposure and political participation. Therst two columns list the study’s author(s) and year of

cience Journal xxx (2014) xxx–xxx 3

publication. The third column provides the number ofwaves of data analyzed and shows that of 21 studies allbut two consider only one or two waves of data. The finalthree columns list the type of media examined, the measureof political participation used, and the mediating variablesassessed. This table illustrates that trust, external politicalefficacy, and knowledge are the major focus of attention instudies of the effect of media exposure on political engage-ment.

2.2. Political trust

Political trust reflects citizens’ evaluation of the per-formance of political authorities and political institutions(Citrin & Luks, 2001; Hetherington & Husser, 2012; Miller,1974; Wilkes, 2011). Media exposure has a complicatedrelationship with political trust. Some studies find thatmedia exposure is associated with reduced trust becauseits negative and critical portrayals of political leaders andcandidates can create a feeling of malaise (Hallin, 1992;Hart, Smith-Howell, & Llewellyn, 1990; Kleinnijenhuis,van Hoof, & Oegema, 2006). Even entertainment program-ming involves generally negative portrayals of the politicalprocess (Lichter, Lichter, & Rothman, 1994) and has thepotential to cultivate cynical attitudes toward government(Grossman, 1995). This type of coverage, in turn, can leadto lower levels of trust in political figures and institutions(Capella & Jamieson, 1996; except Moy & Scheufele, 2000),and as a consequence, could discourage engagement inpolitical acts (Kleinnijenhuis et al., 2006; Patterson, 1993).

Another set of studies finds a more positive relationshipbetween media exposure and political trust (Gross, Aday, &Brewer, 2004; Hetherington, 1998; Moy & Scheufele, 2000;Norris, 2002). The negative coverage and focus on conflictand scandal, which is increasingly the norm in modernmedia, may be keeping the audience enthralled (Newton,1999). Political radio is important in this process. Barkerand Knight (2000) show that listening to political talk radio,including call-in shows that emphasize discussion of poli-tics, elections, and public policy issues, polarizes attitudesabout public figures, groups, and policies (Barker, 1998).Political talk radio might be particularly important becauseof its deviation from the objective journalism norms foundin more traditional broadcast and print news. However,although this type of news consumption may make a morecynical population, some levels of distrust and cynicismmay enhance participation (De Vreese & Semetko, 2002).

Citizens are much more likely to engage with the polit-ical system if they have confidence in it (Hadjar & Beck,2010). A lack of trust in the political system is associated

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

2 Putnam shows that trust and civic engagement are strongly corre-lated. However, the role of trust is complicated by the fact that, at leasttheoretically, a high level of political trust could lead to a false senseof security or remove feelings of individual responsibility, which couldreduce levels of participation (Nyækowiak, 2009).

Page 4: Media exposure and the engaged citizen: How the media shape political participation

Social S

1998; Patterson, 2001; Scheufele & Nisbet, 2002; Shah,Kwak, et al., 2001; Shah, McLeod, et al., 2001).4 Patterson(2001) calls attention to the dramatic growth in television

4 C. Corrigall-Brown, R. Wilkes / The

Therefore, we hypothesize that H2: Media attention isassociated with decreased political trust and this lowerlevel of trust leads to a lower propensity to both vote andprotest.

2.3. Political efficacy

Media consumption is also associated with feelingsof efficacy, the belief that one is capable of the specificbehaviors required to produce a desired outcome in agiven situation (Gecas, 2000). A high level of efficacy indi-cates that individuals are self-confident in their ability tochange and respond to their environment. Efficacy directedat the political system can either reflect feelings aboutone’s ability to understand and engage with that sys-tem (internal political efficacy) or a confidence that onecan shape political outcomes (external political efficacy)(Acock, Clarke, & Stewart, 1985; Miller, Miller, & Schneider,1980).3

Iyengar (1991) argues that because exposure to certaintypes of media, particularly newspapers, encourages peo-ple to reason with issues or about the role of government,this can increase feelings of both internal and externalpolitical efficacy. Newspapers are a particularly importantmedium for fostering feelings of internal and external effi-cacy. This is because they have a more local focus and, as aresult, work to connect individuals to their community andincrease their feelings that they can shape this community(Kim & Ball-Rokeach, 2006).

Without feelings of competency and a belief in theconsequentiality of one’s actions, there is little incen-tive to participate in politics (Abramson & Aldrich, 1982).Consistent with this argument, empirical studies demon-strate a positive relationship between efficacy and voting(Corrigall-Brown, Snow, Smith, & Quist, 2009; Corrigall-Brown, 2011, 2012; Dalton, Van Sickle, & Weldon, 2009;Gastil & Xenos, 2010; Jung, Kim, & Homero, 2011;Klandermans et al., 2008; Verba et al., 2002; Weldon &Dalton, 2014, chap. 7). Pinkleton, Weintraub, and Fortman(1998), for example, show that feelings of political effi-cacy are positively associated with voting behavior and areshaped by the active use of political media. McLeod et al.(1999) also find that reading hard news in the newspaperis associated with both increased knowledge and politicalefficacy and that these factors are the strongest predictorsof institutional political participation.

Some research in the area shows that both internaland external feelings of efficacy are related to politicalparticipation. Stephen and Maggiotto (1981), for example,find that increased internal and external efficacy are bothrelated to participation in elections and political campaigns

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

(Waldner, 2001). However, external political efficacy mightbe a more important predictor of engagement than inter-nal efficacy: many studies show a larger effect of external

3 One’s feeling that the government is responsive to one’s interests is animportant predictor of engagement in both protesting and voting. As Dahlwrites, “A key characteristic of democracy is the continued responsivenessof the government to the preferences of the people” (1971, p. 1). The roleof political responsiveness cannot be tested in this paper as this variableis not available in the Jennings data set.

cience Journal xxx (2014) xxx–xxx

than of internal efficacy on participation in electoral politics(Abramson & Aldrich, 1982; Finkel, 1985; Shaffer, 1981).This may be because external political efficacy shapes polit-ical sentiments and, consequently political participation,while internal political efficacy is more likely to affect self-esteem and self-efficacy, which might not translate intopolitical engagement (Yeich & Levine, 1994).

Media exposure is expected to increase feelings ofexternal efficacy, which, in turn, should increase thepropensity to both vote and protest. We hypothesize thatH3: Media exposure is associated with increased exter-nal political efficacy and, as a consequence, a higherpropensity to both vote and protest. Based on the liter-ature that highlights the significance of external politicalefficacy, as opposed to internal political efficacy, in thispaper, this relationship is tested with a measure of externalpolitical efficacy.

2.4. Political knowledge

It has been widely established that those who knowmore about politics are more likely to engage with thepolitical system (Kenski & Jomini Stroud, 2006; Putnam,2000; Sotirovic & McLeod, 2001). While political knowl-edge can be gained through a variety of sources, such asformal education (Delli Carpini & Keeter, 2003; Nie, Junn,& Stehlik-Barry, 1996) and interpersonal discussion aboutpolitics (Delli Carpini & Keeter, 2003), traditional newsmedia consumption is also strongly related to the acquisi-tion of this type of knowledge (Chaffee, Zhao, and Leshner,1994; Scheufele & Nisbet, 2002).

News media inform and educate the public aboutpolitical issues and debates, though this is contingenton the type of news outlet examined. Newspaper andradio consumption has the strongest relationship withpolitical knowledge in large part because it offers morein-depth coverage than other media, particularly televi-sion (Livingstone & Markham, 2008; Patterson & McClure,1976; Scheufele, Shanahan, & Sei-Hill, 2002; Shah, Cho,Eveland, & Kwak, 2005; Smith, 1986). Newspapers arealso more likely to focus on local issues and, therefore, tofoster the acquisition of local political information (Kim& Ball-Rokeach, 2006). This relationship is not as strongfor television exposure. This is, in part, because of themore recent trend in television, whereby, increasingly highentertainment content can inhibit the acquisition of politi-cally relevant information (McLeod et al., 1999; McQuail,

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

reporting of soft news unrelated to public affairs or policy

4 More recent work assesses how the Internet fosters engagement. Shahand colleagues (2001) find, for example, that Internet use for informationexchange (as opposed to social recreation and product consumption) ispositively related to interpersonal trust and political participation, evenafter controlling for demographics, contextual factors, and traditionalmedia use. Tolbert and McNeal (2003) also show that those who haveaccess to the Internet and to online election news were significantly morelikely to vote in the 1996 and 2000 U.S. elections. In contrast, Livingstoneand Markham (2008) find no effect of Internet use on participation.

Page 5: Media exposure and the engaged citizen: How the media shape political participation

Social S

tomo(Tmr

3

epasrtwiaelSbudi

c

uf11wsd1rycpTit2

pcbibi

Bdtaa

mdw

C. Corrigall-Brown, R. Wilkes / The

hat is typically more sensational, personality- or celebrity-riented, and incident-based.5 Exposure to this kind ofedia is, therefore, less likely to foster the acquisition

f political knowledge and, in turn, political engagementexcept Baum, 2003). Therefore, we hypothesize that H4:he relationship between media attention and voting isediated by political knowledge. We do not expect this

elationship to hold for protesting.

. Data and methods

We use the Jennings and Stoker (2004) dataset toxamine the relationship between media exposure andolitical participation. This dataset comprises a nation-lly representative longitudinal probability sample of highchool seniors (N = 1669) first interviewed in 1965 ande-interviewed in 1973, 1982, and 1997 (N = 934, 56% ofhe original sample).6,7 While the dataset was collectedith the original aim of studying political socialization, it

s ideal for our purposes because it contains informationbout media usage, trust, external political efficacy, knowl-dge, and political participation.8 It also differs from otherarge-scale longitudinal data sets, such as the U.S. Generalocial Survey and the American National Election Studies,ecause it follows the same people over time.9 This allows

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

s to examine how changes in features of one individualirectly impact their changing propensity to engage in pol-

tics.

5 At the same time, political campaigns are becoming more candidate-entered and less focused on broad issues (Wattenberg, 1992).

6 The Jennings and Stoker data (2004) are available at the Inter-niversity Consortium for Political and Social Research through theollowing catalog numbers: 1965 (ICPSR07286-v3), 1973 (ICPSR07779),982 (ICPSR09553), and 1997(ICPSR04023-v2). To achieve the sample of965 high school seniors, a selection of American schools was first madeith probability proportionate to size. About nine in ten of the schools

elected agreed to take part in the study. Within schools, systematic ran-om samples were taken. In total, 1669 students were interviewed in965, an average of 17 seniors per school. Of those seniors selected, theesponse rate was over 99%. As a result of this sampling procedure, theouth panel is a probability sample of the American high school seniorlass of 1965. Jennings (2002) has identified the presence of a slightanel bias over the course of the project due to respondent dropout.he panelists who remained in the sample were slightly more engagedn and knowledgeable about politics and had slightly more liberal atti-udes (ibid). In no case, however, did panel status account for more than

% of the variance in these scores.7 The strength of this data is that it allows us to consider the same peo-

le at three different time points. This allows us to better assess causallaims and examine how features of individuals at time 1 can affect theirehavior at time 2. Of course, some caution is needed when interpret-

ng the impact of the independent variables at one time on subsequentehavior because of the other changes that might have occurred in an

ndividual’s life between periods.8 We used the Amelia program designed by Honaker, King, andlackwell (version 1.2–15) to create multiple imputations of the missingata. We ran the model on each of five imputed data sets and combinedhe results for the final findings (Honake et al., version 1.2–1.5). Separatenalyses using listwise deletion yielded very similar results (results arevailable from the authors upon request).9 As illustrated in Appendix A, studies of the relationship betweenedia exposure and voting and/or protesting have yet to employ panel

ata with three waves (except Boulianne, 2011), the minimum number ofaves of data required to consider the impact of changes over time.

cience Journal xxx (2014) xxx–xxx 5

We measure political participation with two variables:voting and participation in protest. Respondents wereasked about each specific Congressional and Presidentialelection in the survey period, if they were eligible to vote,and if they did vote. Since there was a variable rangeof elections for each time period surveyed, this variableis calculated as the overall percentage (0–100) of elec-tions in a particular period in which each respondentvoted (mean = 0.67, standard deviation = 0.35). The protestvariable is dichotomous, indicating whether the respon-dent engaged in a protest event since the last survey(mean = 0.11, standard deviation = 0.31). Because our inter-est is in political participation, and questions on this topicwere not asked until the second wave,10 the 1965 data arenot included in these analyses.

The primary predictor variables are media use, politi-cal trust, external political efficacy, and knowledge. Mediais assessed with three variables measuring the frequencyof newspaper, radio, and television use related to pol-itics. Responses ranged from never (0) to almost daily(4).11 Political trust is a composite scale of five ques-tions: do you think the government is crooked, do youthink the government wastes money, do you trust thegovernment, do you think that the government is smart,and do you think the government works for everyone?These questions are re-coded in the same direction andsummed to create a scale ranging from 0 (low trust ingovernment) to 5 (high trust in government) (mean = 2.71,standard deviation = 1.31).12 The measure of external polit-ical efficacy is based on two questions asking respondentswhether they agreed or disagreed with the following:“I don’t think public officials care much what peoplelike me think” and “People like me don’t have any sayabout what the government does.” External efficacy rangesfrom 0 (low) to 4 (high) (mean = 2.00, standard devia-tion = 0.72). Political knowledge is measured on a scaleranging from 0 to 6 representing the number of correctanswers to six questions about politics (mean = 3.42, stan-

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

dard deviation = 1.46).13 We also include controls for familyincome, education, political ideology, sex, and race. Familyincome was standardized into 1997 dollars and then logged

10 Respondents were too young to vote in the initial survey and thesurvey did not ask about engagement in protest until the 1973 wave.

11 The question wordings were as follows: “Do you read about publicaffairs and politics in any newspapers? If yes, how often?”; “How abouttelevision – do you watch any programs about public affairs, politics, andthe news on television? If yes, how often?”; and “How about the radio– do you listen to any programs about public affairs, politics, and thenews on the radio? If yes, how often?”. Television viewership (mean = 3.03,standard deviation = 1.26), Radio (mean = 1.91, standard deviation = 1.76),Newspaper (mean = 2.87, standard deviation = 1.42).

12 The five measures of political trust are also found in the NES. TheCronbach’s alpha score ranges from 0.67 in 1965 to 0.55 in 1997.

13 The questions are as follows: “About how many years does a U.S. sen-ator serve?”, “Marshal Tito is a leader in what country?”, “Do you happento know about how many members there are on the U.S. Supreme Court?”,“Who is the governor of (name of this state) now?”, “During World War II,which nation had a great number of concentration camps for Jews?”, and“Do you happen to remember whether President Franklin Delano Roo-sevelt was a Republican or Democrat? (If necessary): Which?” Cronbach’salpha for these questions ranges from 0.61 in 1965 to 0.62 in 1997.

Page 6: Media exposure and the engaged citizen: How the media shape political participation

Social S

6 C. Corrigall-Brown, R. Wilkes / The

(mean = 55,172.21, standard deviation = 33,958.48).14 Edu-cation is the number of years of post-secondary education(ranging from 0 to 12) that the respondent has completedsince graduating high school at the time of the initial survey(mean = 1.13, standard deviation = 0.32). Political ideologyis an index averaging two component variables: (1) self-placement on a labeled 7-point ideological scale like theone used in the NES and (2) the difference in feelingthermometer ratings of conservatives and liberals. Thisresults in a scale from 0 to 7 with 0 indicating politi-cally left (mean = 4.76, standard deviation = 0.46).15 Sex andrace are dummy coded (male and white are the referencecategories) (50% female; 91% White, 7% Black, 2% Other).Appendix B provides a summary of the descriptive statisticsfor the independent and dependent variables.

In order to assess the relationship between media usageand political participation, we use a combination of randomand fixed effects models.16 These are specified as linear inthe case of voting, because the voting variable refers to thepercentage of elections in which an individual could voteand did within the time period. The models are specifiedas logistic in the case of protesting because the variableindicates whether the respondent participated in a protestevent or not in the time period.

The linear random effects model, which is specified asyit = ˇ1xit + εit , represents both cross-sectional and over-time variability and provides a general picture of the factorsassociated with voting and protesting by looking at bothcross-sectional and over-time variance. In essence, thesemodels show why some people participate while othersdo not and why individuals move in and out of engage-ment over time. In contrast, the linear fixed effects modelsexamine the factors that explain over-time variance inparticipation and show why the same individual may par-ticipate at one point in time but not at another (Wooldridge,2001). This model takes the form: yit − yi = ˇ1(xit − xi) +εit − εit , where i is the grouping variable and t is the indi-vidual observations at the three different time points. Bycomparing the results of the random effects model, whichrepresents both over-time and cross-sectional variance,and the fixed effects model, which only examines over-time variance, we are able to isolate the factors that arerelated to differences across and within individuals.

We begin by estimating a model with the mediavariables predicting each mode of political participation,voting, and protesting. This provides a baseline for assess-ing the relative importance of each media variable onvoting and protesting. Then, in order to ascertain the extentto which there are indirect effects, trust, external efficacy,and knowledge are added via a series of nested models. Thechange in the media coefficients represents the portion of

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

the media’s effect that operates indirectly via these othercharacteristics.

14 Family income was logged in order to reduce its positive skew.15 These two variables are correlated at 0.77 in 1997.16 Because the data follow the same individuals over time with occasions

nested in subjects, observations within individuals are not independent(Rabe-Hesketh & Skrondal, 2008). These models account for potential cor-related errors that could arise from this lack of independence across theobservations for each individual (Wooldridge, 2001).

cience Journal xxx (2014) xxx–xxx

4. Findings

4.1. Voting

Table 1 shows the random effects models predictingthe effect of media use, trust, external political efficacy,and knowledge on voting behavior, controlling for income,education, political ideology, sex, and race. These mod-els examine both the factors that predict who will voteas well as an individual’s changing propensity to voteover time. The baseline model (model 1) includes onlythe control variables. In this model, income, years of post-secondary education and conservative ideology each havea significant positive effect on voting. Model 2 adds themedia exposure variables (newspapers, radio, and televi-sion). Newspaper exposure is the most important factorpredicting voting in this model. Increasing the number oftimes per week that a respondent reads the newspaper byone point on the 0–4 scale increases his or her chance ofvoting by 3.1%. Listening to the radio is also positive andsignificant, although this relationship is weaker. Increasingradio exposure by one point on the 0–4 scale increases thelikelihood of voting by 0.9%. Watching political television isnot significantly related to voting behavior. The inclusion ofthe three media variables increases the explanatory powerof the model from 11.6% to 16.5%, showing the importanteffect that media exposure has on voting.

Models 3 and 4 add the political trust and external effi-cacy variables. These models show a small positive effectof trust on voting and a positive effect of external efficacyand voting. A one point increase in trust on the 0–5 scaleleads to a 1.5% increase in the likelihood of voting and a onepoint increase on the 0–4 external efficacy scale increasesthe likelihood of voting by 2.9%. The coefficients for eachof the three media variables remain virtually unchangedwith the inclusion of trust and external efficacy in models3 and 4. This indicates that the impact of media use on vot-ing is not because of the relationship between the formerand either trust or external efficacy. While the inclusion oftrust in model 3 and external efficacy in model 4 increasethe model’s explanatory power, the change in the overallR2 is relatively small when compared to the model withonly the media variables. For example, a 0.04% improve-ment for the trust model and 0.22% improvement for theexternal efficacy model.

Model 5 includes the effect of political knowledge,which is positively associated with voting. An increase ofone point on the 0–6 political knowledge scale increasesthe chance that someone will vote by 8.0%. The inclusion ofthis variable sizably decreases the magnitude of the directrelationship between newspaper readership and voting. Inthis model, a one-point increase in newspaper readershipon the 0–4 scale only increases one’s chance of voting by2.3%, compared with about 3.1% in the earlier models with-out knowledge. This suggests that the relationship betweenexposure to some forms of media and voting is partlyaccounted for by how reading the newspaper and listening

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

to the radio increases political knowledge. It is also impor-tant to note that, once political knowledge is included in themodel, political ideology is no longer a significant predic-tor of voting. The inclusion of political knowledge notably

Page 7: Media exposure and the engaged citizen: How the media shape political participation

C. Corrigall-Brown, R. Wilkes / The Social Science Journal xxx (2014) xxx–xxx 7

Table 1Linear panel models predicting voting behavior, random effects.

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

MediaNewspaper 0.031*** 0.032*** 0.030*** 0.023*** 0.023***

(0–4) (0.004) (0.004) (0.004) (0.004) (0.004)Radio 0.009* 0.009* 0.008* 0.006* 0.006*

(0–4) (0.003) (0.003) (0.003) (0.003) (0.003)Television 0.003 0.003 0.003 0.004 0.004(0–4) (0.005) (0.004) (0.005) (0.004) (0.004)

Individual characteristicsTrust 0.015** 0.006(0–5) (0.004) (0.004)Efficacy 0.029*** 0.010(0–4) (0.009) (0.008)Knowledge 0.080*** 0.079***

(0–6) (0.004) (0.004)Controls

Income (Log) 0.099*** 0.095*** 0.095*** 0.093*** 0.068*** 0.068***

(0.008) (0.008) (0.008) (0.008) (0.008) (0.008)Education 0.028*** 0.028*** 0.027*** 0.027*** 0.017*** 0.017***

(0–12 years) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003)Ideology 0.022*** 0.019*** 0.017*** 0.016*** 0.001 0.001(0–7) (0.003) (0.003) (0.003) (0.004) (0.003) (0.003)Female −0.002 0.014 0.014 0.019 0.053*** 0.055***

(1 = female) (0.015) (0.014) (0.014) (0.014) (0.014) (0.014)Race – Black 0.051+ 0.045+ 0.036 0.052+ 0.111*** 0.108***

(1 = Black) (0.029) (0.027) (0.027) (0.027) (0.026) (0.027)Race – Other −0.088 −0.077 −0.075 −0.075 −0.067 −0.066(1 = other) (0.058) (0.056) (0.056) (0.056) (0.054) (0.054)

Intercept −0.515*** −0.581*** −0.543*** −0.601*** −0.575*** −0.566***

(0.084) (0.084) (0.085) (0.085) (0.080) (0.081)R2 0.1164 0.1648 0.1652 0.1670 0.2297 0.2298

N 1333 1333 1333 1333 1333 1333

+ p < 0.1.* p < 0.05.

**

it

iieceTiiiko

Fbaamemn

h

p < 0.01.*** p < 0.001.

ncreases the R2 by 6.49%, indicating the much better fit ofhis model in comparison with models 1–4.

The final full model, model 6, shows a strong and pos-tive effect of newspaper readership, a weak but positivempact of listening to the radio, and a non-significanceffect of watching television on voting. While knowledgeontinues to have a positive effect on voting, trust andxternal efficacy become non-significant in this full model.he overall explanatory power of this model is virtuallydentical to that of model 5, 22.97% versus 22.98%, indicat-ng that political knowledge and media use are the mostmportant variables explaining voting. In fact, politicalnowledge is the most important and significant predictorf voting in these analyses.

We highlight two main findings from these models.irst, media exposure is an important predictor of votingehavior. This is particularly true of reading newspaperslthough listening to political radio is also associated with

higher propensity to vote. Second, knowledge is theost important predictor of voting behavior. When trust,

xternal efficacy, and knowledge are included in the same

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

odel, knowledge is the only predictor that retains its sig-ificant effect.

The fixed effects models, presented in Table 2, showow life-course changes in media consumption and other

individual characteristics impact within-individual partic-ipation over time. In general, these models show the samerelationships as the random effects models, indicating thatthe same factors predict both who votes as well as an indi-vidual’s changing propensity to vote over time. Thus, forexample, having higher levels of political knowledge notonly makes one more likely to vote, but if an individualincreases their knowledge over time their propensity tovote also increases.

The value of the fixed effects model is shown in twoways. First, an individual’s changing use of political radiois not associated with a changing propensity to vote overtime, even though it is associated with who votes andwho does not (Table 1). This indicates that the relationshipbetween listening to the radio and voting is correlational.When an individual changes their patterns of listening, itdoes not increase the likelihood of voting. These two vari-ables are related, as those who listen to the radio also tendto vote, as in the random effects model. However, they donot appear to be causally linked. Second, there is a signifi-cant effect of trust on voting in the full fixed effects model.

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

While the level of trust is not related to who votes and whodoes not, as indicated in the full random effects model pre-sented in Table 1, there is a within individual effect of truston the propensity to vote.

Page 8: Media exposure and the engaged citizen: How the media shape political participation

8 C. Corrigall-Brown, R. Wilkes / The Social Science Journal xxx (2014) xxx–xxx

Table 2Linear panel models predicting voting behavior, fixed effects.

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

MediaNewspaper 0.011* 0.013* 0.015* 0.013* 0.014**

(0–4) (0.005) (0.005) (0.006) (0.005) (0.005)Radio −0.001 0.001 −0.001 −0.001 −0.001(0–4) (0.003) (0.003) (0.003) (0.003) (0.003)Television −0.006 −0.004 −0.005 −0.005 −0.003(0–4) (0.005) (0.005) (0.005) (0.005) (0.005)

Individual characteristicsTrust 0.025*** 0.016*

(0–5) (0.004) (0.005)Efficacy 0.018*** 0.016(0–4) (0.011) (0.011)Knowledge 0.083*** 0.080***

(0–6) (0.006) (0.005)Controls

Income (Log) 0.094*** 0.094*** 0.092*** 0.093*** 0.063*** 0.061***

(0.009) (0.009) (0.009) (0.009) (0.009) (0.009)Education 0.029*** 0.029*** 0.029*** 0.029*** 0.020*** 0.019***

(0–12 years) (0.003) (0.003) (0.003) (0.003) (0.004) (0.003)Ideology 0.004 0.004 0.003 0.004 0.001 −0.001(0–7) (0.005) (0.005) (0.005) (0.005) (0.004) (0.004)

Intercept −0.388*** −0.396*** −0.316*** −0.419*** −0.407*** −0.375***

(0.102) (0.104) (0.105) (0.105) (0.099) (0.101)R2 0.0922 0.1105 0.0984 0.1159 0.1927 0.1864

N 935 935 935 935 935 935

+ p < 0.1.* p < 0.05.

** p < 0.01.*** p < 0.001.

4.2. Protesting

The next set of models examines the effect of the sameindependent variables on protesting, and these modelsshow notably different results from the models predict-ing voting behavior (Table 3). Model 2 includes the mediaexposure measures. There is a positive and significant effectof both newspaper readership and listening to the radio onprotesting. An increase of one point on the 0–4 newspaperscale increases one’s odds of participating in protest by 1.12and an increase of 1 point on the 0–4 radio scale increasesone’s odds of participating in protest by 1.10. The amountof political television watched has no effect on the odds ofengaging in protest. As a whole, the three media variablesdo not dramatically alter the log likelihood of this model incomparison with model 1. Consequently, while the modelis a statistically significant improvement over the baselinemodel (1) (�2 test p < 0.01), the magnitude of the change issmall, indicating that the overall effect of exposure to themedia in predicting engagement in protest events is notlarge.

Model 3 shows that trust has a positive relationshipwith protesting. A one-point increase in trust increasesone’s chance of protesting by 0.81. Controlling for trustreduces the log likelihood of the model from −1069.70 to−1061.73 (�2 test p < 0.01), indicating a modestly betterfit of this model in comparison with the media exposure-

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

only model. Model 4 adds the external efficacy measure.This variable is positive and significant and a one-pointincrease in external efficacy increases the odds of protest-ing by 1.49. The relative change in the overall log likelihood

shows that external efficacy is almost twice as important asmedia exposure (newspapers and radio) or trust in predict-ing engagement in acts of protest. The inclusion of externalefficacy also reduces the direct effect of media exposure onprotesting. While in the original model without efficacy,increasing newspaper readership by one point increasedthe chance of protesting by 1.12 (p < 0.05), this effect isnow only 1.09 (p < 0.10). This indicates that the relation-ship between media use and external efficacy accounts fora portion of the effect of media use on protesting.

Model 5 adds the measure of political knowledge butit does not have a significant effect on protesting. Model 6includes all the variables and shows the positive impact ofnewspaper and radio exposure, trust, external efficacy, andknowledge on engagement in protest events. The smallerlog likelihood indicates that this model provides the bestfit and is a significant improvement on models 1–5 (�2 testp < 0.001).

In summary, the results presented in Table 3 show thatexposure to the media is a less important predictor ofprotesting than it is of voting. External efficacy has thestrongest effect of the three social–psychological variableson protest. Finally, trust is positively related to protest-ing and is an important differentiator between those whoprotest and those who do not.

The fixed effects models predicting protesting are pre-sented in Table 4. These models focus only on the factors

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

predicting the changing participation of individuals overtime. There are three principle findings from this table.First, the media variables are not significant predictorsof shifting engagement over time. While exposure to the

Page 9: Media exposure and the engaged citizen: How the media shape political participation

C. Corrigall-Brown, R. Wilkes / The Social Science Journal xxx (2014) xxx–xxx 9

Table 3Logistic panel models predicting protesting activity, random effect.

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

MediaNewspaper 0.117* 0.124* 0.095+ 0.109* 0.101*

(0–4) (0.051) (0.051) (0.051) (0.051) (0.051)Radio 0.100** 0.101** 0.088* 0.098** 0.089**

(0–4) (0.036) (0.036) (0.036) (0.036) (0.036)Television −0.055 −0.038 −0.050 −0.054 −0.031(0–4) (0.050) (0.051) (0.050) (0.050) (0.050)

Individual characteristicsTrust 0.206*** 0.219***

(0–5) (0.053) (0.052)Efficacy 0.399*** 0.421***

(0–4) (0.095) (0.097)Knowledge 0.065 0.015(0–6) (0.051) (0.052)

ControlsIncome (Log) −0.320*** −0.334*** −0.324*** −0.364*** −0.350*** −0.360***

(0.079) (0.078) (0.078) (0.078) (0.079) (0.079)Education 0.119*** 0.113*** 0.117*** 0.103*** 0.106*** 0.104***

(0–12 years) (0.029) (0.029) (0.030) (0.029) (0.030) (0.030)Ideology −0.399*** −0.378*** −0.362*** −0.340*** −0.359*** −0.316***

(0–7) (0.049) (0.049) (0.049) (0.049) (0.050) (0.051)Female −0.196 −0.132 −0.121 −0.064 −0.098 −0.042(1 = female) (0.136) (0.135) (0.134) (0.136) (0.137) (0.136)Race – Black 1.243*** 1.226*** 1.085*** 1.321*** 1.276*** 1.183***

(1 = Black) (0.215) (0.212) (0.212) (0.212) (0.214) (0.215)Race – Other 0.789 0.798+ 0.815+ 0.809+ 0.796+ 0.826+

(1 = other) (0.484) (0.477) (0.475) (0.475) (0.474) (0.472)Intercept −0.674 −0.823 −0.400 −1.136 −0.853 −0.694

(0.854) (0.852) (0.857) (0.847) (0.844) (0.850)Log likelihood −1076.89 −1069.70 −1061.73 −1060.79 −1068.91 −1051.54

N 1333 1333 1333 1333 1333 1333

+ p < 0.1.* p < 0.05.

**

mncteimaewdFopotbkdapc

Tt

p < 0.01.*** p < 0.001.

edia is associated with who will protest and who willot, this exposure is not associated with an individual’shanging propensity to protest over time, suggesting thathis relationship is not truly causal. In addition, the role ofxternal efficacy is different when only examining changesn protest within individuals, as is done in the fixed effects

odels. While external efficacy predicts which individu-ls will protest, it is not a significant predictor of changingngagement in protest over time. This is also true of trust,hich predicts who will protest and who will not, butoes not predict changing protest propensity over time.or both efficacy and trust, individuals with higher levelsf each of these features are more likely to participate inrotest. However, changes in these features of individualsver time are not associated with a changing propensityo engage in protest – they seem to prime participation,ut are not causally related to it over time. Finally, whilenowledge is not associated with who protests and whoid not, increasing political knowledge over time is associ-ted with a higher propensity to protest. An increase of oneoint on the 0–6 political knowledge scale increases one’s

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

hance of engaging in protest by 1.47%.How well do the findings conform to our hypotheses?

he first hypothesis is that media exposure is a more impor-ant predictor of voting than of protesting. The results

show this is the case. Newspapers are, as expected, themost important medium for predicting either voting orprotesting, with listening to the radio being the secondmost important medium for predicting each mode of par-ticipation. Television exposure is not related to eithervoting or protesting. It is important to note, however, thatmedia exposure is not a negative predictor of protesting.The second hypothesis is that political trust mediates therelationship between media exposure and political partic-ipation. The models do not support this hypothesis. Whiletrust is positively related to both voting and protesting,it is not a mediating factor in the relationship betweenmedia exposure and engagement. Our third hypothesis isthat external political efficacy mediates the effect of mediaexposure on engagement and is partially supported. Whileincreased feelings of external efficacy mediate the relation-ship between media exposure and protesting, they do notmediate the media–voting relationship. Our final hypoth-esis is that political knowledge mediates the relationshipbetween media and voting but not protesting. This hypoth-esis is supported. In sum, the results fail to show support

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

for one of our hypotheses but show partial or full supportfor the remaining three. While the media are an impor-tant predictor of political engagement, the routes throughwhich it shapes voting and protesting differ.

Page 10: Media exposure and the engaged citizen: How the media shape political participation

10 C. Corrigall-Brown, R. Wilkes / The Social Science Journal xxx (2014) xxx–xxx

Table 4Logistic panel models predicting protesting activity, fixed effects.

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

MediaNewspaper −0.041 −0.039 −0.041 −0.049 −0.046(0–4) (0.076) (0.076) (0.076) (0.076) (0.076)Radio 0.022 0.024 0.022 0.041 0.042(0–4) (0.051) (0.052) (0.051) (0.053) (0.054)Television 0.026 0.030 0.026 0.019 0.022(0–4) (0.070) (0.070) (0.070) (0.072) (0.072)

Individual characteristicsTrust 0.107 0.118(0–5) (0.079) (0.082)Efficacy −0.002 0.028(0–4) (0.152) (0.156)Knowledge 0.380*** 0.385***

(0–6) (0.083) (0.084)Controls

Income (Log) −0.375*** −0.369*** −0.359** −0.369** −0.227* −0.218+

(0.110) (0.110) (0.110) (0.111) (0.114) (0.114)Education −0.083+ −0.081+ −0.083+ −0.081+ −0.047 −0.049(0–12 years) (0.044) (0.044) (0.044) (0.044) (0.046) (0.046)Ideology −0.162** −0.163** −0.153** −0.164** −0.193*** −0.182**

(0–7) (0.064) (0.064) (0.065) (0.064) (0.065) (0.066)Log likelihood −261.84 −261.56 −260.64 −261.56 −250.68 −249.46

N 748 748 748 748 748 748

Table reports betas and (standard errors).+ p < 0.1.

* p < 0.05.

** p < 0.01.*** p < 0.001.

5. Conclusion

While a large body of research demonstrates that expo-sure to the mass media is a predictor of voting (Gastil &Xenos, 2010; Livingstone & Markham, 2008; McLeod et al.,1999), less is known about how the media shape otherpolitical behaviors, such as protesting. In this article, weassess the direct effect that media exposure has on par-ticipation in these two modes of political engagement andconsider the extent to which these relationships are medi-ated by trust, external efficacy, and knowledge. The dataused in this study is the only panel data set of its kind and assuch allows us to consider these relationships across threetime points.

The results demonstrate that any discussion of themedia–political participation relationship must be qual-ified by reference to the type of political participationexamined. Media exposure has a greater effect on insti-tutional than on non-institutional political engagement.17

In particular, while reading the newspaper and listeningto the radio are positively related to voting and protest-ing (Scheufele & Nisbet, 2002; Shah, Kwak, et al., 2001;Shah, McLeod, et al., 2001), the magnitude of this effectis much stronger for voting. In addition, while increasedindividual-level media consumption, particularly reading

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

the newspaper more often, increases the likelihood of vot-ing, it does not affect the likelihood of protesting over time.Thus, in the case of protesting, the media’s effect is not truly

17 Future research that employs a broader set of indicators of protestactivity would be useful to provide additional tests of the effect of thestrength of the protest measure.

causal, as changing an individual’s media consumption isnot related to a changing propensity to protest over time.

The media’s stronger relationship with voting could bethe result of differential coverage provided to these twomodes of political engagement (Boyle, McLuskey, McLeod,& Stein, 2005; Boyle & Schmierbach, 2009). The media aremuch more likely to cover electoral politics and, for thisreason, may be more likely to educate the public abouthow they can vote as opposed to how they could protest. Inaddition, media coverage of voting tends to be more posi-tive than its coverage of contentious participation (McLeod,2000; McLeod & Detenber, 1999). As a result, individualsreading the newspaper or listening to the radio may feelencouraged to vote but discouraged from engaging in actsof protest.

Our results also show that trust, external efficacy, andknowledge function differently as intermediaries betweenmedia exposure and the two types of political participation.Media exposure leads to voting because it increases politi-cal knowledge. This relationship held in both the random-and fixed-effects models, indicating that it is not simplythat those individuals who have more knowledge thatare more likely to vote but also that increasing politicalknowledge over time increases an individual’s propensityto vote. In contrast, the relationship between media use andprotesting is mediated by external political efficacy. Thisrelationship is, however, only significant in the random-effect models, indicating that while people with higher

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

levels of external efficacy are more likely to protest thanthose who feel less efficacious, an increased sense of exter-nal efficacy does not also increase the propensity to protestover time.

Page 11: Media exposure and the engaged citizen: How the media shape political participation

Social S

aousobtNtohtiatticw

C. Corrigall-Brown, R. Wilkes / The

The findings of this study provide important evidences to how media exposure is connected to two key formsf political participation: voting and protesting. The datased for these analyses consist of surveys of U.S. highchool seniors who were followed over a major periodf their lives from 1965 until 1997. Clearly, there haveeen large-scale changes in the media environment inhe years since the final wave of this data collection.ewspaper readership, the most important media predic-

or of engagement in this study, has and continues to ben the decline. The nature of television and radio newsas also changed. The rise of conservative talk shows onhe radio, such as Rush Limbaugh, and the 1996 found-ng of explicitly political television news stations, suchs MSNBC and Fox News, increasingly challenge the ideahat the media’s role is one of neutral observer. Finally,

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

he Internet differs from the traditional media examinedn this study in that the Internet newsreader has morehoice about what they consumes. Individuals frustratedith how institutional and non-institutional politics are

Author(s) Year Waves Media

Bouilanne 2011 3 Tv, online, print news

Boyle and Schmierbach 2009 1 Newspaper, tv, internet

Eveland et al. 2005 2 News use

Eveland and Scheufele 2000 1 Tv, newspaper use

Gastil and Xenos 2010 2 Public affairs media use

Holbert and Stephenson 2003 1 Debate viewing

Jung et al. 2011 1 News media use

Kenski and Jomini Stroud 2006 1 Tv, radio, paper, internet use

Kwiazek et al. 2010 1 General news media use

Livingstone and Markham 2008 1 Media and news consumptionO’Neil 2009 1 Tv, newspaper, ratio

Pinkleton et al. 1998 1 Active use of media for politicsShah 1998 1 Newspaper, broadcast viewingShah et al. 2001 2 Print, broadcast, internet mediScheufele 2002 1 Newspaper and tv public affairScheufele et al. 2002 1 Local media use

Scheufele et al. 2003 2 Newspapers and tv news

Sotirovic and McLeod 2001 2 Newspaper, tv entertainment

Stromback and Shehata 2010 3 News media use

Tolbert and McNeal 2003 3a Internet, tv, newspaper use

Zhang and Chia 2006 1 Newspaper, tv, internet use

a Analyzed separately.

cience Journal xxx (2014) xxx–xxx 11

covered by mainstream print, television, and radio newsmay be more likely to use the Internet for alternativeviews and opinions (Tolbert & McNeal, 2003). In fact, 29%of respondents who went online to get news and infor-mation about the 2000 election did so because they feltthat “you don’t get all the news and information you wantfrom traditional news sources such as the daily newspaperor networks news” (Pew Research Center for the People,2000 in Tolbert & McNeal, 2003). These new media andnews sources, including Twitter (Peterson, 2012), providemore diverse news content and could include more cov-erage, and more positive coverage, of the multiplicity ofways that citizens can engage in politics, from the insti-tutional to the contentious. Future work on later cohortsof individuals should test how these changes in the mediaenvironment have altered the relationship between mediaexposure and political participation, and future researchwith a more recent panel is needed to confirm whetherthe strong connection between media usage and politicalknowledge found with the Jennings data continues to hold.

Appendix A. Studies of media usage and politicalparticipation

Political participation Attitudes

Talk politics Political interestVoting, heavy protest –Political discussion Political knowledgeVoting, general participation Knowledge (via education)Voting, group involvement Civic faith, efficacy, civic virtueTraditional participation Political salienceOnline political participation Knowledge, efficacyTraditional participation Political efficacy and knowledgeCivic participation –

Voting and political actions Political efficacyCivic engagement General trust

Voting, group involvement Political efficacy, cynicism Civic engagement Trust in governmenta Civic participation –s Index of political participation Political knowledge and efficacy

Local political involvement –Political participation Knowledge, efficacy

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

Political participation KnowledgeInterest in politics –Voting Political interest, efficacyPolitical, civic participation Social capital

Page 12: Media exposure and the engaged citizen: How the media shape political participation

12 C. Corrigall-Brown, R. Wilkes / The Social Science Journal xxx (2014) xxx–xxx

Appendix B. Descriptive statistics by year,1973–1997

1973 1982 1997 Total

DependentVoting (0 = Never, 1 = All in period) 0.509 0.723 0.838 0.669

(0.34) (0.34) (0.27) (0.35)Protesting (0 = No, 1 = Yes) 0.15 0.07 0.09 0.11

(0.36) (0.25) (0.29) (0.31)Independent

Television (0 = Low, 4 = High) 3.16 3.00 2.87 3.03(1.21) (1.26) (1.29) (1.26)

Radio (0 = Low, 4 = High) 2.00 1.90 1.79 1.91(1.78) (1.78) (1.42) (1.76)

Newspaper (0 = Low, 4 = High) 2.89 2.86 2.87 2.87(1.42) (1.42) (1.42) (1.42)

Political trust (0 = Low, 5 = High) 2.99 2.53 2.53 2.71(1.39) (1.25) (1.18) (1.31)

Efficacy (0 = Low, 4 = High) 1.98 2.00 2.05 2.00(0.71) (0.76) (0.74) (0.72)

Political knowledge (0 = Low, 6 = High) 2.70 3.87 3.95 3.42(1.19) (1.38) (1.47) (1.46)

ControlsFamily income 1997 Dollars 42,021.45 53,618.17 76,018.35 55,172.21

(26,031.94) (29,923.58) (38,299.08) (33,958.48)Education (yrs. post-secondary) 0.67 1.04 0.95 1.13

(0.04) (0.40) (0.04) (0.32)Political ideology (0 = Left, 7 = Right) 4.56 4.76 4.85 4.76

(0.54) (0.45) (0.35) (0.46)Gender Female 50%Race White 91%

Black 7%Other 2%

Notes: Numbers in parentheses are standard deviations.

References

Abramson, P. R., & Aldrich, J. H. (1982). The decline of electoral participa-tion in America. American Political Science Review, 76, 502–521.

Acock, A., Clarke, H. D., & Stewart, M. C. (1985). A new model for old mea-sures: A covariance structure analysis of political efficacy. Journal ofPolitics, 47(4), 1062–1084.

Altheide, D. L., & Snow, R. P. (1991). Media worlds in the postjournalism era.New York, NY: Aldine/de Gruyter.

Barker, D. C. (1998). The talk-radio community: Non-traditional socialnetworks and political participation. Social Science Quarterly, 79(2),261–272.

Barker, D. C., & Knight, K. (2000). Political talk radio and public opinion.Public Opinion Quarterly, 64, 149–170.

Baum, M. A. (2003). Soft news goes to war: Public opinion and Americanforeign policy in the new media age. Princeton, NJ: Princeton UniversityPress.

Bélanger, E., & Nadeau, R. (2005). Political trust and the vote in multipartyelections. European Journal of Political Research, 44, 121–146.

Boulianne, S. (2011). Stimulating or reinforcing political interest. PoliticalCommunication, 28, 147–162.

Boyle, M. P., Mcluskey, M. R., Devanathan, N., Stein, S. E., & McLeod, D. M.(2004). The influence of level of deviance and protest type on coverageof social protest in Wisconsin from 1960 to 1999. Mass Communicationand Society, 7, 43–60.

Boyle, M. P., McLuskey, M. R., McLeod, D. M., & Stein, S. E. (2005). News-papers and protest: An examination of newspaper coverage of socialprotest from 1960 to 1999. Journalism and Mass Communication Quar-terly, 82, 638–653.

Boyle, M. P., & Schmierbach, M. (2009). Media use and protest. Communi-cation Quarterly, 57(1), 1–17.

Cho, J., Boyle, M. P., Keum, H., Shevy, M. D., McLoed, D. M., & Shah, D. V.(2003). Media, terrorism, and emotionality: Affective dimensions ofnews content and effects after September 11th. Journal of Broadcastingand Electronic Media, 47, 309–327.

Citrin, J., & Luks, S. (2001). Political trust revisited: Deja vu all over again?In J. R. Hibbing, & E. Theiss-Morse (Eds.), What is it About Governmentthat Americans Dislike? (pp. 9–27). New York: University of CambridgePress.

Corrigall-Brown, C., Snow, D. A., Smith, K., & Quist, T. (2009). Explainingthe puzzle of homeless mobilization: An examination of differentialparticipation. Sociological Perspectives, 52(3), 309–335.

Corrigall-Brown, C. (2011). Patterns of protest: Trajectories of par-ticipation in social movements. Stanford, CA: Stanford UniversityPress.

Corrigall-Brown, C. (2012). From the balconies to the barricades, and back?Trajectories of participation in contentious politics. Journal of CivilSociety, 8(1), 17–38.

Corrigall-Brown, C., & Wilkes, R. (2012). Picturing protest: The visual fram-ing of collective action by first nations in Canada. American BehavioralScientist, 56(2), 223–243.

Dahl, R. A. (1971). Polyarchy: Participation and opposition. Chelsea, MI: YaleUniversity Press.

Dalton, R. J., Van Sickle, A., & Weldon, S. (2009). Theindividual–institutional nexus of protest behavior. British Journal ofPolitical Science, 28(5), 931–951.

De Vreese, C. H., & Semetko, H. A. (2002). Cynical and engaged: Strategiccampaign coverage, public opinion, and mobilization in a referendum.Communication Research, 29(6), 615–641.

Delli Carpini, M. X., & Keeter, S. (2003). The Internet and an informed cit-izenry. In D. M. Anderson, & M. Cornfields (Eds.), The civic web: Onlinepolitics and democratic values (pp. 129–153). Lanham, MD: Rowan and

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

Capella, J. N., & Jamieson, K. H. (1996). News frames, political cynicism,and media cynicism. Annals of the American Academy of Political andSocial Science, 546, 71–85.

Chaffee, S. H., & Kanihan, S. F. (1997). Learning about politics from themass media. Political Communication, 14, 421–430.

Chaffee, S. H., Zhao, X., & Leshner, G. (1994). Political knowledge and thecampaign media of 1992. Communication Research, 21, 305–324.

Littlefield.Entman, R. N. (2005). The nature and sources of news. In G. Overholser,

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

& K. H. Jamieson (Eds.), The press (pp. 48–65). New York, NY: OxfordUniversity Press.

Eveland, W. P., Hayes, A. F., Shah, D. V., & Kwak, N. (2005). Understand-ing the relationship between communication and political knowledge.Political Communication, 22, 423–446.

Page 13: Media exposure and the engaged citizen: How the media shape political participation

Social S

E

F

G

G

G

G

G

H

H

H

H

H

H

I

J

J

J

K

K

K

K

K

L

L

L

M

M

M

M

C. Corrigall-Brown, R. Wilkes / The

veland, W. P., & Scheufele, D. A. (2000). Connecting news media usewith gaps in knowledge and participation. Political Communication,17, 215–237.

inkel, S. E. (1985). Reciprocal effects of participation and political effi-cacy: A panel analysis. American Journal of Political Science, 29(4),891–913.

astil, J., & Xenos, M. (2010). Of attitudes and engagement. Journal ofCommunication, 60, 318–343.

ecas, V. (2000). Value identities, self motives and social movements. InS. Stryker, T. J. Owens, & R. W. White (Eds.), Self, identity, and socialmovements (pp. 93–109). Minneapolis, MN: University of MinnesotaPress.

iltin, T. (1980). The whole world is watching: Mass media in the making andunmaking of the New Left. Berkeley, CA: University of California Press.

ross, K., Aday, S., & Brewer, P. R. (2004). A panel study of media effectson political and social trust after September 11, 2001. Harvard Inter-national Journal of Press/Politics, 9(4), 49–73.

rossman, L. K. (1995). The electronic republic: Reshaping democracy in theinformation age. New York, NY: Viking.

adjar, A., & Beck, M. (2010). Who does not participation in elections inEurope and why is this? European Societies, 12(4), 521–542.

allin, D. C. (1992). Sound bite news: Television coverage of elections1968–1988. Journal of Communication, 42(spring), 5–24.

art, R. P., Smith-Howell, D., & Llewellyn, J. (1990, October–December).Evolution of presidential news coverage. Political Communication andPersuasion, 7, 213–230.

etherington, M. (1998). The political relevance of political trust. AmericanPolitical Science Review, 92(4), 791–807.

etherington, M., & Husser, J. (2012). How trust matters: The changingpolitical relevance of political trust. American Journal of Political Sci-ence, 56, 312–325.

onaker, J., King, G. & Blackwell, M. Version 1.2-15. Amelia program.Gking.harvard.edu/Amelia.

yengar, S. (1991). Is anyone responsible? How television frames politicalissues. Chicago, IL: University of Chicago Press.

ennings, M. K. (2002). Generational units and the student protest move-ment in the United States: An intra- and inter-generational analysis.Political Psychology, 23(2), 303–324.

ennings, M. K., & Stoker, L. (2004). Youth-parent socialization panelstudy, 1965–1997: Youth wave IV, 1997. Retrieved from the NationalArchive of Computerized Data on Aging Web Site: http://www.icpsr.umich.edu/icpsrweb/NACDA/studies/4037/detail

ung, N., Kim, H., & Homero, G. (2011). The mediating role of knowledgeand efficacy in the effects of communication on political participation.Mass Communication and Society, 14, 407–430.

aniss, L. L. (1991). Making local news. Chicago, IL: The University ofChicago Press.

enski, K., & Jomini Stroud, N. (2006). Connections between Internet useand political efficacy, knowledge, and participation. Journal of Broad-casting and Electronic Media, 50, 173–192.

im, Y. C., & Ball-Rokeach, S. J. (2006). Community storytelling network,neighbourhood context, and civic engagement: A multilevel approach.Human Communication Research, 32, 411–439.

landermans, B., van der Toorn, J., & van Stekelenburg, J. (2008). Embed-dedness and identity: How immigrants turn grievances into action.American Sociological Review, 73(6), 992–1012.

leinnijenhuis, J., van Hoof, A. M. J., & Oegema, D. (2006). Negative newsand the sleeper effect of distrust. Harvard International Journal ofPress/Politics, 11(2), 86–104.

awson, C., & McCann J.A.F J.A. (2005). Television news Mexico’s 2000elections and media effects in emerging democracies. British Journalof Political Science, 35(1), 1–30.

ichter, S. R., Lichter, L. S., & Rothman, S. (1994). Prime time: How TV por-trays American culture. Washington, DC: Regency Publishing.

ivingstone, S., & Markham, T. (2008). The contribution of media coverageto civic participation. British Journal of Sociology, 59(2), 351–371.

cCarthy, J. D., McPhail, A., & Smith, J. (1996). Images of protest:Dimensions of selection bias in media coverage of Washingtondemonstrations, 1982 to 1991. American Sociological Review, 61(3),478–499.

cLeod, D. M. (2000). The protest paradigm and news coverage of the‘right to party’ movement. In D. Schultz (Ed.), It’s show time (pp. 29–50).New York: Peter Lang.

cLeod, D. M., & Detenber, B. (1999). Framing effects of television news

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

coverage of social protest. Journal of Communication, 49, 3–23.cLeod, D. M., & Hertog, J. K. (1992). The manufacture of ‘public opinion’

by reporters: Informal cues for public perceptions of protest groups.Journal of Communication, 49, 3–23.

cience Journal xxx (2014) xxx–xxx 13

McLeod, J. M., Scheufele, D. A., & Moy, P. (1999). Community, communi-cation, and participation: The role of mass media and interpersonaldiscussion in local political participation. Political Communication,16(3), 315–336.

McQuail, D. (1998). New roles for new times? In E. E. Dennis, & R. W.Snyder (Eds.), Media and democracy (pp. 13–20). New Brunswick, NJ:Transaction Publishers.

McVeigh, R., & Smith, C. (1999). Who protests in America: An analysisof three political alternatives – Inaction, institutionalized politics, orprotest. Sociological Forum, 14, 685–702.

Meyrowitz, J. (1985). No sense of place. New York, NY: Oxford UniversityPress.

Miller, A. H. (1974). Political issues and trust in government: 1964-1970.The American Political Science Review, 68(3), 951–972.

Miller, W. E., Miller, A. H., & Schneider, E. J. (1980). American nationalelection studies data sourcebook 1952–1978. Cambridge, MA: HarvardUniversity Press.

Moy, P., & Scheufele, D. A. (2000). Media effects on political andsocial trust. Journalism and Mass Communication Quarterly, 77(4),744–759.

Newton, K. (1999). Mass media effects: Mobilization or media malaise?British Journal of Political Science, 29, 577–599.

Nie, N. H., Junn, J., & Stehlik-Barry, K. (1996). Education and democraticcitizenship in America. Chicago, IL: University of Chicago Press.

Norris, P. (2002). Social capital and the news media. Harvard InternationalJournal of Press/Politics, 7(1), 3–8.

Nyækowiak, J. (2009). Political activity: Is trust in democratic institutionsreally a relevant determinant? International Journal of Sociology, 39(1),49–61.

Patterson, T. E. (1993). Out of order. New York, NY: Knopf.Patterson, T. E. (2001). Doing well and doing good. Cambridge, MA: Shoren-

stein Center for Press, Politics, and Public Policy.Patterson, T. E., & McClure, R. D. (1976). The unseeing eye: The myth of

television power in national elections. New York, NY: G. P. Putnam.Peterson, R. D. (2012). To tweet or not to tweet: Exploring the determi-

nants of early adoption of Twitter by House members in the 111thCongress. Social Science Journal, 49(4), 430–438.

Pew Research Center for the People and the Press. (2000). Campaign andInternet survey. October 10–November 25, N = 1,435.

Pinkleton, B. E., Weintraub, A., & Fortman, K. (1998). Relationships ofmedia use and political disaffection to political efficacy and votingbehavior. Journal of Broadcasting and Electronic Media, 42, 34–49.

Putnam, R. D. (2000). Bowling alone: The collapse and revival of Americancommunity. New York, NY: Simon and Schuster.

Rabe-Hesketh, S., & Skrondal, A. (2008). Multilevel and longitudinal model-ing using Stata (2nd ed.). College Station, TX: Stata Press.

Shaffer, S. (1981). A multivariate explanation of decreasing turnout inpresidential elections 1960–1976. American Journal of Political Science,25, 68–69.

Scheufele, D. A., & Nisbet, M. C. (2002). Being a citizen online: New oppor-tunities and dead ends. Harvard International Journal of Press/Politics,7(3), 55–75.

Scheufele, D. A., Shanahan, J., & Sei-Hill, K. (2002). Who cares aboutlocal politics? Journalism and Mass Communication Quarterly, 79(2),427–444.

Schussman, A., & Soule, S. A. (2006). Process and protest: Accounting forindividual protest participation. Social Forces, 84, 1083–1108.

Shah, D. V., Cho, J., Eveland, W. P., & Kwak, N. (2005). Information andexpression in a digital age: Modeling Internet effects on civic partici-pation. Communication Research, 32, 531–565.

Shah, D. V., Kwak, N., & Holbert, R. (2001). ‘Connecting’ and ‘disconnecting’with civic life. Political Communications, 18, 141–162.

Shah, D. V., McLeod, J. M., & Yoon, S. H. (2001). Communication, context,and community. Communication Research, 28, 464–506.

Smith, J., McCarthy, J. D., McPhail, C., & Augustyn, B. (2001). From protest toagenda building: Description bias in media coverage of protest eventsin Washington, DC. Social Forces, 79(4), 1397–1423.

Smith, H. H., III. (1986). Newspaper readership as a determinant of politicalknowledge and activity. Newspaper Research Journal, 7, 47–54.

Sotirovic, M., & McLeod, J. M. (2001). Values, communication behavior,and political participation. Political Communication, 18(3), 273–300.

Stephen, C. C., & Maggiotto, M. A. (1981). Political discontent and politicalaction. Journal of Politics, 43(2), 514–522.

Tolbert, C. J., & McNeal, R. S. (2003). Unraveling the effects of the

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

Internet on political participation. Political Research Quarterly, 56(2),175–185.

Verba, S., Schlozman, K. L., & Brady, H. E. (2002). Voice and equality: Civicvoluntarism in American politics. London, UK: Harvard University Press.

Page 14: Media exposure and the engaged citizen: How the media shape political participation

Social S

14 C. Corrigall-Brown, R. Wilkes / The

Waldner, L. (2001). Lesbian and gay political activism: An analysis of vari-ables predicting political participation. Research in Political Sociology,9, 59–81.

Wattenberg, M. P. (1992). The rise of candidate-centered politics. Cam-bridge, MA: Harvard University Press.

Weaver, D. A., & Scacco, J. M. (2013). Revisiting the protest paradigm.

Please cite this article in press as: Corrigall-Brown, C., & Wilkemedia shape political participation. The Social Science Journal (2

International Journal of Press/Polity, 18(1), 61–84.Weldon, S., & Dalton, R. J. (2014). Democratic structures and democratic

participation: The limits of consociational theory. In J. Thomassen(Ed.), Elections and democracy: Representations and accountability.Oxford, UK: Oxford University Press.

cience Journal xxx (2014) xxx–xxx

Wilkes, R. (2011). Re-thinking the decline in trust: A compari-son of black and white. Americans Social Science Research, 40,1596–1610.

Wilkes, R., Corrigall-Brown, C., & Ricard, D. (2010). Nationalism and mediacoverage of Indigenous People’s collective action. American Indian Cul-ture and Research Journal, 34, 41–59.

s, R. Media exposure and the engaged citizen: How the014), http://dx.doi.org/10.1016/j.soscij.2014.03.009

Wooldridge, J. M. (2001). Econometric analysis of cross section and paneldata. Cambridge, UK: MIT Press.

Yeich, S., & Levine, R. (1994). Political efficacy: Enhancing the constructand its relationship to mobilization of people. Journal of CommunityPsychology, 22, 259–271.