Agenda Setting in the Partisan TV News Context

Abstract

This study examines the agenda setting of candidates’ attributes and its relationship with polarized candidate evaluation among TV news viewers. Content analyses of candidates’ affective attributes during the 2012 presidential election indicate partisan imbalance from CNN’s Anderson Cooper and Fox’s Special Report. NBC Nightly News was relatively balanced. Watching a particular program was positively associated with attribute agenda setting by each program. Also, agenda setting by the Fox program was positively related to viewers’ polarized candidate evaluations, whereas agenda setting by the NBC program was negatively associated. Implications of the partisan TV news context for agenda-setting theory are discussed.

Keywords

election partisan news polarization selective exposure

Attribute agenda-setting research has documented that attributes or traits of candidates emphasized in the news media influence voters’ images about those candidates (e.g., Coleman & Wu, 2010; Golan & Wanta, 2001; Kim & McCombs, 2007; King, 1997; McCombs, Llamas, Lopez-Escobar, & Rey, 1997; McCombs, Lopez-Escobar, & Llamas, 2000; Wu & Coleman, 2009). The current news environment, characterized by partisan selective exposure, however, requires researchers to consider potentially distinct agenda setting that may vary according to news sources’ political orientations. Also, the assumed partisan filtering by individual news outlets may influence subsequent effects of agenda setting on audience attitudes toward candidates.

To explore implications of the partisan news context for agenda-setting theory, this study investigates attribute agenda-setting influence and its relationship with subsequent candidate evaluation by analyzing broadcast and cable TV news and their audiences. TV news, in particular, merits scholarly attention as it has undergone significant partisan self-selection and polarization (Hollander, 2008; Iyengar & Hahn, 2009; Pew Research Center, 2009; Stroud, 2010). Continuing public reliance on TV news as a main source of electoral information, despite the growing prominence of the Internet (Pew Research Center, 2012), further justifies examination of its agenda-setting effects.

The first goal of this study is to investigate separate agenda setting of candidate attributes by different TV news outlets. A growing body of data indicates contrasting partisan imbalance in coverage of candidates among different TV news sources, particularly cable news (Farnsworth & Lichter, 2011; Morris & Francia, 2010; Project for Excellence in Journalism, 2004, 2008, 2012) that may lead to splitting of the agenda-setting process. Combining content analysis and survey data from the 2012 U.S. presidential election, this study examines how TV news programs from NBC, CNN, and Fox News differently emphasized affective attributes of opposing presidential candidates and whether those attributes that are salient in particular programs influenced their viewers’ perception of candidates.

The second goal of the current study is to address the potential role of attribute agenda setting in polarizing audience evaluation of competing candidates as a part of broader attitudinal outcomes of agenda setting. When polarization refers to divergence of opinions toward partisan or ideological extremes (Mutz, 2002; Stroud, 2010), the concept involves directional strength of attitudes. This polarization concept connects with the literature about attitudinal consequences of agenda setting, which has illustrated that agenda-setting effects can strengthen as well as direct public opinion about candidates, either positively or negatively (Balmas & Sheafer, 2010; Kiousis, 2005, 2011; Kiousis & McCombs, 2004; Moon, 2013). As a certain news source consistently highlights positive attributes of one candidate while emphasizing negative attributes of the other, audience attitudes toward the candidates may strengthen in bipolar directions, leading to polarized attitudes.

This study tests that possibility by linking affective agenda setting of candidate attributes to polarized audience evaluation of the candidates. By doing so, we aim to demonstrate that polarizing or depolarizing effects of TV news programs rely on viewers’ susceptibility to the effects of the affective dimension present in specific programs, rather than on simple exposure to those programs. Three TV news programs were selected for analysis: NBC Nightly News, CNN’s Anderson Cooper 360, and Fox News’ Special Report with Bret Baier. For a rigorous test, an individual-level analysis was used to determine whether agenda setting and its subsequent effects persist even after controlling for viewers’ political predispositions and general news use variables.

Partisan TV News

Selective exposure to likeminded news and political information has been suggested as one of the main reasons for rearrangement of partisan compositions of TV news viewers among different channels. As people tend to avoid cognitive dissonance, which arises when they encounter information that disagrees (Festinger, 1957; Klapper, 1960), they likely seek news sources congenial to their existing political orientations. In a highly competitive TV news market, individual channels are motivated to capitalize on this tendency by strategically aligning news content with the political predispositions of targeted audiences (Baron, 2006; Gentzkow & Shapiro, 2006). Accordingly, Republican and conservative viewers have continued to move to Fox News, whereas Democrats and liberals increasingly select CNN and MSNBC (Hollander, 2008; Pew Research Center, 2009).

The presumed partisan and ideological slants of some news sources have become a cause for concern among commentators who worry about partisan intervention into standard norms of objective journalism. Criticism of news bias is more commonly aimed at Fox News (Alterman, 2004; Brock, 2004; Kitty, 2005), but CNN (Bozell, 2004) and MSNBC (Stanley, 2012) also receive bias accusations. More systematic content analyses lend some support to criticism of partisan inclinations in cable news. Fox News in particular was reported to select news items responsive to Republican or conservative positions and ideologies more frequently than news wires such as the Associated Press and Reuters (Baum & Groeling, 2008). In reporting politically divisive issues, Fox News was more favorable to the Bush administration’s war efforts in Iraq than was NBC (Aday, 2010), CNN, or MSNBC (Muddiman, Stroud, & McCombs, 2014), and more suspicious of claims of climate change than CNN and MSNBC (Feldman, Maibach, Roser-Renouf, & Leiserowitz, 2012).

Similar partisan inclinations in TV news were also noted in electoral contexts, especially when such inclinations were measured in terms of evaluative tones given to competing candidates. According to a series of content analyses (Project for Excellence in Journalism, 2004, 2008, 2012), Fox News tended to be more favorable to Republican than Democratic candidates, with this tendency reversed on CNN and MSNBC. One noteworthy pattern is that the partisan leaning of Fox was relatively stable regardless of the distinctive atmospheres of individual election cycles. During the 2008 election, for example, TV news in general gave more negative coverage to McCain than Obama. Even Fox gave a considerable amount of unfavorable coverage to McCain, but the channel was still less critical of McCain relative to other news channels (Farnsworth & Lichter, 2011; Project for Excellence in Journalism, 2008).

In sum, a growing body of evidence points to considerable partisan and ideological divergence in both content and audience of TV news. The shifting landscape of TV news has generated much research interest regarding its impact on viewers’ political opinions and attitudes (e.g., Morris & Francia, 2010; Smith & Searles, 2014; Stroud, 2010). Given that partisan imbalance is most evident in tonal aspects of election coverage, the attribute agenda-setting process, including an affective dimension as an integral part of the extended agenda-setting process, can provide a theoretical framework to explain the influence of partisan TV news viewing on the electorate.

Attribute Agenda Setting in Electoral Contexts

The basic hypothesis of attribute agenda setting posits that attributes emphasized in the news become prioritized in the public’s mind in much the same way that traditional agenda setting works. While traditional agenda setting deals with the salience of an object, attribute agenda setting focuses on traits and properties that comprise an object, whether a political issue, event, or figure. Attributes are further divided into substantive and affective dimensions: The substantive dimension refers to cognitive elements constituting an object, whereas the affective dimension means evaluative components involving positive, negative, or neutral tones (McCombs, 2014; McCombs et al., 1997; McCombs et al., 2000).

Based on the framework of attribute agenda setting, studies have demonstrated that news media shape the public’s images of political candidates. When news media differ in emphasizing various attributes or traits of candidates, people tend to respond to the ordering of attributes presented by the media. Although attribute agenda setting has been replicated in many studies, the current partisan news situation has rarely been considered. As audience options for news sources with distinctive political orientations proliferate, the media agenda of a few elite news sources may no longer serve as a reliable proxy to measure the overall media agenda (Chaffee & Metzger, 2001).

Even before the rise of partisan news, correlations between agendas of different news media have never been perfect. In their seminal agenda-setting research, McCombs and Shaw (1972) noted that a lack of consensus of media agendas might arise from the “point of view” or even “extreme bias” of each news medium (p. 184). This observation suggests that editorial policy and culture of a particular news organization can cause dissimilarity of news agendas. This perspective coincides with the hierarchy of influences model, proposing that organizational characteristics can lead to variations in news content among different news outlets (Shoemaker & Reese, 2014).

A partisan TV news environment may further increase organizational influence on the news-making process that could undermine intermedia similarity of agendas. Balanced reporting is more loosely practiced in cable news, with more single-viewpoint news present in cable news than in broadcast news (Project for Excellence in Journalism, 2005). Across all three cable news channels, the amount of opinion-based news has continued to rise at the expense of factual reporting (Project for Excellence in Journalism, 2013). Deviation from fact-oriented, balanced reporting allows more room for partisan or ideological interpretation in the news.

These changes in the TV news landscape establish a need for agenda-setting research that involves a more refined measurement of media and public agendas across different news outlets and their audiences. Supporting this point, the magnitude of attribute agenda setting tends to be greater when candidate attributes of a specific news source are paired with attributes of that source’s audience rather than with other sources’ audience (McCombs et al., 2000). Experiment-based agenda-setting research clearly supports the idea that agenda-setting effects arise from the content of news messages to which individuals are exposed. Given evidence indicating different emphases of tonal aspects of candidate image by different TV news outlets (Farnsworth & Lichter, 2011; Morris & Francia, 2010; Project for Excellence in Journalism, 2004, 2008, 2012), agenda-setting effects depend on the particular TV news programs on which individuals rely. Muddiman et al. (2014) confirmed the separate attribute agenda setting of the three cable news channels by demonstrating that different emphases of affective tones about the Iraq War on the three channels were related to a viewer’s overall positive–negative opinions about the war. Therefore, we postulated the following hypothesis:

H1: Viewing a particular news program will be positively associated with the attribute agenda setting of the program.

Attribute Agenda Setting and Attitude Polarization

Partisan selective exposure has generated much research interest regarding its impact on political polarization (Mutz, 2006; Prior, 2013; Sunstein, 2001). Some data based on a panel survey (Stroud, 2010) and experiment (Jones, 2002; Taber & Lodge, 2006) show that a causal direction flows from partisan selective exposure to attitude polarization. What is lacking in these studies, though, is attention to the content of news messages that may cause polarization in real-world contexts.

Research on attitudinal consequences of agenda setting complements the extant literature by taking potentially polarizing news content into account. Polarization is often defined as divergence of political opinion toward partisan or ideological extremes that is operationalized by folding continuous measures of political attitudes (Mutz, 2002; Stroud, 2010). This definition suggests that the concept includes strength as well as direction of opinion pertaining to a certain political object. The concept includes tonal direction as it involves opposition or conflict of opinion among the public regarding political orientations, issues, or figures. Polarization is also a manifestation of attitude strength, particularly attitude extremity, identified at the bipolar ends of positive or negative evaluation (Abelson, 1995).

These aspects of strength and direction inherent in polarization connect with agenda-setting research, which has documented the strengthening and directing of opinion as major attitudinal outcomes of agenda setting. News attention to a certain object generates more frequent and engaged thinking, which functions to strengthen opinions and attitudes about the object (McCombs & Reynolds, 2002; Weaver, 1984). Empirical studies supported this hypothesis by confirming that media salience is positively correlated with various dimensions of attitude strength, such as opinion strength about political issues (Weaver, 1991), opinionation, and extreme attitudes about political figures (Kiousis, 2011; Kiousis & McCombs, 2004). More relevant to the current study, the role of media salience in inducing stronger attitudes has also been observed in the attribute agenda setting of candidates (Kiousis, 2005; Moon, 2013).

Furthermore, studies that analyzed the influence of valence of media agenda discovered that news attention can have an impact on the direction of public opinion. Attribute agenda-setting research elaborates directional consequences of agenda setting by incorporating the affective element as an essential part of the agenda-setting process. Specifically, affective agenda setting posits that tones about particular attributes in the news can guide public opinion in either positive or negative ways. Mostly conducted in electoral contexts, prior research confirmed that tones of political figures in the news corresponded to public evaluation of those figures (Coleman & Wu, 2010; Kim & McCombs, 2007; McCombs et al., 1997; McCombs et al., 2000; Wu & Coleman, 2009). In addition, through the affective priming process, the valence of candidate attributes in the news predicted affective salience among the public, which in turn related to public judgment of candidates’ suitability for public office (Balmas & Sheafer, 2010).

Within the context of partisan TV news, these prior studies imply that partisan imbalance of certain programs may cause polarized attitudes toward competing candidates. That is, when cable news channels consistently deliver slanted news, offering positive attributes of one candidate while emphasizing negative attributes of the other, audience attitudes may move toward bipolar ends. However, simple exposure to partisan news sources may not be a sufficient condition for partisan news to induce polarized attitudes. Even among viewers who watch the same partisan source, the level of attitude polarization may vary beyond political predispositions: Some viewers will be more affected by the evaluative tones presented in news than others.

Attribute agenda setting can tap into individual levels of susceptibility to the influence of watching particular news programs. In an aggregate-level analysis, media salience was found to affect individuals’ attitude strength, including polarized feelings toward competing candidates (Kiousis, 2011; Kiousis & McCombs, 2004). Furthermore, an individual-level analysis discovered that attribute agenda setting mediated the relationship between individuals’ news use and attitude strength about candidates (Moon, 2013). In the current TV news environment characterized by partisan inclinations of cable news channels, it can be expected that agenda setting by partisan news programs will contribute to the strengthening of candidate evaluations in bipolar directions. Specifically, those who are more susceptible to agenda setting by cable news should favor one candidate more strongly and, at the same time, have greater negative feelings toward the other. However, susceptibility to attribute agenda setting by broadcast TV news should depolarize viewer attitudes because relatively balanced coverage of candidates may cancel out the effects of opposing tones of different candidates. Therefore, we postulated the following two hypotheses:

H2: Agenda setting of candidates’ attributes by CNN’s Anderson Cooper and Fox’s Special Report will be positively related to polarized attitudes toward candidates.

H3: Agenda setting of candidates’ attributes by NBC Nightly News will be negatively related to polarized attitudes toward candidates.

Method

Public Opinion Data

The 2012 American National Election Studies (ANES) data were analyzed. The cross-sectional survey used a face-to-face method and the Internet to interview a total of 5,916 respondents. All variables used in this study were obtained from the preelection wave. Notably, the 2012 data cover a wide array of questions regarding respondents’ news media use. Questionnaires included the use of specific newspapers and TV and radio programs, in addition to conventional questions about general news media use. Statistical analysis in the current research relied upon weighted data to adjust the difference between the sample and national demographic characteristics.

Content Analysis

Sampling

News programs on the three TV channels aired in July and August 2012 were analyzed: NBC Nightly News, CNN’s Anderson Cooper 360, and Fox News’ Special Report with Bret Baier. The time frame of sampling, covering 2 months prior to the beginning of the ANES survey, was based on the time lag of agenda-setting effects, observed with 4 to 8 weeks as the optimal period (McCombs, 2014).

Despite the proliferation of Internet news in recent years, TV continues to be the main news source for Americans. Pew Research Center (2012) reported that 67% of voters relied on TV as the main source of campaign news, and more people selected cable news rather than broadcast news as their primary source. Among broadcast TV news channels, NBC had more viewers than its competitors. Because cable channels do not have typical primetime news programs like broadcast TV does, Special Report was selected as the flagship news program of Fox News (Groseclose & Milyo, 2005). Anderson Cooper was the only CNN news program included in ANES data.

Stories were collected using keyword searches from transcripts of LexisNexis Academic (http://www.lexisnexis.com). Keywords were Obama, president, Romney, former Massachusetts governor, and former governor.

Coding

Attributes of Obama and Romney and respective tones were analyzed. The coding scheme was matched with ANES questionnaires asking about the two candidates’ personal traits. While ANES inquired about six personal attributes of each candidate—particularly morality, leadership, care, knowledge, intelligence, and honesty—we combined questions about knowledge and intelligence in the same category as they were not clearly distinguished in pilot coding. Responses from the survey showed a high correlation between the two items at the .001 level (Obama, r = . 80; Romney, r = .76). Other interrelated attributes were morality and honesty. In today’s electoral context, however, morality represents an attribute related to fundamental and religious values. Typically, issues involving moral values in elections are gay marriage, abortion, and religion (Hillygus & Shields, 2005; Lovett & Jordan, 2005). Hence, if an article evaluated the candidates based on those issues, it was coded as morality. Honesty, however, is a broader concept than morality, at least in the election. When mentions were made, such as “he is ethical” or “he is truthful,” those were coded as honesty.

Next, the affective tone of each mention of an attribute was coded by determining whether it was positive or negative. Neutral tone was deleted after pilot coding because it was difficult to identify an attribute directly related to the evaluation of candidates as being neutral in the electoral context. Coders were asked to judge the tones based on the impression they would receive from the perspective of TV viewers. The unit of analysis was a mention. When a sentence included multiple attributes, all of them were coded separately. Intercoder reliability between two coders was checked based on a 10% subsample. The overall agreement was 96%. Following are the values of Scott’s pi for each variable: Obama’s morality (.74), leadership (.75), care (.90), intelligence (.87), and honesty (.86); Romney’s morality (.80), leadership (.80), care (1), intelligence (1), and honesty (1).

Measures

TV news viewing

Respondents were first asked whether “they heard about presidential campaign on TV news/talk/public affairs/news analysis programs.” If they gave a positive answer, they were further instructed to check the TV programs they watch regularly, at least once a month. Viewership of CNN’s Anderson Cooper (n = 480), Fox’s Special Report (n = 288), and NBC Nightly News (n = 1,216) was measured separately (1 = watching the program, 0 = not watching).

Attribute agenda-setting index

To examine the individual level of the agenda-setting effects of candidates’ affective attributes, correspondence between the program’s affective tones and respondents’ evaluations of candidates’ attributes was calculated. ANES asked respondents how well each statement about a candidate’s attributes or traits (e.g., “he is moral”) described the candidate on a 5-point scale (1 = extremely well, 5 = not well at all). After combining the attributes of knowledge and intelligence, five attributes for each candidate were obtained. Each score of the total 10 attribute items was reversed for alignment with attribute scores of the programs.

Each TV channel’s candidate attribute scores were calculated using the Janis–Fadner Coefficient of Imbalance (Janis & Fadner, 1943). The Janis–Fadner method works well for the operationalization of candidate attributes because it incorporates both strength (frequency) and direction (tone) of the message, which are two important dimensions of attribute agenda setting (Sheafer, 2007). The coefficient ranges from −1 when all candidate attributes in a particular program are negative to +1 when all attributes are positive.¹

The final step was determining the level of correspondence of the 10 attribute scores between a particular TV program and a respondent’s evaluation using rank-order correlation analysis which has been the conventional approach in agenda-setting research since the original study by McCombs and Shaw (1972). Each respondent had three values of Spearman’s rho because there were three matching news programs. A higher value of rho indicated a closer match between candidates’ affective attributes as presented by the program and respondents’ perception of those attributes.

Polarization

Based on prior research (Mutz, 2002; Stroud, 2010), polarized attitudes toward candidates were measured by calculating the difference of 100-degree thermometer feelings toward the two candidates, with higher degrees indicating more favorable feelings. A feeling thermometer has been employed to assess attitude strength in previous agenda-setting research (Kiousis, 2005, 2011; Kiousis & McCombs, 2004). Attitude polarization was an absolute value obtained by subtracting Obama’s rating from Romney’s, with a possible range from 0 to 100 (M = 53.49, SD = 30.09). If a respondent felt the same degree of feeling for both candidates, the polarization score was 0. If a respondent felt 100 degrees for Obama and 0 degree for Romney (or vice versa), the score was 100.

Control variables

News attention was measured on a 5-point scale (1 = none, 5 = a great deal). To separate effects of the individual news programs from those of general news use, attention to TV news (M = 2.78, SD = 1.54), newspaper (M = 1.72, SD = 1.67), Internet news (M = 2.15, SD = 1.61), and radio news (M = 1.84, SD = 1.71) was controlled.

Five demographic variables were also controlled: gender (female = 52%), age (Mdn = 8, 50-54 years old), education (Mdn = 3, some post–high school, no bachelor’s degree), income (Mdn = 13, US$40,000-US$44,999), and race (White/non-Hispanic = 59%). Two political variables were also controlled: partisanship and political interest. These variables were found to relate to the selection of partisan or neutral news sources and the magnitude of attribute agenda-setting effects (Camaj, 2014). Partisanship was measured on a 7-point scale, ranging from 1 (strong Democrat) to 7 (strong Republican). Political interest was a single-item question asking, “How often do you pay attention to what’s going on in government and politics?” based on a 5-point scale (1 = never, 5 = always).

Analytic Procedure

Ordinary least squares multiple regressions were conducted to test the hypotheses. The first set of analyses estimated the predictive value of viewing particular TV programs on attribute agenda setting of those programs. The second set of analyses assessed how attribute agenda setting of each program related to viewers’ polarized attitudes toward the two candidates. We emphasize that the cross-sectional nature of the analyses does not allow us to determine causal directions among news viewing, attribute agenda setting, and polarization. The design of this study does not rule out the possibility of reverse causal directions from polarized attitudes to the selection of news programs and attribute agenda setting. It should be cautioned, therefore, that the study’s findings ought to be interpreted against this methodological limitation.

Results

A total of 1,800 mentions of candidates’ attributes in the news were analyzed. Fox’s Special Report (n = 836) and CNN’s Anderson Cooper (n = 654) had more mentions than NBC Nightly News (n = 310). Mentions about Obama (n = 920) occurred slightly more often than those about Romney (n = 880). The most frequently mentioned traits were leadership and honesty, whereas morality was mentioned the least for both candidates.

As expected, there were clear differences in the balance of positive and negative attributes given to the two candidates among the three programs. In Fox’s Special Report, the balance was clearly tilted in favor of Romney; the overall Janis–Fadner coefficient for Romney attributes was .31 whereas that of Obama attributes was −.68. The uniform negative or positive signs of all five attributes for the two candidates also presented contrasting tones given to the candidates by the program. To a lesser extent, CNN’s Anderson Cooper also showed an imbalance in the reverse direction, offering more positive attributes to Obama (.04) than to Romney (−.21). NBC Nightly News was marginally more positive to Obama (.04) than to Romney (−.04), but the imbalance was the smallest among the three programs (see Table 1).

Table 1.

Attributes of Presidential Candidates by TV News Programs.

	Obama			Romney
	Negative	Positive	Coefficient^a	Negative	Positive	Coefficient
CNN
Morality	1	6	.016	5	5	.000
Leadership	62	59	−.006	60	29	−.055
Caring	9	20	.028	29	8	−.043
Intelligence	21	37	.037	39	37	−.003
Honesty	34	25	−.019	118	50	−.126
Total	127	147	.039	251	129	−.212
Fox
Morality	7	1	−.010	2	7	.012
Leadership	235	25	−.377	29	111	.196
Caring	14	10	−.005	20	34	.027
Intelligence	65	9	−.098	9	47	.096
Honesty	125	13	−.201	35	38	.005
Total	446	58	−.681	95	237	.305
NBC
Morality	0	2	.014	0	11	.065
Leadership	43	36	−.027	32	21	−.040
Caring	5	24	.111	19	7	−.050
Intelligence	4	8	.019	13	14	.003
Honesty	14	6	−.039	26	25	−.003
Total	66	76	.038	90	78	−.038

Janis–Fadner’s Coefficient of Imbalance.

Additional chi-square analyses also showed different candidate portrayals by the three programs. Among six pairs (2 Candidates × 3 Programs), five comparisons yielded significant chi-square values. Fox demonstrated significant differences in affective tones at the .001 level in comparison with NBC (Obama, χ² = 118; Romney, χ² = 29) and with CNN (Obama, χ² = 162; Romney, χ² = 99). NBC and CNN also showed significant differences in terms of tones given to Romney (χ² = 7.7, p < .01), but they exhibited no significant differences in coverage of Obama attributes. Overall, the content analysis supported the argument of partisan imbalance on the two cable channels.

H1 focused on the distinctive agenda-setting effects of candidates’ attributes by different TV news programs. As predicted, viewing a particular program was a positive and significant predictor of individual-level agenda-setting effects. Watching CNN’s Anderson Cooper, Fox’s Special Report, and NBC Nightly News was positively associated with the agenda-setting index of each program. Among the three programs, viewing Special Report (β = .06) and Anderson Cooper (β = .05) was significant at the .001 level, whereas watching Nightly News (β = .03) was significant at the .05 level. Even after controlling for media and political variables, the attribute agenda-setting effects of particular news programs were confirmed. Not surprisingly, respondents’ partisanship was the strongest predictor of the match between each program’s attribute agenda and respondents’ attribute perceptions in the expected direction. The stronger the identification with the Republican Party, the higher the congruence of a respondent’s perceptions of candidate attributes with the candidate attributes emphasized in Fox’s Special Report. The stronger the identification with the Democratic Party, the closer the match between a respondent’s attribute perceptions and the attribute agenda of CNN’s Anderson Cooper (see Table 2). The results imply that candidate attributes emphasized by both Fox’s Special Report and CNN’s Anderson Cooper tend to correspond to partisan respondents’ candidate evaluations.²

Table 2.

Regression Analyses Predicting Attribute Agenda-Setting Effects.

	CNN agenda setting		Fox agenda setting		NBC agenda setting
	β	t value	β	t value	β	t value
Age	−.09***	−7.49	.07***	6.30	.00***	0.16
Gender (1 = female)	−.02^†	−1.67	−.00	−0.38	−.02	−1.57
Education	.05***	4.44	−.02	−1.49	.08***	5.19
Income	−.00	−0.12	.03**	2.76	.02	1.28
Race	−.02	−1.58	.08***	8.17	.04**	2.65
Political interest	−.01	−0.69	.02^†	1.69	.05**	2.63
Partisanship	−.71***	−65.71	.73***	72.97	−.44***	−30.94
TV attention	−.02	−1.57	.00	0.27	−.00	−0.08
Internet attention	.01	1.20	−.02*	−1.98	.01	0.39
Newspaper attention	.01	1.02	−.02*	−2.43	−.01	−0.38
Radio attention	−.03**	−2.86	.02*	1.97	−.02	−1.50
CNN’s Anderson Cooper	.05***	4.47	—	—	—	—
Fox’s Special Report	—	—	.06***	5.63	—	—
NBC Nightly News	—	—	—	—	.03*	2.29
R ²	.54		.62		.19
N	4,513		4,513		4,513

Note. Standardized coefficients.

†

p < .1. *p < .05. **p < .01. ***p < .001.

H2 and H3 focused on the attitudinal consequences of agenda-setting effects of candidates’ affective attributes. Because agenda-setting effects of a specific medium or outlet are led by the use of that medium, only the respondents who watched a specific program were included in each statistical analysis. The original 7-point scale of partisanship was collapsed to a 4-point scale to match the dependent variable measured as an absolute value of gap between feelings toward Obama and Romney. Strong Democrats and strong Republicans were recoded to 4, for instance, and midpoint independents were recoded to 1.

H2 predicted that agenda setting of candidates’ attributes by CNN’s Anderson Cooper and Fox’s Special Report would be positively related to polarized attitudes toward the candidates. As hypothesized, the agenda-setting index of Fox’s Special Report was significantly related to polarized attitudes (β = .30, p < .001). That is, those who are more prone to attribute agenda setting by the Fox program had more polarized attitudes toward the opposing candidates. In contrast, the agenda-setting index of CNN’s Anderson Cooper (β = −.06, p = .15) was not significantly related to polarization.

H3 posited that the agenda setting of NBC Nightly News would be inversely associated with polarized attitudes toward candidates. As predicted, the Nightly News agenda-setting index was a negative predictor of polarization (β = −.06, p < .05). People who are more susceptible to the attribute agenda-setting effects of Nightly News had less polarized attitudes. H3 was supported (see Table 3).

Table 3.

Regression Analyses Predicting Polarized Candidate Evaluation.

	CNN polarization		Fox polarization		NBC polarization
	β	t value	β	t value	β	t value
Age	.04	0.86	−.08	−1.31	.08**	2.83
Gender (1 = female)	−.01	−0.21	.04	0.71	.02	0.80
Education	.02	0.52	.04	0.61	.01	0.52
Income	−.08^†	−1.90	−.22***	−3.68	−.01	−0.20
Race	−.08^†	−1.81	−.13*	−2.44	−.10***	−3.76
Political interest	.16**	3.12	.11^†	1.76	.08**	2.62
Partisan strength	.35***	8.51	.23***	4.26	.33***	12.96
TV attention	.10^†	1.89	.09	1.46	.17***	5.37
Internet attention	.04	0.93	.12*	2.03	.02	0.66
Newspaper attention	.03	0.78	.02	0.42	−.06*	−2.08
Radio attention	.01	0.16	.04	0.65	.03	0.96
H2a. CNN agenda setting	−.06	−1.46	—	—	—	—
H2b. Fox agenda setting		—	.30***	5.14	—	—
H3. NBC agenda setting		—	—	—	−.06*	−2.44
R ²	.24		.23		.22
n	511		299		1,297

Note. Standardized coefficients.

†

p < .1. *p < .05. **p < .01. ***p < .001.

Discussion

To elaborate the agenda-setting process in the partisan TV news context, the present study explored distinct attribute agenda-setting effects and attitudinal outcomes by NBC, CNN, and Fox News. Overall, analyses of candidates’ attributes during the 2012 presidential election indicate that distinctive emphases of affective attributes by the different programs were related to separate patterns of agenda-setting effects and attitudinal consequences.

First, we found that a partisan imbalance existed in the portrayal of candidates’ attributes on the three programs. In particular, Fox’s Special Report provided near one-sided coverage favorable to Romney rather than to Obama in terms of affective attributes. To a much lesser degree, CNN’s Anderson Cooper also showed an imbalance but in the opposite direction, giving more favorable coverage to Obama. NBC’s Nightly News was relatively balanced. Accordingly, disparity in affective attributes in the different programs resulted in separate agenda-setting processes. As the basic agenda-setting hypothesis posits, watching a specific program was found to be associated with the agenda-setting influence of that program. Notably, this relationship still held even after controlling for other media and political variables.

In addition, the current research discovered that agenda setting by the different news programs was related to distinct attitudinal outcomes in regard to polarization. Specifically, the agenda setting of Fox’s Special Report was associated with polarized opinions among viewers. Conversely, the agenda setting of NBC Nightly News predicted depolarized attitudes. This discrete function of attribute agenda setting for polarization seems to reflect contrasting patterns of candidate coverage by the programs and ensuing agenda-setting influence. Special Report was clearly one-sided, showing overwhelming negativity toward Obama and considerable positivity toward Romney. This clear-cut contrast in valence should push viewer attitudes toward bipolar extremes, widening the gap of viewer attitudes about the competing candidates. Meanwhile, the relatively balanced candidate treatment on NBC Nightly News should work against polarization by neutralizing the effects of the opposing tones of candidate attributes. The relative ambivalence toward the candidates in terms of affective attributes by CNN’s Anderson Cooper may explain why CNN’s agenda setting did not relate to polarized attitudes. Although the CNN program leaned more toward Obama than Romney, the imbalance was relatively minor compared with the Fox program.

The current investigation extends an emerging scholarly interest in the role of agenda setting in the partisan news context. Our study is consistent with prior research by Muddiman et al. (2014) that found distinctive affective agenda setting by different cable news channels. While their study dealt with the political issue of the Iraq War as an object, this study discovered attribute agenda setting of political candidates during a presidential campaign. Also, we employed a more rigorous test of affective agenda setting by matching the salience order of different attributes in the news with that of public perceptions, instead of comparing overall positive–negative tones in the news with general public evaluations.

In addition, the present research expands previous investigations regarding attitudinal consequences of agenda setting. Although past examinations documented that agenda-setting effects strengthen and direct attitudes toward objects and attributes, few studies have looked at attitudinal outcomes from the perspective of polarization, which is often regarded as the most significant political consequence caused by partisan selective exposure. Building upon the literature about subsequent influence of agenda setting, we illustrated that polarization or depolarization of voters’ candidate evaluations varies in relation to people’s susceptibility to the balance of news tones given to competing candidates by individual news programs.

This research contributes to knowledge about polarization because it is a rare investigation connecting the content of news messages and associated public opinions. In many survey-based studies, a partisan slant of cable news has often been presumed rather than examined. Accordingly, use of a partisan news source was assumed to relate to different types of political attitudes encompassing partisanship, ideology, issue opinion, and candidate evaluation without delving into the associated news content. By focusing on the specific content dimension of news, the current study demonstrates that one aspect of polarization (i.e., attitudes toward candidates) does indeed relate to variations of news content and its effects (i.e., agenda setting of candidate attributes).

More broadly, this research contributes to explicating agenda-setting theory in the shifting TV news environment characterized by a multitude of news channels and partisan selection of news. Some scholars have argued that proliferation of news sources would diminish agenda-setting effects, as we can no longer assume that mass audience would receive similar news content, or more particularly a similar news agenda (e.g., Bennett & Iyengar, 2008; Blumler & Kavanagh, 1999; Chaffee & Metzger, 2001). This study provides some empirical evidence to examine the claim that agenda-setting power of the news media is being weakened. At the affective level, the three TV news programs clearly diverged in their emphasis of candidates’ attributes. The results seem to support the argument that intermedia similarity of news agenda in certain contexts cannot be taken for granted.³ However, the current investigation also shows that the central proposition of agenda setting, transfer of salience from media to public, persists even in the partisan TV news context.

Our results have an important theoretical implication, given that self-selection of news sources is often thought to diminish media effects, including agenda setting. When an audience’s political predispositions guide news choice, the impact of news use might be only a reflection, or reinforcement of preexisting beliefs and opinions (Bennett & Iyengar, 2008). The findings of the present study, however, call for caution before coming to a hasty conclusion that media influence is weakened in the current news setting. Although viewers’ partisanship tended to correspond to the news agenda of the individual program that reflected their political orientations, each program still had a unique agenda-setting influence at the individual level. In this respect, the outcome of our investigation is in line with prior research outside the agenda-setting tradition, noting the possibility that imbalanced news sources may exert strong influence on public opinion as one-sided information can be more persuasive than mixed or balanced messages (e.g., Dalton, Beck, & Huckfeldt, 1998; Feldman, 2011; Zaller, 1992).

Limitations of the current research should be recognized, the first being in regard to the causality presumed in this study. Although media influence on the public is well established in prior agenda-setting research, one must pay careful attention to the growing trend of partisan selective exposure. When an audience selects news sources based on its own political predispositions, identifying causal influence by the news media becomes more elusive. That is, audience evaluation of candidate traits may direct audience selection of particular TV programs. Similarly, those having polarized opinions about candidates may choose certain programs congenial to their opinions. To address this concern, future studies might consider overtime analysis with multiple time lags to examine whether use of particular TV news programs leads to attribute agenda setting and its ensuing effects.

The use of rank-order correlations is another drawback of the present study. As McCombs and Shaw (1972) first used rank-order correlations in their original agenda-setting study, the measurement has been widely used in subsequent agenda-setting research. However, rank-order measurement can reduce more refined differences in news content and public opinion to rather rough rank orders, undermining the predictive power of the variables. This research also has limitations related to the use of secondary data. Because of the unavailability of proper opinion data about MSNBC viewers, its agenda setting and polarization effects were not included in the analyses. In light of observations noting MSNBC’s partisan slant, comparable with Fox News in the reverse direction (Project for Excellence in Journalism, 2008, 2012), one might speculate that MSNBC could have a function similar to Fox: Watching MSNBC may lead to agenda setting of candidates’ attributes, which relates to polarized candidate evaluations.

It should also be noted that this study looked at the content and effects of particular programs. On one hand, this can be seen as a strength because we were able to match content of specific programs directly with opinions of their viewers, instead of making the broad assumption that emphases of candidates’ attributes would be homogeneous across different programs within the same channel. On the other hand, our method restricts generalizability of the findings. Therefore, the results should be interpreted as the effects of specific programs, not as the effects of TV channels in general.

In addition, our focus on specific programs did not allow us to generate a sufficient sample size to examine how watching multiple programs of opposing political slants relates to agenda setting and polarization. In our data, only about 1% of respondents reported having watched both CNN’s Anderson Cooper and Fox’s Special Report. According to previous studies, however, many people cross-watch cable news channels with opposing partisan inclinations (Prior, 2013; Stroud, 2011). Viewing multiple programs repeating similar attributes can augment agenda-setting effects, whereas cross-exposure to programs emphasizing opposing attributes may neutralize each program’s agenda-setting influences (Muddiman et al., 2014). Future research could expand the number of programs to explore the impact of cross-viewing on agenda setting and polarization.

Our findings reinforce a concern that polarization, along with partisan self-exposure, may undermine common ground for public deliberation. The implications of partisan news use for democracy, however, need to be examined further because those who use partisan sources may participate more actively in politics than others (e.g., Dilliplane, 2011; Stroud, 2011). In this regard, we expect that follow-up research would address the role of attribute agenda setting and polarization for public engagement in politics. Given previous studies documenting the positive association of agenda setting with political behaviors such as voting choice (Roberts, 1992; Wu & Coleman, 2009) and campaign activities (Moon, 2013), attribute agenda setting and polarization by partisan news sources may either mediate or moderate those sources’ influence on political participation.

Finally, scholars need to consider how emerging third-level agenda setting, also termed the network agenda-setting model, can further illuminate the role of agenda setting in the partisan news environment. Focusing on networked relationships among the elements of agendas, this new model posits that news media can transfer the salience of relationships in the news to the public, and this hypothesis has been supported at both issue and attribute levels (Guo & McCombs, 2011a, 2011b; Vargo, Guo, McCombs, & Shaw, 2014; Vu, Guo, & McCombs, 2014). This new approach can extend to research on the agenda-setting effects of partisan news sources. Future studies can explore the potentially distinct process of network agenda setting by different partisan news outlets and its relationship to subsequent attitudes seen as associations of related attitudes, rather than as discrete, separate attitudes.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Ki Deuk Hyun (PhD, University of Texas at Austin) is an assistant professor in the School of Communications at Grand Valley State University, Allendale, Michigan, USA. His research interests include content and effects of news, new media technologies, and international communication.

Soo Jung Moon (PhD, University of Texas at Austin) is an associate professor in the Department of Mass Communications at the University of West Georgia, Carrollton, Georgia, USA. Her research interests include media effects and new communication technologies.

References

Abelson

(1995). Attitude extremity. In Petty

Krosnick

(Eds.), Attitude strength: Antecedents and consequences (pp. 25-41). Mahwah, NJ: Lawrence Erlbaum.

Aday

(2010). Chasing the bad news: An analysis of 2005 Iraq and Afghanistan War coverage on NBC and Fox News Channel. Journal of Communication, 60, 144-164. doi:10.1111/j.1460-2466.2009.01472.x

Alterman

(2004). What liberal media? The truth about bias and the news. New York, NY: Basic Books.

Balmas

Sheafer

(2010). Candidate image in election campaigns: Attribute agenda setting, affective priming, and voting intentions. International Journal of Public Opinion Research, 22, 204-229. doi:10.1093/ijpor/edq009

Baron

D. P.

(2006). Persistent media bias. Journal of Public Economics, 90, 1-36. doi:10.1016/j.jpubeco.2004.10.006

Baum

M. A.

Groeling

(2008). News media and the polarization of American political discourse. Political Communication, 25, 345-365. doi:10.1080/10584600802426965

Bennett

W. L.

Iyengar

(2008). A new era of minimal effects? The changing foundations of political communication. Journal of Communication, 58, 707-731. doi:10.1111/j.1460-2466.2008.00410.x

Blumler

J. G.

Kavanagh

(1999). The third age of political communication: Influences and features. Political Communication, 16, 209-230. doi:10.1080/105846099198596

Bozell

L. B.

(2004). Weapons of mass distortion: The coming meltdown of the liberal media. New York, NY: Three Rivers Press.

10.

Brock

(2004). The republican noise machine: Right wing media and how it corrupts democracy. New York, NY: Crown.

11.

Camaj

(2014). Need for orientation, selective exposure, and attribute agenda-setting effects. Mass Communication and Society, 17, 689-712. doi:10.1080/15205436.2013.835424

12.

Chaffee

Metzger

(2001). The end of mass communication? Mass Communication and Society, 4, 365-379. doi:10.1207/S15327825MCS0404_3

13.

Coleman

H. D.

(2010). Proposing emotion as a dimension of affective agenda setting: Separating affect into two components and comparing their second-level effects. Journalism & Mass Communication Quarterly, 87, 315-327. doi:10.1177/107769901008700206

14.

Dalton

R. J.

Beck

P. A.

Huckfeldt

(1998). Partisan cues and the media: Information flows in the 1992 presidential election. American Political Science Review, 92, 111-126. Retrieved from https://www-jstor-org.web.bisu.edu.cn/stable/2585932

15.

Dilliplane

(2011). All the news you want to hear: The impact of partisan news exposure on political participation. Public Opinion Quarterly, 75, 287-316. doi:10.1093/poq/nfr006

16.

Farnsworth

Lichter

R. S.

(2011). Network television’s coverage of the 2008 presidential election. American Behavioral Scientist, 55, 354-370. doi:10.1177/0002764211398066

17.

Feldman

(2011). The opinion factor: The effects of opinionated news on information processing and attitude change. Political Communication, 28, 163-181. doi:10.1080/10584609.2011.565014

18.

Feldman

Maibach

E. W.

Roser-Renouf

Leiserowitz

(2012). Climate on cable: The nature and impact of global warming coverage on Fox News, CNN, and MSNBC. International Journal of Press/Politics, 17, 3-31. doi:10.1177/1940161211425410

19.

Festinger

(1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.

20.

Gentzkow

Shapiro

J. M.

(2006). Media bias and reputation. Journal of Political Economy, 114, 280-316. doi:10.1086/499414

21.

Golan

Wanta

(2001). Second-level agenda setting in the New Hampshire primary: A comparison of coverage in three newspapers and public perceptions of candidates. Journalism & Mass Communication Quarterly, 78, 247-259. doi:10.1177/107769900107800203

22.

Groseclose

Milyo

(2005). A measure of media bias. Quarterly Journal of Economics, 120, 1191-1237. doi:10.1162/003355305775097542

23.

Guo

McCombs

(2011a, May). Network agenda setting: A third level of media effects. Paper presented at the International Communication Association annual conference, Boston, MA.

24.

Guo

McCombs

(2011b, August). Toward the third level of agenda setting theory: A network agenda setting model. Paper presented at the Association for Education in Journalism and Mass Communication annual conference, St. Louis, MO.

25.

Hillygus

D. S.

Shields

T. G.

(2005). Moral issues and voter decision making in the 2004 presidential election. Political Science & Politics, 38, 201-209. doi:10.1017/S1049096505056301

26.

Hollander

B. A.

(2008). Tuning out or tuning elsewhere? Partisanship, polarization, and media migration from 1998 to 2006. Journalism & Mass Communication Quarterly, 85, 23-40. doi:10.1177/107769900808500103

27.

Iyengar

Hahn

K. S.

(2009). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59, 19-39. doi:10.1111/j.1460-2466.2008.01402.x

28.

Janis

I. L.

Fadner

R. H.

(1943). A coefficient of imbalance for content analysis. Psychometrika, 8, 105-119. doi:10.1007/BF02288695

29.

Jones

D. A.

(2002). The polarizing effect of new media messages. International Journal of Public Opinion Research, 14, 158-174. doi:10.1093/ijpor/14.2.158

30.

Kim

McCombs

M. E.

(2007). News story descriptions and the public’s opinions of political candidates. Journalism & Mass Communication Quarterly, 84, 299-314. doi:10.1177/107769900708400207

31.

King

(1997). The press, candidate images, and voter perceptions. In McCombs

Shaw

D. L.

Weaver

(Eds.), Communication and democracy: Exploring the intellectual frontiers in agenda-setting theory (pp. 29-40). Mahwah, NJ: Lawrence Erlbaum.

32.

Kiousis

(2005). Compelling arguments and attitude strength: Exploring the impact of second-level agenda setting on public opinion of presidential candidate images. Harvard International Journal of Press/Politics, 10(2), 3-27. doi:10.1177/1081180X05276095

33.

Kiousis

(2011). Agenda-setting and attitudes: Exploring the impact of media salience on perceived salience and public attitude strength of US presidential candidate from 1984-2004. Journalism Studies, 12, 359-374. doi:10.1080/1461670X.2010.501149

34.

Kiousis

McCombs

M. E.

(2004). Agenda-setting effects and attitude strength: Political figures during the 1996 presidential election. Communication Research, 31, 36-57. doi:10.1177/0093650203260205

35.

Kitty

(2005). Outfoxed: Rupert Murdoch’s war on journalism. New York, NY: Disinformation Limited.

36.

Klapper

J. T.

(1960). The effects of mass communication. Glencoe, IL: The Free Press.

37.

Lovett

B. J.

Jordan

A. H.

(2005). Moral values, moralism, and the 2004 presidential election. Analyses of Social Issues and Public Policy, 5, 165-175. doi:10.1111/j.1530-2415.2005.00062.x

38.

McCombs

M. E.

(2014). Setting the agenda: The mass media and public opinion (2nd ed.). Cambridge, UK: Polity Press.

39.

McCombs

M. E.

Llamas

J. P.

Lopez-Escobar

Rey

(1997). Candidate images in Spanish elections: Second-level agenda-setting effects. Journalism & Mass Communication Quarterly, 74, 703-717. doi:10.1177/107769909707400404

40.

McCombs

M. E.

Lopez-Escobar

Llamas

J. P.

(2000). Setting the agenda of attributes in the 1996 Spanish general election. Journal of Communication, 50, 77-92. doi:10.1111/j.1460-2466.2000.tb02842

41.

McCombs

M. E.

Reynolds

(2002). News influence on our pictures of the world. In Bryant

Zillmann

(Eds.), Media effects (pp. 1-16). Mahwah, NJ: Lawrence Erlbaum.

42.

McCombs

M. E.

Shaw

D. L.

(1972). The agenda-setting function of the mass media. Public Opinion Quarterly, 36, 176-187. doi:10.1086/267990

43.

Moon

S. J.

(2013). Attention, attitude, and behavior: Second-level agenda-setting effects as a mediator of media use and political participation. Communication Research, 40, 698-719. doi:10.1177/0093650211423021

44.

Morris

J. S.

Francia

P. L.

(2010). Cable news, public opinion, and the 2004 party conventions. Political Research Quarterly, 63, 834-849. doi:10.1177/1065912909338463

45.

Muddiman

Stroud

N. J.

McCombs

(2014). Media fragmentation, attribute agenda setting, and political opinions about Iraq. Journal of Broadcasting & Electronic Media, 58, 215-233. doi:10.1080/08838151.2014.906433

46.

Mutz

D. C.

(2002). Cross-cutting social networks: Testing democratic theory in practice. American Political Science Review, 96, 111-126. doi:10.1017/S0003055402004264

47.

Mutz

D. C.

(2006). How the mass media divide us. In Nivola

P. S.

Brady

D. W.

(Eds.), Red and blue nation? Characteristics and causes of America’s polarized politics (Vol. 1, pp. 223-242). Washington, DC: Brookings Institution Press.

48.

Pew Research Center. (2009). Partisanship and cable news audiences. Retrieved from http://www.pewresearch.org/2009/10/30/partisanship-and-cable-news-audiences/

49.

Pew Research Center. (2012). News sources, election night and views of press coverage. Retrieved from http://www.people-press.org/2012/11/15/section-4-news-sources-election-night-and-views-of-press-coverage/

50.

Prior

(2013). Media and political polarization. Annual Review of Political Science, 16, 101-127. doi:10.1146/annurev-polisci-100711-135242

51.

Project for Excellence in Journalism. (2004). The debate effect: How the press covered the pivotal period of the 2004 presidential campaign. Retrieved from http://www.journalism.org/files/legacy/debateffect.pdf

52.

Project for Excellence in Journalism. (2005). The state of the news media 2005. Retrieved from http://stateofthemedia.org/2005/cable-tv-intro/content-analysis/

53.

Project for Excellence in Journalism. (2008). The color of news: How different media have covered the general election. Retrieved from http://www.journalism.org/files/legacy/The%20COLOR%20OF%20NEWS%20FINAL.pdf

54.

Project for Excellence in Journalism. (2012). The final days of the media campaign 2012. Retrieved from http://www.journalism.org/2012/11/19/final-days-media-campaign-2012/

55.

Project for Excellence in Journalism. (2013). The changing TV news landscape. Retrieved from http://stateofthemedia.org/2013/special-reports-landing-page/the-changing-tv-news-landscape/

56.

Ragas

M. W.

(2014a). Agenda-setting in the corporate sphere: Synthesizing findings and identifying new opportunities in this growing domain. In Johnson

(Ed.), Agenda setting in a 2.0 world: A tribute to Maxwell McCombs (pp. 256-280). New York, NY: Routledge.

57.

Ragas

M. W.

(2014b). Intermedia agenda setting in business news coverage. In Hart

(Ed.), Communication and language analysis in the public sphere (pp. 335-357). Hershey, PA: IGI Global.

58.

Roberts

M. S.

(1992). Predicting voting behavior via the agenda-setting tradition. Journalism Quarterly, 69, 878-892. doi:10.1177/107769909206900408

59.

Sheafer

(2007). How to evaluate it: The role of story-evaluative tone in agenda setting and priming. Journal of Communication, 57, 21-39. doi:10.1111/j.0021-9916.2007.00327.x

60.

Shoemaker

P. J.

Reese

S. D.

(2014). Mediating the message in the 21 century: A media sociology perspective (3rd ed.). New York, NY: Routledge.

61.

Smith

Searles

(2014). Who let the (attack) dogs out? New evidence for partisan media effects. Public Opinion Quarterly, 78, 71-99. doi:10.1093/poq/nft082

62.

Stanley

(2012, August 31). How MSNBC became Fox’s liberal evil twin. The New York Times. Retrieved from http://www.nytimes.com/2012/08/31/us/politics/msnbc-as-foxs-liberal-evil-twin.html

63.

Stroud

N. J.

(2010). Polarization and partisan selective exposure. Journal of Communication, 60, 556-576. doi:10.1111/j.1460-2466.2010.01497.x

64.

Stroud

N. J.

(2011). Niche news: The politics of news choice. New York, NY: Oxford University Press.

65.

Sunstein

C. R.

(2001). Republic.com . Princeton, NJ: Princeton University Press.

66.

Taber

C. S.

Lodge

(2006). Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50, 755-769. doi:10.1111/j.1540-5907.2006.00214.x

67.

Vargo

Guo

McCombs

M. E.

Shaw

(2014). Network issue agendas on Twitter during the 2012 U.S. presidential election. Journal of Communication, 64, 296-316. doi:10.1111/jcom.12089

68.

H. T.

Guo

McCombs

M. E.

(2014). Exploring “the world outside and the pictures in our heads”: A network agenda-setting study. Journalism & Mass Communication Quarterly, 91, 669-686. doi:10.1177/1077699014550090

69.

Weaver

D. H.

(1984). Media agenda-setting and public opinion: Is there a link? In Bostrom

(Ed.), Communication yearbook 8 (pp. 680-691). Newberry Park, CA: SAGE.

70.

Weaver

D. H.

(1991). Issue salience and public opinion: Are there consequences of agenda-setting? International Journal of Public Opinion Research, 3, 53-68. doi:10.1093/ijpor/3.1.53

71.

D. H.

Coleman

(2009). Advancing agenda-setting theory: The comparative strength and new contingent conditions of the two levels of agenda-setting effects. Journalism & Mass Communication Quarterly, 86, 775-789. doi:10.1177/107769900908600404

72.

Zaller

J. R.

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