Extending the Impacts of Hostile Media Perceptions

Abstract

Researchers recently have begun to examine how hostile media perceptions may promote discursive activities aimed at correcting the media’s perceived negative influence. We use nationally representative survey data to test a moderated-mediation model that finds that hostile media perceptions significantly affect support for climate mitigation policies through the mediator of discussion and that the link between discussion and policy support is moderated in a three-way interaction with social network heterogeneity and political ideology. Discussion in homogeneous social networks increases opinion polarization between liberals and conservatives by intensifying conservatives’ opinions, whereas discussion in heterogeneous social networks decreases polarization by moderating liberals’ opinions.

Keywords

climate change hostile media perceptions opinion polarization environmental communication science communication

Over the past 15 years public opinion on global climate change in the United States has become increasingly polarized (McCright & Dunlap, 2011), with Democrats and liberals generally more supportive of climate mitigation policy and Republicans and conservatives generally more opposed. At the same time, hostile perceptions of media coverage of climate change have become more likely from partisans on both sides of the issue (Kim, 2011). Although existing research and theory suggest that the intensely partisan nature of the climate change debate may give rise to both hostile media perceptions and public opinion polarization (Hansen & Kim, 2011; Kim, 2011), we know surprisingly little about the relationship between these two variables. The present study addresses this research gap by examining how hostile media perceptions influence ideological polarization around the issue of climate change, both directly and indirectly, via effects on discussion about climate change.

Since the hostile media effect was first documented by Vallone, Ross, and Lepper (1985), studies have primarily focused on demonstrating the persistence of this effect across contexts, as well as on the theoretical mechanisms that cause hostile media perceptions to arise (e.g., Coe et al., 2008; Dalton, Beck, & Huckfeldt, 1998; Perloff, 1989). Researchers only very recently have begun to address the impacts of hostile media perceptions. Initial studies in this area have found that hostile media perceptions can affect multiple outcomes, including perception of public opinion (Choi, Yang, & Chang, 2009; Gunther, Christen, Liebhart, & Chia, 2001), feelings of minority alienation (Tsfati, 2007), and trust in the media (Tsfati & Cohen, 2005). In addition, related to the present study, recent research has found that when individuals perceive a hostile media bias they are more likely to engage in expressive actions, including interpersonal discussion, in an effort to correct the bias perceived in the media (Hwang, Pan, & Sun, 2008; Rojas, 2010). We build on this line of research to investigate whether hostile media perceptions influence individuals’ discussion about the issue of climate change. However, we also go a step further to examine the subsequent effect of that discussion on ideological polarization around support for climate mitigation policies, and how this effect is moderated by the issue-relevant opinion heterogeneity of one’s social network. In this way, the current research helps expand our understanding of how hostile media perceptions may drive political behavior and polarization.

Using climate change as a substantive domain of inquiry, we explicate and test the theoretical relationship between hostile media perceptions and public opinion polarization. Specifically, we utilize nationally representative survey data to examine: (1) a mediated pathway in which (a) hostile media perceptions influence levels of discussion about the issue and (b) levels of discussion, in turn, influence policy support, moderated by both political ideology and social network heterogeneity, and (2) a conditional effect of hostile media perceptions on policy support, moderated by political ideology. This is the first study we are aware of to investigate this theoretical model. Overall, the study offers a nuanced examination of how hostile media perceptions may amplify or attenuate political polarization on the issue of global climate change.

Hostile Media Perceptions

Hostile media perceptions arise when partisan individuals perceive the media coverage of an issue to be more biased against their personally held views than the media coverage actually is (Vallone et al., 1985). The effect has been observed in response to objective, balanced coverage of an issue (Perloff, 1989; Vallone et al., 1985), such that partisans on either side of the issue perceive the coverage to be biased against their views. It also has been observed in response to biased coverage of an issue (Coe et al., 2008; Feldman, 2011; Gunther & Christen, 2002), such that the level of bias is viewed to be in relative opposition to personal views (i.e., the perception of an article biased toward an individual’s position would be seen as relatively less biased, and the perception of an article biased away from an individual’s position would be seen as relatively more biased). Hostile media perceptions exemplify the fact that individuals often selectively interpret media information based on previously held beliefs.

The first empirical investigation into hostile media perceptions was conducted by Vallone et al. (1985). Undergraduates were shown U.S. coverage of events that led to a 1982 massacre of Palestinians by a Lebanese militia group; the coverage raised questions about Israeli responsibility for the attack. Pro-Israel students saw an anti-Israel bias in the coverage, whereas pro-Arab students saw a pro-Israel bias.

Since the initial study, research has primarily investigated four explanatory factors for hostile media perceptions: partisan involvement, message processing mechanisms, source characteristics, and perceived reach. Partisan involvement is often a critical factor in driving hostile media perceptions. For example, Dalton et al. (1998) examined reactions to newspaper coverage of the 1992 presidential election and found hostile media perceptions for strong Democrats and Republicans. Hansen and Kim (2011) found similar results, such that hostile media perceptions were stronger for highly involved partisans but exist even for individuals with very low partisan involvement.

Research examining message processing mechanisms have focused on whether (1) selective recall, (2) selective categorization, or (3) holding different standards of fair coverage contributes to hostile media perceptions. While early research found support for selective recall (Vallone et al., 1985), subsequent research has not found selective recall to be a reliable driving mechanism and has instead found mixed support for selective categorization and different standards. Comparing the three mechanisms, Giner-Sorolla and Chaiken (1994) only found evidence for different standards, although the measurement adopted in their study has been criticized (Schmitt, Gunther, & Liebhart, 2004). In contrast, Schmitt et al. (2004) and Gunther and Liebhart (2006) found that of the three proposed mechanisms, selective categorization had the most support.

In addition to partisanship and message mechanisms, the source characteristics of media coverage can drive hostile media perceptions. Different sources for a media story may drive heuristic judgments for expected bias in the story, which can in turn affect perceptions of the media content. Multiple studies (Ariyanto, Hornsey, & Gallois, 2007; Arpan & Raney, 2003; Baum & Gussin, 2008) have found that the perceived bias of a news story is influenced by the perceived bias of the source of the story.

Finally, recent research has suggested that the perceived reach of a message may moderate hostile media perceptions such that messages perceived as having a wide reach are more likely to trigger perceptions of bias. For example, Gunther and Liebhart (2006) found that when information was presented as a national news article it strengthened hostile media perceptions compared to identical information presented as a college newspaper article. Gunther, Miller, and Liebhart (2009) found that this moderation may operate on a continuum, such that a national article strengthens hostile media perceptions more than a regional news article, which, in turn, has a greater effect than a student paper.

As mentioned previously, there is a large amount of research focused on the causal factors that drive perceptions of bias in the media, but there is still only a limited amount of research examining the impacts of hostile media perceptions, especially in the realm of how hostile media perceptions may drive political behavior and polarization. Therefore, in the present study we examine the potential outcomes of hostile media perceptions.

We use climate change as the substantive domain to investigate the links between hostile media perceptions, discussion, and policy support. Climate change has become an increasingly polarizing issue in the United States (McCright & Dunlap, 2011), with Democrats generally more concerned about the issue and more supportive of policy action to mitigate negative impacts than Republicans. This polarization has been driven, in part, by a conservative mobilization to question the veracity of climate change claims starting in the early 1990s (McCright & Dunlap, 2000), as well as political conflict over attempts to create international agreements to address the issue (McCright & Dunlap, 2003). This polarization is exhibited in how conservatives and liberals frame the issue (Jang & Hart, 2015) as well as how they may respond to messaging about the issue (M. C. Nisbet, 2009). Thus, the issue of climate change serves as a useful domain of inquiry as it is a timely issue with significant societal consequences that is also highly polarizing with a strong chance of generating hostile media perceptions for individuals who support and oppose policy action to address climate change. Moreover, discussion in the public sphere about climate change has the potential to attenuate some of the polarization through consideration of alternative viewpoints but may also amplify polarization as motivated reasoning hardens previously held beliefs (Hart & Nisbet, 2012).

Corrective Action and Policy Support

Although still a new area of research, recent studies suggest that individuals are motivated to take more issue-relevant action in the public sphere as their hostile media perceptions increase (Hwang et al., 2008; Rojas, 2010). This effect arises out of synergies between hostile media perceptions and the third-person effect (Davison, 1983). The third-person effect predicts that someone will often believe that negative information in the media will have a greater influence on other individuals than on the self (Tsfati, Cohen, & Gunther, 2011). Thus, when an individual believes that coverage of an issue in the media is more negative toward his or her own position, the individual may also, in turn, believe that the press will influence others’ opinions to move away from his or her own. In other words, greater hostile media perceptions can cause individuals to believe that, due to the perceived influence of the hostile information, public opinion is more opposed to the position that the individual holds (Choi et al., 2009; Gunther, 1998; Gunther et al., 2001).

How do individuals respond when they believe that the media is biased against them and that this media bias is influencing public opinion as well? Two recent studies suggest that individuals may attempt to take corrective action in the public sphere to counteract the presumed negative influence arising from hostile media perceptions (Hwang et al., 2008; Rojas, 2010). Hwang et al. (2008) found that hostile media perceptions increased the feeling of media indignation—that the media was being unfair in the way an issue was covered—which caused individuals to engage in actions, such as discussing the issue with others, to try and correct the public perception of an issue. Rojas (2010) found similar results: When individuals believed that the media was biased against them, they were more likely to enter the public sphere and engage with others in order to counteract the perceived negative media influence. In the present study, we focus on discussion with friends and family as previous research has found that people most commonly and frequently talk about political issues such as climate change with friends and family (Klofstad, McClurg, & Rolfe, 2009; Mutz, 2006). This provides an opportunity to investigate how hostile media perceptions may encourage transactive exchanges about an issue, which may, in turn, affect policy support. Based on the studies conducted by Hwang et al. (2008) and Rojas (2010), we formally propose the following:

Hypothesis 1: Hostile media perceptions will be correlated with higher levels of discussion about climate change with friends and family.

To the best of our knowledge, no study has examined how willingness to engage in discussion with others, as affected by hostile media perceptions, may subsequently influence another variable. Here, we examine how the exchange of ideas from discussion driven by hostile media perceptions may affect individuals’ support for policies to address climate change, and how the influence of discussion on policy support may vary as a function of individuals’ social network heterogeneity and political ideology.

Interpersonal discussion is considered a critical mechanism for the strengthening and polarization of attitudes (Moscovici & Zavalloni, 1969). This is because discussion involves accessing and expressing one’s personal views (Tesser & Conlee, 1975), initiates social comparison processes that can serve to corroborate attitudes (Baron et al., 1996), and exposes participants to novel, persuasive arguments from others (Burnstein & Vinokur, 1977)—all of which can contribute to opinion formation, strengthening, and sometimes polarization. However, scholars recognize that the effects of discussion on individual attitudes and opinions likely depend on the nature of the social context, namely, the relative heterogeneity of one’s discussion network (e.g., Huckfeldt, Mendez, & Osborn, 2004; Mutz, 2002a; Visser & Mirabile, 2004).

To examine the heterogeneity of an individual’s social network, we adopt the approach taken by Mutz (2002b) and Scheufele, Nisbet, Brossard, and Nisbet (2004), in which heterogeneous “crosscutting” networks are identified by how much individuals perceive that members of their network differ from them in orientation toward the issue under consideration. In doing so, we focus specifically on opinion heterogeneity, which is often not captured by traditional measures of network heterogeneity such as ethnicity, socioeconomic status, or gender (Feldman & Price, 2008). Thus, mentions of social network heterogeneity throughout the article are made with respect to like-minded or diverse views with respect to climate change.

Sunstein (2001), in particular, has argued that engagement with social networks with low self-other heterogeneity—by limiting the pool of arguments and creating normative pressure to conform to the group—can reinforce beliefs, amplify political polarization, and lower political tolerance. In support of this idea, Binder, Dalrymple, Brossard, and Scheufele (2009) found that discussion in like-minded social networks led to the development of more extreme attitudes about stem cell research. Moreover, because their study used longitudinal data, they were able to show that discussion leads to attitude extremity, rather than vice versa. Using experimental and correlational designs, Visser and Mirabile (2004) also demonstrated that people in more homogeneous networks were more resistant to attitude change than were people in heterogeneous networks.

In contrast, and consistent with the tenets of deliberative theory, if individuals engage in discussion within a politically heterogeneous network, they are likely to come into contact with opposing views that may encourage them to reconsider their initial position and ultimately arrive at a more sound, reasoned conclusion (Arendt, 1968; Delli Carpini, Cook, & Jacobs, 2004; Habermas, 1989). Indeed, studies have shown that diversity within one’s discussion network increases awareness of rationales for others’ viewpoints, which, in turn, fosters tolerance for the other side (Mutz, 2002b; Price, Cappella, & Nir, 2002). Moreover, as people are exposed to diverse perspectives as a result of discussion within heterogeneous networks, their political attitudes become more ambivalent, unstable, and less polarized (Huckfeldt et al., 2004; Levitan & Visser, 2009; Mutz, 2002a). Parsons (2010) also found that exposure to divergent viewpoints in political conversations depolarizes partisans’ emotions toward political candidates. Specifically, political disagreement increased the intensity of negative emotions toward one’s preferred candidate, while increasing the intensity of positive emotions toward the opposing party’s candidate. Additional evidence suggests that attitude depolarization in diverse social networks is due not only to the information value of these networks (i.e., the availability of opposing viewpoints) but also to differences in how new information is processed. For example, Levitan and Visser (2008) demonstrated that people in heterogeneous networks are more likely to carefully scrutinize attitude-relevant information and will moderate their attitudes in response to strong counterarguments.

Drawing on this prior theory and research, we would expect that individuals who discuss climate change with friends and family who largely share their views on the issue may develop more polarized views toward climate policy. On the other hand, if discussing climate change with friends and family who hold different views on climate change, the exposure to a diversity of ideas may have the opposite effect and reduce polarization. Because public opinion on climate change is divided along ideological lines (Hart & Nisbet, 2012; McCright & Dunlap, 2011; M. C. Nisbet & Kotcher, 2009), the direction of amplified or diminished polarization will likely be contingent on political ideology (i.e., in the case of polarization, liberals will increase their policy support and conservatives will decrease their support; in the case of depolarization, liberals will decrease their policy support and conservatives will increase their support). We thus propose a three-way interaction between discussion, social network heterogeneity, and political ideology:

Hypothesis 2a: For individuals with low social network heterogeneity, discussion will be associated with greater political polarization, with liberals becoming more supportive and conservatives less supportive of climate mitigation policies as discussion increases.

Hypothesis 2b: For individuals with high social network heterogeneity, discussion will be associated with lower political polarization, with liberals becoming less supportive and conservatives more supportive of climate mitigation policies as discussion increases.

To summarize, we expect that individuals who perceive a hostile bias in media coverage of climate change will be motivated to engage in discussion about climate change in an effort to correct what they see as a potentially negative media influence on public opinion. In turn, discussion—by allowing for the expression and exchange of personal views—is likely to influence individuals’ support for climate mitigation policies. However, the nature of discussion’s influence on policy support is expected to depend on the heterogeneity of one’s discussion network in conjunction with ideology. In like-minded (i.e., low heterogeneity) networks, where opinion reinforcement is likely, both liberals and conservatives are likely to become more extreme, or polarized, in their viewpoints as a result of discussion; specifically, liberals are likely to become more supportive of climate policies and conservatives less supportive. In heterogeneous networks, which expose individuals to diverse viewpoints and thus invite reflection and reconsideration, discussion is likely to moderate the opinions of liberals and conservatives, such that liberals will become less supportive of climate policies and conservatives more supportive. In this case, polarization is attenuated. Following from these proposed relationships, we expect that hostile media perceptions will influence policy support indirectly, via discussion; however, because the link between discussion and policy support is moderated by network heterogeneity and ideology (as specified in Hypothesis 2), the indirect effect will likewise be conditional on network heterogeneity and ideology. Specifically, we hypothesize the following:

Hypothesis 3a: For individuals with low social network heterogeneity, hostile media perceptions—through their effect on discussion—will indirectly correlate with increased policy support among liberals and indirectly correlate with decreased policy support among conservatives, thereby amplifying political polarization.

Hypothesis 3b: For individuals with high social network heterogeneity, hostile media perceptions—through their effect on discussion—will indirectly correlate with decreased policy support among liberals and indirectly correlate with increased policy support among conservatives, thereby attenuating political polarization.

Finally, some previous research has found that hostile media perceptions—independent of their effects on discussion—can encourage individuals to engage in behaviors that reinforce values and beliefs. For example, Kim (2011) found that when individuals perceived a news story about global warming to have a hostile media bias, they tended to engage in selective exposure—that is, they preferred to view a news story that confirmed their position on the issue as opposed to a story that contradicted their position. Additional studies (Hwang et al., 2008; Morris, 2007) found that hostile media perceptions can cause individuals to avoid diverse news perspectives and instead seek out news sources that are more likely to align with their existing attitudes and beliefs. This selective exposure to like-minded information, in turn, has been found to drive attitude extremity and polarization (Feldman, Myers, Hmielowski, & Leiserowitz, 2014; Stroud, 2011). In this way, hostile media perceptions have the potential to reinforce and strengthen previously held values and beliefs, thereby exacerbating polarization between liberals and conservatives on the issue of climate change.

The link between hostile media perceptions and polarization may be particularly likely given the recent proliferation of partisan news sources on cable television and the Internet. For example, prior research has shown stark divides in how Fox News covers climate change relative to CNN and MSNBC, with coverage on Fox News more dismissive of the science of climate change and coverage on CNN and MSNBC more accepting (Feldman, Maibach, Roser-Renouf, & Leiserowitz, 2012). According to research on the “relative” hostile media perception, when news stories are slanted rather than balanced, audiences are more likely to identify bias in coverage that does not cohere with their political perspective—for example, conservatives will perceive less bias in a conservatively leaning news story and more bias in a liberally leaning story, with the reverse true for liberals (Arceneaux, Johnson, & Murphy, 2012; Coe et al., 2008; Feldman, 2011; Gunther & Christen, 2002). In other words, people become mistrustful of news content that does not come from their ideologically preferred sources. This, in turn, may make it easier for partisans to validate their personal political beliefs—by embracing information that reinforces their views and rejecting counterattitudinal content (Taber & Lodge, 2006).

In addition, Ladd (2011)—although not looking specifically at hostile media perceptions—has found that people who distrust the media in general are more resistant to new information about public affairs and instead rely on their partisan predispositions when forming and updating their beliefs about political candidates and issues. As a result, the beliefs of Republicans and Democrats are more divergent when they distrust the media than when they trust the media. Generalizing to the present context, this finding suggests that perceptions of a hostile bias in media coverage of climate change should reinforce liberals’ predispositions to support climate mitigation policies and conservatives’ predispositions to oppose such policies. Similarly, just as social identity provides a filter through which people perceive media (Perloff, 1989), once media coverage is deemed hostile, this likely reinforces one’s initial social identity by strengthening feelings of affinity with members of one’s own group and feelings of hostility toward the other group (Slater, 2007). This, too, may make it especially likely that individuals who hold hostile media perceptions will rely on their political identity as a cue when forming their opinions on issues such as climate change. Following from these ideas, we propose a final hypothesis:

Hypothesis 4: Higher levels of hostile media perceptions will be associated with higher levels of polarization between liberals and conservatives. Specifically, higher levels of hostile media perceptions will be associated with greater support for climate policies among liberals and less support among conservatives.

Taken together, these hypotheses offer a nuanced view of the relationship between hostile media perceptions and political polarization on the issue of climate change, proposing that hostile media perceptions may directly amplify political polarization, indirectly amplify political polarization through discussion for individuals with low social network heterogeneity, but also indirectly attenuate political polarization when encouraging individuals with high levels of social network heterogeneity to engage in discussion on the issue. Figure 1 shows how these hypotheses fit together in a conceptual map for testing.

Figure 1.

Conceptual map for testing.

Method

The analysis for this study uses nationally representative survey data of 1,000 Americans collected in late October and early November 2011. Participants for the study were randomly sampled from a 50,000 member online panel maintained by Knowledge Networks. Knowledge Networks populates its panel through a recruitment process utilizing random digit dialing and address-based sampling. This means that recruitment includes participants who have both listed and unlisted phone numbers, telephone, nontelephone, and cell phone–only households. Panelists typically complete about two 5- to 20-minute surveys each month and receive an incentive of $4 to $6 in return. Knowledge Networks provides a free netbook and Internet service to make sure individuals without computers/Internet access are included in the panel. The survey had a completion rate of 65.1% and a cumulative response rate of 6.2%.¹ The sample was 47% male and 74% White, with a mean age of 47.37 (SD =17.00) and an average of 13.89 years of education (SD = 2.64). Median annual household income was $50,000 to $59,999.

Measures

Independent Variable

Our measure of hostile media perceptions consisted of two items, adapted from Hwang et al. (2008). Respondents were asked to indicate their level of agreement, where 1 = strongly disagree and 4 = strongly agree, with the following two statements: “Media coverage of global warming is biased against my views on the issue” and “Media coverage of global warming is distorted by powerful interest groups.” Responses were averaged together (r = .50, p < .001; M = 2.70, SD = 0.73).

Mediating Variable

Discussion about climate change was measured by asking respondents, “How often do you discuss global warming with family and friends?” Respondents answered on a scale that ranged from 1 (never) to 4 (often; M = 2.07, SD = 0.83).

Moderating Variables

Political ideology was measured by asking respondents if they see themselves as liberal or conservative, where 1 = very liberal and 5 = very conservative (M = 3.16, SD = 1.04).

Global warming social network heterogeneity was measured by asking respondents, “How many of your friends share your views on global warming?” Respondents answered on a scale that ranged from 1 (all) to 5 = (none; M = 3.30, SD = 1.10), meaning that higher scores signified higher levels of social network heterogeneity.

Dependent Variables

Policy support for addressing climate change was measured by asking respondents how much they supported a variety of climate mitigation measures on a scale that ranged from 1 (strongly opposed) to 4 (strongly support). Respondents offered support responses to the following five proposals: (1) “Regulate carbon dioxide (the primary greenhouse gas) as a pollutant” (M = 2.90, SD = 0.87); (2) “Require electric utilities to produce at least 20% of their electricity from wind, solar, or other renewable energy sources, even if it cost the average household an extra $100 a year” (M = 2.73, SD = 0.95); (3) “Sign an international treaty that requires the United States to cut its emissions of carbon dioxide by 90 by the year 2050” (M = 2.75, SD = 0.91); (4) “Fund more research into renewable energy sources, such as solar and wind power” (M = 3.10, SD = 0.87); and (5) “Provide tax rebates for people who purchase energy-efficient vehicles or solar panels” (M = 3.02, SD = 0.86). Responses for the five items, which a factor analysis confirmed loaded on a single factor, were then averaged into a single policy support scale (Cronbach’s α = .88; M = 2.90, SD = 0.73).

Control Variables

In addition to controlling for the sociodemographic variables discussed earlier (age, gender, race, education, and income), we controlled for attention to global warming issues in the news, perceived understanding of global warming, and use of various media sources for news. For global warming news attention, respondents indicated how much attention they pay to news stories about global warming (1 = none, 4 = a lot; M = 2.41, SD = 0.91). For perceived understanding of global warming, respondents indicated how well informed they feel personally about global warming (1 = not at all informed, 4 = very well informed; M = 2.58, SD = 0.74). For preferred news media source, respondents indicated which of the following media sources they turned to most often to keep up with current news and world events: Internet, television, radio, print newspapers, and magazines. Three dummy variables were computed to represent preferences for broadcast (television and radio; 65.4%), print (newspapers and magazines; 12.9%), and the Internet (21.7%).

Missing Data

Some participants did not respond to all of the variables included in the analysis. In order to reduce the amount of missing data, we adopted a hot deck imputation procedure as recommended by Myers (2011). For the imputation procedure, survey respondents were randomly permuted by sex, income, and education. Missing responses for variables were imputed with the response of the nearest respondent in the randomly permuted file with identical responses for sex, age, income level, and education as the respondent with the missing variable. This means that missing responses were imputed through a random sampling without replacement procedure that drew from individuals with the same sex, age, income level, and education. The vast majority of respondents (90.8%) did not require any imputation, and all variables had an imputation rate of less than 5.6%.

Analysis

Our analysis was conducted using the SPSS PROCESS macro developed by Hayes (2013). The PROCESS macro uses an ordinary least squares (OLS) regression-based path analytic framework to test hypothesized direct, indirect, and conditional effects. In addition to providing OLS regression coefficients for the analysis, the PROCESS macro provides bootstrapped confidence intervals for the indirect effect of hostile media perceptions on policy support via discussion at varying levels of political ideology and social network heterogeneity, while accounting for all control variables. The use of bootstrapped confidence intervals to estimate mediated effects is considered superior to alternatives such as the Sobel test or causal steps approach (Hayes, 2013). Bootstrap intervals were calculated using 10,000 sampling iterations and bias-corrected estimates. The 95% confidence intervals may be interpreted such that if the lower and upper confidence intervals (LCIs and UCIs, respectively) are either both below or both above zero there is a significant indirect effect; if the LCIs and UCIs straddle zero then there is no significant effect.

Results

Two OLS regressions were first conducted to separately predict (1) the mediator (discussion) from the independent variable (hostile media perceptions) and (2) the dependent variable (policy support) from both (a) hostile media perceptions, accounting for an interaction between hostile media perceptions and political ideology, and (b) from discussion, accounting for an interaction between discussion, political ideology, and social network heterogeneity. The results of these analyses are provided in Table 1.

Table 1.

Ordinary Least Squares Regression Results Predicting Discussion (Mediator) and Policy Support (Dependent Variable).

Variable	Discussion, Β (SE)	Policy support, Β (SE)
Control variables
Age	.00 (.00)	.002 (.001)*
Gender (male)	−.08 (.04)	−.05 (.04)
Race	−.09 (.03)	−.01 (.05)
Education	.03 (.01)**	.02 (.01)*
Income	−.001 (.005)	−.004 (.005)
Perceived understanding	.25 (.03)***	−.03 (.03)
Attention to climate change news	.41 (.03)***	.27 (.03)***
Preferred news media source^a
Broadcast	−.12 (.05)*	.08 (.05)
Print	.05 (.08)	.09 (.07)
Conceptual variables
Hostile media perceptions	.09 (.03)**	−.02 (.03)
Discussion		−.06 (.03)
Political ideology		−.21 (.02)***
Network heterogeneity		−.01 (.02)
Political ideology × network heterogeneity		.05 (.02)**
Discussion × political ideology		−.02 (.02)
Discussion × network heterogeneity		−.02 (.02)
Discussion × political ideology × network heterogeneity		.05 (.04)**
Hostile media perceptions × political ideology		−.07 (.02)**
Constant	−1.80 (0.16)***	1.89 (0.16)***
R ²	.34	.32
N	989	989

Note. Unstandardized regression coefficients are reported. Variables were mean-centered prior to computing the interaction terms.

Internet is the reference category.

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

The analysis reveals that hostile media perceptions were positively related to discussion about global warming (unstandardized Β = .09, p < .01), in support of Hypothesis 1.

Discussion, in turn, had a significant relationship with support for policies to address climate change, and this relationship was conditional on a significant three-way interaction with political ideology and social network heterogeneity (unstandardized Β = .05, p < .01). Figure 2 depicts the interaction between levels of discussion, political ideology, and social network heterogeneity. A post hoc examination of the three-way interaction was conducted for all combinations of liberal, moderate, and conservative political ideology; and low, medium, and high levels of social network heterogeneity (both calculated at 1 standard deviation below the mean, the mean, and 1 standard deviation above the mean for the respective variables). This revealed that at low levels of social network heterogeneity (i.e., more homogenous networks), higher levels of discussion did not affect support for climate mitigation policies among liberals (p = .33, LCI = −.047, UCI = .138) or moderates (p = .42, LCI = −.103, UCI = .043) but lowered levels of policy support among conservatives (p < .05, LCI = −.198, UCI = −.013). This partially supports Hypothesis 2a, as conservatives in more homogenous social networks were less supportive of climate policy, as predicted. However, liberals in more homogenous social networks were not significantly more supportive of climate policy, which contrasts what was predicted in Hypothesis 2a.

Figure 2.

Interaction between level of discussion and political ideology at low, medium, and high levels of network heterogeneity in predicting support for policies to address global climate change.

At medium levels of social network heterogeneity, higher levels of discussion did not affect policy support among liberals (p = .33, LCI = −.112, UCI = .038) or moderates (p = .057, LCI = −.114, UCI = .002), but decreased policy support among conservatives (p < .05, LCI = −.151, UCI = −.001). Finally, at high levels of social network heterogeneity, higher levels of discussion led to decreased policy support among liberals (p < .05, LCI = −.220, UCI = −.019) and moderates (p < .05, LCI = −.160, UCI = −.006), but not conservatives (p = .38, LCI = −.151, UCI = .059). This partially supports Hypothesis 2b, as liberals were less supportive as predicted by Hypothesis 2b, but conservatives did not exhibit the positive change in policy support that was predicted by Hypothesis 2b.

Thus, higher levels of discussion led to increased polarization in policy support at low and medium levels of social network heterogeneity, as conservatives became more opposed to climate mitigation polices while liberal positions remained unchanged. At high levels of social network heterogeneity, political polarization was reduced as discussion increased, with the reduction in polarization accomplished primarily by liberals and moderates lowering their support for global warming policy closer to conservative levels of support.

In a test of Hypothesis 3a and Hypothesis 3b, the PROCESS SPSS macro was used to assess the bootstrapped indirect effect of hostile media perceptions on policy support through discussion at varying levels of political ideology and social network heterogeneity. Indirect effects were calculated for all combinations of liberal, moderate, and conservative political ideology; and low, medium, and high levels of social network heterogeneity (as with the post hoc examination of the three-way interaction, both calculated at 1 SD below the M, the M, and 1 SD above the M for the respective variables). The results, shown in Table 2, reveal that through the mediated pathway of discussion, hostile media perceptions had a significant negative impact on policy support for the following groups: (1) liberals with high levels of social network heterogeneity (LCI = −.027, UCI = −.001), in partial support of Hypothesis 3b; (2) moderates with medium (LCI = −.014, UCI = −.001) to high levels of social network heterogeneity (LCI = −.019, UCI = −.001); and (3) conservatives with low levels of social network heterogeneity (LCI = −.026, UCI = −.001), in partial support of Hypothesis 3a. The indirect effects of hostile media perceptions were not significant for any other combinations of the various levels of political ideology and social network heterogeneity.

Table 2.

Conditional Indirect Effect of Hostile Media Perceptions on Policy Support via Discussion at Values of Network Heterogeneity and Political Ideology.

Network heterogeneity	Political ideology	Indirect effect (boot SE)	95% Lower confidence interval	95% Upper confidence interval
Low	Liberal	.004 (.005)	−.003	.016
Low	Moderate	−.003 (.004)	−.013	.003
Low	Conservative	−.009 (.006)	−.026	−.001
Medium	Liberal	−.003 (.004)	−.013	.003
Medium	Moderate	−.005 (.003)	−.014	−.001
Medium	Conservative	−.006 (.004)	−.019	.001
High	Liberal	−.010 (.006)	−.027	−.001
High	Moderate	−.007 (.004)	−.019	−.001
High	Conservative	−.004 (.006)	−.017	.006

Note. For network heterogeneity and political ideology, respectively, “medium”/“moderate” values correspond to the sample mean, “low”/“liberal” values correspond to 1 SD below the mean, and “high”/“conservative” values correspond to 1 SD above the mean. Bootstrapped standard errors and confidence intervals were computed using 10,000 bootstrap samples. Boldface text is used to denote significant effects.

Finally, there was a significant direct relationship between hostile media perceptions and policy support that was moderated by political ideology (unstandardized Β = −.07, p < .01), consistent with Hypothesis 4. Figure 3 shows the nature of this interaction: As hostile media perceptions increased, the gap in policy support between liberals and conservatives widened. The interaction was probed using the Johnson-Neyman technique (Hayes, 2013), which found that the negative relationship between hostile media perceptions and policy support was significant for political ideology values at or above 4.0. This means that higher levels of hostile media perceptions did not directly impact climate mitigation policy support for liberals and moderates (political ideology values 1-3) but did significantly lower policy support for conservatives (political ideology values of 4 and 5). Thus, this found partial support for Hypothesis 4, as conservatives were less supportive—and thus more extreme—as predicted, but liberals’ policy support—although demonstrating a positive trend—did not manifest a significant, direct relationship with hostile media perceptions.

Figure 3.

Interaction between hostile media perceptions and political ideology in predicting support for policies to address global climate change.

Discussion

This study is the first that we are aware of to investigate a model in which hostile media perceptions may influence public opinion through the mediator of issue-based discussion with friends and family. All proposed hypotheses were supported or partially supported.

Hostile media perceptions had a significant positive relationship with predispositions to discuss the issue of climate change with friends and family, in support of Hypothesis 1.

The study demonstrated a nuanced relationship between discursive actions and support for climate mitigation policies, with a significant three-way interaction between levels of discussion, social network heterogeneity, and political ideology, which partially supported Hypothesis 2. For people with relatively homogeneous social networks, discussion about climate change amplified polarization between liberals and conservatives; conservatives were less supportive of climate mitigation policies in this condition, while unexpectedly, liberal positions were unchanged instead of having higher levels of policy support. The lack of a significant positive relationship between discussion and policy support for liberals with more homogeneous social networks (Figure 2 shows a nonsignificant positive relationship) may have been due to a ceiling effect, as liberals with homogeneous social networks showed high levels of policy support across levels of discussion. For people with relatively heterogeneous social networks, discussion reduced polarization. However, this reduction in polarization was achieved primarily by liberals and moderates moving their levels of policy support closer to the lower levels held by conservatives.

Looking next to overall indirect effects, the analysis revealed a significant indirect effect of hostile media perceptions on policy support via discussion that was conditional on political ideology and social network heterogeneity. Among those with low social network heterogeneity, hostile media perceptions increased polarization indirectly through discussion; however, this indirect path was significant only among conservatives. Among those with high social network heterogeneity, hostile media perceptions decreased polarization indirectly through discussion; in this case, however, the indirect effect was significant only among liberals and moderates.

The study also found evidence that, independent of the mediated pathway through discussion, hostile media perceptions were related to higher levels of political polarization on the issue of climate change. Although higher levels of hostile media perceptions were unrelated to policy support among liberals, among conservatives, higher levels of hostile media perceptions were related to lower levels of policy support. Thus, as hostile media perceptions increased, the gap in policy support between liberals and conservatives widened.

As with any study, there are limitations that should be taken into account. Data used in this study are cross-sectional, making unequivocal statements about causality impossible. However, past research using experimental and multiwave survey data confirms the causal direction running from discussion in heterogeneous social networks to political attitudes (Binder et al. 2009; Levitan & Visser, 2009; Visser & Mirabile, 2004). Still, future research may build from the present study with experimental approaches to directly examine how hostile media perceptions influence discursive activities and polarization, and how levels of discussion with groups of varying levels of social network heterogeneity may influence policy perspectives. Future research also may examine how additional variables, such as selective exposure, may mediate the link between hostile media perceptions and public opinion. We also note some limitations in the wording of the survey questions. The question measuring levels of discussion is not temporally specific, and it is possible that different respondents may have interpreted the question differently (e.g., past vs. future discussion). In addition, the discussion question jointly asks about discussion with friends and family, and these two types of social connections may not always overlap in terms of political discussion. The measure examining network heterogeneity asks about whether friends share the respondent’s views on climate change, whereas the discussion question asks about friends and family; while both questions concern the types of networks that are most commonly used to talk about political issues (Klofstad et al., 2009; Mutz, 2006), it is possible that this difference in wording (friends vs. friends and family) introduced noise into the analysis.

Finally, we note that the size of the coefficients is small. This is likely for two reasons. First, climate change is a highly polarizing issue with beliefs about climate change increasingly serving as an identity marker of what it means to be a liberal or conservative (M. C. Nisbet, 2009), making members of these groups highly resistant to change. Second, the mediated model examines how policy preferences may change as a result of hostile media perceptions driving more discussion on the issue. It is likely that individuals engaging in discussion on the issue as a result of hostile media perceptions are more partisan and are engaging in these types of discussion as a “corrective action” (Rojas, 2010) to account for perceived negative media bias; this dynamic means that these individuals are likely to be more resistant to changing their policy preferences in light of conversations they may have. Nevertheless, results presented here do show significant effects from this mediated pathway.

Overall, the results presented here contribute to the still nascent area of research examining the impacts that hostile media perceptions may have on political behavior and public opinion. First, the positive relationship we find between hostile media perceptions and discussion lends additional support to previous literature, suggesting that individuals with higher levels of hostile media perceptions are more likely to engage in “corrective actions” in order to counteract the presumed negative influence of media coverage (Hwang et al., 2008; Rojas, 2010). Second, to our knowledge, this is the first study to explicitly link hostile media perceptions to issue-based opinion polarization. As we show, hostile media perceptions are associated with increased attitude polarization, both directly and indirectly through discussion within homogeneous social networks. As public trust in the media, in general, declines (Ladd, 2011) and audiences fragment across myriad news outlets—thereby intensifying perceptions of hostility in less familiar and less ideologically friendly outlets (Arceneaux et al., 2012; Feldman, 2011)—these results raise concerns about the potential to reconcile partisan divides and reach compromise on critical political issues. Nonetheless, our findings offer evidence that when the discussion motivated by hostile media perceptions occurs in heterogeneous social networks, some individuals may be willing to moderate their views in light of exposure to alternative perspectives, thereby depolarizing political opinions.

The results also hold implications for science communication in general and, in particular, for efforts to engage strong partisans on the issue of climate change. Prior studies have shown that claims of media bias from political elites can drive public perceptions of media bias (Ladd, 2011; Watts, Domke, Shah, & Fan, 1999). This study suggests that this rhetorical strategy may, in turn, spur discussion and intensify opinions among like-minded partisans, which has been found to increase participation and activism (Mutz, 2002b). At the same time, our findings suggest that efforts to encourage meaningful deliberation on the issue of climate change by having individuals of diverse political orientations discuss the issue may lower political polarization but may do so by moving liberals and moderates closer to conservative positions in terms of policy support. For climate change advocates this would be viewed as a boomerang effect (Byrne & Hart, 2009; Hart & Nisbet, 2012), with a result that moves opposite of the intended direction. Overall, the results presented here show that conservatives, in particular, are resistant to changing opinions on climate change, as across levels of social network heterogeneity, more engagement on the issue led to either less policy support or no change. This pattern is consistent with the hypothesis—albeit a relatively controversial one—that conservatives are particularly dogmatic and likely to engage in motivated reasoning (Jost, Glaser, Kruglanski, & Sulloway, 2003), although this may be specific to certain topics such as climate change (E. C. Nisbet, Cooper, & Garrett, 2015). It is also in line with other research on political discussion, which has found that individuals holding conservative positions on issues are more likely to polarize as a result of discussion whereas those supporting liberal positions tend to accommodate their views (Wojcieszak & Price, 2010). This may pose a problem from the perspective of deliberative democracy, which requires all members of a public to be willing to listen to and seek compromise with the others.

In conclusion, this study demonstrates the importance for researchers to not only investigate the precursors to hostile media perceptions but also examine the effects of hostile media perceptions for politically charged issues. Significantly, this study is the first we are aware of to offer evidence that hostile media perceptions may drive individuals to engage in behavior such as discursive activities and that these behaviors can subsequently shift policy support on an issue. For critical issues such as global climate change, it is imperative that researchers continue to examine how perceptions of media coverage of the issue can affect attitudes and behavior as well as how best to engage with individuals with varying political ideologies.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Yale Project on Climate Change Communication and the George Mason University Center for Climate Change Communication and was funded by the Surdna Foundation, the 11th Hour Project, and the Grantham Foundation for the Protection of the Environment. Drs. Hart and Feldman were supported by the National Science Foundation (SES-1419604).

Notes

Author Biographies

P. Sol Hart is an assistant professor at the University of Michigan, Ann Arbor.

Lauren Feldman is an assistant professor at Rutgers University.

Anthony Leiserowitz is a research scientist and director of the Yale Project on Climate Change Communication at Yale University.

Edward Maibach is a university professor and director of the Center for Climate Change Communication at George Mason University.

References

Arceneaux

Johnson

Murphy

(2012). Polarized political communication, oppositional media hostility, and selective exposure. Journal of Politics, 74(1), 174-186. doi:10.1017/S002238161100123X

Arendt

(1968). Between past and future: Eight exercises in political thought. New York, NY: Viking Press.

Ariyanto

Hornsey

M. J.

Gallois

(2007). Group allegiances and perceptions of media bias taking into account both the perceiver and the source. Group Processes & Intergroup Relations, 10, 266-279. doi:10.1177/1368430207074733

Arpan

L. M.

Raney

A. A.

(2003). An experimental investigation of news source and the hostile media effect. Journalism & Mass Communication Quarterly, 80, 265-281. doi:10.1177/107769900308000203

Baron

R. S.

Hoppe

S. I.

Kao

C. F.

Brunsman

Linneweh

Rogers

(1996). Social corroboration and opinion extremity. Journal of Experimental Social Psychology, 32, 537-560. doi:10.1006/jesp.1996.0024

Baum

M. A.

Gussin

(2008). In the eye of the beholder: How information shortcuts shape individual perceptions of bias in the media. International Quarterly Journal of Political Science, 3(1), 1-31.

Binder

A. R.

Dalrymple

K. E.

Brossard

Scheufele

D. A.

(2009). The soul of a polarized democracy testing theoretical linkages between talk and attitude extremity during the 2004 presidential election. Communication Research, 36, 315-340. doi:10.1177/0093650209333023

Burnstein

Vinokur

(1977). Persuasive argumentation and social comparison as determinants of attitude polarization. Journal of Experimental Social Psychology, 13, 315-332. doi:10.1016/0022-1031(77)90002-6

Byrne

Hart

P. S.

(2009). The boomerang effect: A synthesis of findings and a preliminary theoretical framework. In Beck

(Ed.), Communication yearbook (Vol. 33, pp. 3-38). Mahwah, NJ: Lawrence Erlbaum.

10.

Callegaro

DiSogra

(2008). Computing response metrics for online panels. Public Opinion Quarterly, 72(5), 1008-1032. doi:10.1093/poq/nfn065

11.

Choi

Yang

Chang

J. J.

(2009). Elaboration of the hostile media phenomenon the roles of involvement, media skepticism, congruency of perceived media influence, and perceived opinion climate. Communication Research, 36, 54-75. doi:10.1177/0093650208326462

12.

Coe

Tewksbury

Bond

B. J.

Drogos

K. L.

Porter

R. W.

Yahn

Zhang

(2008). Hostile news: Partisan use and perceptions of cable news programming. Journal of Communication, 58, 201-219. doi:10.1111/j.1460-2466.2008.00381.x

13.

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(1), 111-126. doi:10.2307/2585932

14.

Davison

W. P.

(1983). The third-person effect in communication. Public Opinion Quarterly, 47(1), 1-15. doi:10.1086/268763

15.

Delli Carpini

M. X.

Cook

F. L.

Jacobs

L. R.

(2004). Public deliberation, discursive participation, and citizen engagement: A review of the empirical literature. Annual Review of Political Science, 7, 315-344. doi:10.1146/annurev.polisci.7.121003.091630

16.

Feldman

(2011). Partisan differences in opinionated news perceptions: A test of the hostile media effect. Political Behavior, 33, 407-432. doi:10.1007/s11109-010-9139-4

17.

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(1), 3-31.

18.

Feldman

Myers

Hmielowski

J. D.

Leiserowitz

A. A.

(2014). The mutual reinforcement of media selectivity and effects: Testing the reinforcing spirals framework in the context of global warming. Journal of Communication, 64, 590-611.

19.

Feldman

Price

(2008). Confusion or enlightenment? How exposure to disagreement moderates the effects of political discussion and media use on candidate knowledge. Communication Research, 35, 61-87. doi:10.1177/0093650207309362

20.

Giner-Sorolla

Chaiken

(1994). The causes of hostile media judgments. Journal of Experimental Social Psychology, 30, 165-180. doi:10.1006/jesp.1994.1008

21.

Gunther

A. C.

(1998). The persuasive press inference effects of mass media on perceived public opinion. Communication Research, 25, 486-504. doi:10.1177/009365098025005002

22.

Gunther

A. C.

Christen

C. T.

(2002). Projection or persuasive press? Contrary effects of personal opinion and perceived news coverage on estimates of public opinion. Journal of Communication, 52, 177-195. doi:10.1111/j.1460-2466.2002.tb02538.x

23.

Gunther

A. C.

Christen

C. T.

Liebhart

J. L.

Chia

S. C.-Y.

(2001). Congenial public, contrary press, and biased estimates of the climate of opinion. Public Opinion Quarterly, 65, 295-320. doi:10.1086/322846

24.

Gunther

A. C.

Liebhart

J. L.

(2006). Broad reach or biased source? Decomposing the hostile media effect. Journal of Communication, 56, 449-466. doi:10.1111/j.1460-2466.2006.00295.x

25.

Gunther

A. C.

Miller

Liebhart

J. L.

(2009). Assimilation and contrast in a test of the hostile media effect. Communication Research, 36, 747-764. doi:10.1177/0093650209346804

26.

Habermas

(1989). The theory of communicative action. Boston, MA: Beacon Press.

27.

Hansen

G. J.

Kim

(2011). Is the media biased against me? A meta-analysis of the hostile media effect research. Communication Research Reports, 28, 169-179. doi:10.1080/08824096.2011.565280

28.

Hart

P. S.

Nisbet

E. C.

(2012). Boomerang effects in science communication: How motivated reasoning and identity cues amplify opinion polarization about climate mitigation policies. Communication Research, 39, 701-723. doi:10.1177/0093650211416646

29.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression based approach. New York, NY: Guilford Press.

30.

Huckfeldt

Mendez

J. M.

Osborn

(2004). Disagreement, ambivalence, and engagement: The political consequences of heterogeneous networks. Political Psychology, 25(1), 65-95. doi:10.1111/j.1467-9221.2004.00357.x

31.

Hwang

Pan

Sun

(2008). Influence of hostile media perception on willingness to engage in discursive activities: An examination of mediating role of media indignation. Media Psychology, 11(1), 76-97. doi:10.1080/15213260701813454

32.

Jang

S. M.

Hart

P. S.

(2015). Polarized frames on “climate change” and “global warming” across countries and states: Evidence from Twitter big data. Global Environmental Change, 32, 11-17. doi:10.1016/j.gloenvcha.2015.02.010

33.

Jost

J. T.

Glaser

Kruglanski

A. W.

Sulloway

F. J.

(2003). Political conservatism as motivated social cognition. Psychological Bulletin, 129, 339-375. doi:10.1037/0033-2909.129.3.339

34.

Kim

K. S.

(2011). Public understanding of the politics of global warming in the news media: The hostile media approach. Public Understanding of Science, 20, 690-705. doi:10.1177/0963662510372313

35.

Klofstad

C. A.

McClurg

S. D.

Rolfe

(2009). Measurement of political discussion networks a comparison of two “Name Generator” procedures. Public Opinion Quarterly, 73, 462-483. doi:10.1093/poq/nfp032

36.

Ladd

J. M.

(2011). Why Americans hate the media and how it matters. Princeton, NJ: Princeton University Press.

37.

Levitan

L. C.

Visser

P. S.

(2008). The impact of the social context on resistance to persuasion: Effortful versus effortless responses to counter-attitudinal information. Journal of Experimental Social Psychology, 44, 640-649. doi:10.1016/j.jesp.2007.03.004

38.

Levitan

L. C.

Visser

P. S.

(2009). Social network composition and attitude strength: Exploring the dynamics within newly formed social networks. Journal of Experimental Social Psychology, 45, 1057-1067. doi:10.1016/j.jesp.2009.06.001

39.

McCright

A. M.

Dunlap

R. E.

(2000). Challenging global warming as a social problem: An analysis of the conservative movement’s counter-claims. Social Problems, 47, 499-522.

40.

McCright

A. M.

Dunlap

R. E.

(2003). Defeating Kyoto: The conservative movement’s impact on US climate change policy. Social Problems, 50, 348-373.

41.

McCright

A. M.

Dunlap

R. E.

(2011). The politicization of climate change and polarization in the american public’s views of global warming, 2001–2010. Sociological Quarterly, 52, 155-194. doi:10.1111/j.1533-8525.2011.01198.x

42.

Morris

J. S.

(2007). Slanted objectivity? Perceived media bias, cable news exposure, and political attitudes. Social Science Quarterly, 88, 707-728. doi:10.1111/j.1540-6237.2007.00479.x

43.

Moscovici

Zavalloni

(1969). The group as a polarizer of attitudes. Journal of Personality and Social Psychology, 12, 125-135. doi:10.1037/h0027568

44.

Mutz

D. C.

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

45.

Mutz

D. C.

(2002b). The consequences of cross-cutting networks for political participation. American Journal of Political Science, 46, 838-855. doi:10.2307/3088437

46.

Mutz

D. C.

(2006). Hearing the other side: Deliberative vs. participatory democracy. New York, NY: Cambridge University Press.

47.

Myers

T. A.

(2011). Goodbye, listwise deletion: Presenting hot deck imputation as an easy and effective tool for handling missing data. Communication Methods and Measures, 5, 297-310. doi:10.1080/19312458.2011.624490

48.

Nisbet

E. C.

Cooper

K. E.

Garrett

R. K.

(2015). The partisan brain: How dissonant science messages lead conservatives and liberals to (dis)trust science. Annals of the American Academy of Political and Social Science, 658(1), 36-66. doi:10.1177/0002716214555474

49.

Nisbet

M. C.

(2009). Communicating climate change: Why frames matter for public engagement. Environment, 51(2), 12-23.

50.

Nisbet

M. C.

Kotcher

M. S.

(2009). A two-step flow of influence? Opinion-leader campaigns on climate change. Science Communication, 30, 328-354.

51.

Parsons

B. M.

(2010). Social networks and the affective impact of political disagreement. Political Behavior, 32, 181-204. doi:10.1007/s11109-009-9100-6

52.

Perloff

R. M.

(1989). Ego-involvement and the third person effect of televised news coverage. Communication Research, 16, 236-262. doi:10.1177/009365089016002004

53.

Price

Cappella

J. N.

Nir

(2002). Does disagreement contribute to more deliberative opinion? Political Communication, 19(1), 95-112. doi:10.1080/105846002317246506

54.

Rojas

(2010). “Corrective” actions in the public sphere: How perceptions of media and media effects shape political behaviors. International Journal of Public Opinion Research, 22, 343-363. doi:10.1093/ijpor/edq018

55.

Scheufele

D. A.

Nisbet

M. C.

Brossard

Nisbet

E. C.

(2004). Social structure and citizenship: Examining the impacts of social setting, network heterogeneity, and informational variables on political participation. Political Communication, 21, 315-338. doi:10.1080/10584600490481389

56.

Schmitt

K. M.

Gunther

A. C.

Liebhart

J. L.

(2004). Why partisans see mass media as biased. Communication Research, 31, 623-641. doi:10.1177/0093650204269390

57.

Slater

M. D.

(2007). Reinforcing spirals: The mutual influence of media selectivity and media effects and their impact on individual behavior and social identity. Communication Theory, 17, 281-303. doi:10.1111/j.1468-2885.2007.00296.x

58.

Stroud

N. J.

(2011). Niche news: The politics of news choice (1st ed.). New York, NY: Oxford University Press.

59.

Sunstein

C. R.

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

60.

Taber

C. S.

Lodge

(2006). Motivated skepticism in the evaluation of political beliefs. American Journal of Political Science, 50, 755-769.

61.

Tesser

Conlee

M. C.

(1975). Some effects of time and thought on attitude polarization. Journal of Personality and Social Psychology, 31, 262-270. doi:10.1037/h0076292

62.

Tsfati

(2007). Hostile media perceptions, presumed media influence, and minority alienation: The case of Arabs in Israel. Journal of Communication, 57(4), 632-651. doi:10.1111/j.1460-2466.2007.00361.x

63.

Tsfati

Cohen

(2005). Democratic consequences of hostile media perceptions: The case of Gaza settlers. The Harvard International Journal of Press/Politics, 10(4), 28-51. doi:10.1177/1081180X05280776

64.

Tsfati

Cohen

Gunther

A. C.

(2011). The influence of presumed media influence on news about science and scientists. Science Communication, 33, 143-166.

65.

Vallone

R. P.

Ross

Lepper

M. R.

(1985). The hostile media phenomenon: Biased perception and perceptions of media bias in coverage of the Beirut massacre. Journal of Personality and Social Psychology, 49, 577-585.

66.

Visser

P. S.

Mirabile

R. R.

(2004). Attitudes in the social context: The impact of social network composition on individual-level attitude strength. Journal of Personality and Social Psychology, 87, 779-795. doi:10.1037/0022-3514.87.6.779

67.

Watts

M. D.

Domke

Shah

D. V.

Fan

D. P.

(1999). Elite cues and media bias in presidential campaigns explaining public perceptions of a liberal press. Communication Research, 26, 144-175. doi:10.1177/009365099026002003

68.

Wojcieszak

Price

(2010). Bridging the divide or intensifying the conflict? How disagreement affects strong predilections about sexual minorities. Political Psychology, 31, 315-339. doi:10.1111/j.1467-9221.2009.00753.x