Conservatives and Misinformation

Jessica R. Collier

Since the 2016 U.S. presidential election, concern and interest in misinformation, oft referred to as “fake news,’’has increased exponentially with academics, journalists, and politicians alike being concerned with its consumption and spread. The conversation is compounded when elites like former White House Social Media Director Dan Scavino, state that 85% of media coverage of the Trump administration is “fake news” and that “he [President Trump] is in the White House today because of social media” (O’Connor, 2017). Conflating social media, “fake news” and the Trump administration set alarming precedents for the general public attempting to distinguish between information and misinformation.

This chapter examines whether individuals are concerned with the status quo of news media coverage, specifically with regard to journalistic decisions about what to focus on, and the differentiation between opinion and fact in content. In addition to exploring this question broadly, the chapter uses two-wave panel survey data of American adults from 2016 and 2018 to analyze the interactions of election cycle (2016 versus 2018), partisanship, and individual media diets on the outcomes to address individual-level factors that may influence how individuals perceive and critique news media.

Audience Perspectives on Misinformation

The integral piece of the misinformation puzzle, beyond the algorithms and bots that industry are concerned with, are the individuals who consume, sometimes willingly, and subsequently share false information across predominantly online channels. The public’s definition of the term “misinformation” differs substantially from the ways in which academics choose to define it. The public definitions go beyond the misinformation/disinformation binary and expand broadly to include categories such as satire, false information, propaganda, advertising, hyper-partisan content, and even poor journalism (Nielsen & Graves, 2017). These inclusions cast a much larger net for defining misinformation in addition to raising questions about how individuals perceive news differently than academics and elites.

More audience-centric approaches to examining misinformation, or “fake news,” has cited the importance of individual actors, messages, and affordances across platforms (Marwick, 2018). A recent study utilized ethnographic approaches to observe how Republican groups make sense of mainstream media, concluding that the observed Republicans’ approach to messages mimics the way in which they, as religious conservatives, interpret messages in the Bible (Tripodi, 2018). Epistemological differences in approaches to mediated content interact with how individuals phrase online searches and therefore the results returned to them by algorithms that also associate certain language styles with alt-right groups. This process creates greater potential for false information to reach these individuals, especially if they are the targets of disinformation campaigns. When people vary in how they come to view and trust information, greater concern and attention to individual-level differences is warranted.

Partisan Differences and Misinformation

The ways in which misinformation is consumed and shared differ along partisan lines. This bifurcation of consumption and sharing patterns is attributable to differences between Democrats and Republicans as well as those elicited by partisan media outlets and political elites. From an elite perspective, rhetoric on certain issues can work to polarize the electorate and facilitate misperceptions by either sowing doubt along partisan lines with issues such as global warming/ climate change or by the sheer spread of misinformation as was the case with the Affordable Care Act (Pasek, Sood, & Krosnick, 2015; Tesler, 2018). Misperception belief may be exacerbated by the partisan cues that individuals receive from party elites working to polarize certain issues or policies.

In the case of corrections to misinformation to stem these effects, providing accurate information to conservatives about the existence of weapons of mass destruction (WMDs) in Iraq bolstered their belief in the misperception (Nyhan & Reifler, 2010). Referred to as a backfire or boomerang effect, the rationale for this difference is attributed to motivated reasoning and sticking with party beliefs, but it is worth noting that in many of these cases, conservatives are the group for whom the misperception persists (e.g., Cook & Lewan- dowsky, 2016; Hart & Nisbet, 2012). Additionally, conservatives are more likely to perceive fact-checkers as liberally biased and have negative opinions of fact-checking organizations (Nyhan & Reifler, 2015). Even if fact-checks have negative consequences for conservatives, this group is also the least likely to interact with them (Shin & Thorson, 2017), which suggests that the individuals most in need of fact-checked information are the least likely to be exposed. While the bulk of this research finds that distrust of fact-checks and media broadly is a primarily conservative issue, individuals across the political spectrum respond hostilely to fact-checks that denigrate a member of their political party (Shin & Thorson, 2017). If a hostile media bias exists across parties, a larger issue of disliking attitudinally incongruent information might explain observations thus far.

In addition to issues with getting fact-checked information to partisans, individuals are also selectively exposing themselves to like-minded media content which can perpetuate belief in misperceptions (Kull, Ramsay, & Lewis, 2003) as well as lead to an asymmetric difference in exposure to misinformation (Guess, Nyhan, & Reifler, 2018). Possible explanations for the asymmetry are that feelings of resentment, underdog status, and a desire to see stories that reflect personal beliefs drive conservatives particularly toward partisan media (Cramer, 2016; Hochschild, 2018; Marwick, 2018; Polletta & Callahan, 2017). Tripodi (2018) notes the influence of online videos by conservative commentators on conservative audiences and argues that with messages such as his, “the overall take away is that liberal ideology is formed by disputable claims and emotional appeals instead of fact-based evidence” (p. 39). This line of research indicates a distrust of mainstream media to provide important factual information: Conservatives will have greater agreement than liberals that news should be factual with no opinion or analysis (HIa). Additionally, I expect that conservatives will have greater agreement than liberals that news focuses on scandal and is manipulated by elected officials (Hlb).

Ethnographic and observational research supports findings from quantitative research that identify partisan media, like Fox News, as a symptom of polarization rather than a cause (Arceneaux & Johnson, 2013). Partisan media may not be a cause of polarization, but its ability to reach polarized audiences can work to promote misperceptions (Garrett, Weeks, & Neo, 2016) and prolong the lifespan of misinformation (Shin & Thorson, 2017). Exposure to Fox News is correlated with individuals possessing misperceptions about both climate change (Krosnick & Maclnnis, 2010) and the Iraq War (Kull et al., 2003) though research remains inconclusive on whether partisan media actually increases misperceptions. While Fox News is not synonymous with misinformation, conservatives are more likely to search for nonmainstream media alternatives, and partisan media like Fox are especially susceptible to the intermedia agenda-setting influence of “fake news” websites (Guo & Vargo, 2018). Content analyses of “fake news” reveal that the stories on these sites are not necessarily more partisan, but the right-leaning content utilizes the appeals identified by researchers as being most effective for reaching out to “underdog”conservative audiences (Marwick, 2018). Therefore, the narrative that conservative viewers may be getting from their trusted sources fits with the one that they would also receive from problematic online sources. I propose the following hypothesis: Individuals with conservative media diets will report greater concern for news accuracy than individuals with liberal media diets (H2).

The Role of Online and Social Media

Not only are individuals exposed to problematic content based on partisan- selective exposure, but they may interact with misinformation as a result of the media or sites that they seek out to consume political information. People are receiving substantial amounts of political information from social media like Facebook, even unwittingly (Settle, 2018). Given the widespread nature of information, individuals who seek out information—or at minimum are exposed to information from broader swaths of outlets and platforms—increase their opportunities for consuming as well as disseminating misinformation. This pattern possibility highlights the necessity of a broader perspective on individual media consumption, including traditional outlets as well as social media.

As noted earlier in the chapter, individuals are driven to consume media that aligns with their ideological views. However, individuals are not solely consuming ideologically consonant information, especially in the age of social media. On Twitter and Facebook, individuals have a significant number (~20%) of friends or followers who are ideologically dissimilar from themselves (Bakshy, Messing, & Adamic, 2015) and cross-cutting interaction on social media is actually quite frequent (Barbera, 2015). Individuals have perceptions that disagreement is greater on social media than in face-to-face interactions (Barnidge, 2017). Disagreement is not necessarily negative given that individuals will spend more time reading news content if it is accompanied by a comment containing a different opinion (Sulflow, Schafer, & Winter, 2018).

Despite perceptions of disagreement, individuals utilize social cues and endorsement heuristics that platforms provide. People are more likely to read news with social endorsements (Bakshy, Eckles, Yan, & Rosenn, 2012). Endorsements and social cues also mitigate the likelihood that individuals engage in partisan-selective exposure (Messing & Westwood, 2014). Therefore, individuals with media diets that include social media may be more likely to encounter disagreement, cross-cutting opinions, and media beyond that which they would typically consume. I expect that exposure to different viewpoints would correct for partisan-motivated perceptions such that: Social media users report less bias in media (H3) and less concern for journalistic focus and reporting of opinion (H4).

Differences in audiences, content, and affordances exist across Facebook, Twitter, and Instagram that might impact individuals. An extensive amount of misinformation advertising was shared over social media during the 2016 U.S. presidential campaign (Vosoughi, Roy, & Aral, 2018). The content of that advertising, specifically on Facebook, targeted contentious issues in battleground states and was linked to suspicious groups that indicate the involvement of foreign actors (Kim et al., 2018). Additionally, Facebook founder and CEO, Mark Zuckerberg, testified to the House Intelligence Committee. Facebook launched several initiatives aimed at transparency of paid advertising, and research investigated the role social media might have played in that election (Jamieson, 2018). The wealth of research suggesting that Facebook users had higher likelihoods of exposure to misinformation and the focus of popular culture on Facebook’s involvement and its users’ susceptibility led me to propose the following hypothesis: Facebook users will report greater concern for news and journalism in general than will individuals who are nonusers of social media or who use alternative platforms (H5).

Changes over Election Cycles

Attention to misinformation—specifically, the “fake news” variety'—has increased dramatically from the primaries held in May 2016 to the primaries held in May 2018 with examples in popular culture and multiple convenings of academics and industry professionals interested in stemming the problem. Sheer discussion around the topic might trickle down to individuals. Elite discourse about “fake news” can elicit lower levels of media trust and decreased accuracy in identifying real news (Van Duyn & Collier, 2018). Discourse questioning the veracity of news can have detrimental effects, and given that that discourse is still common, I hypothesize that: Individuals’ concerns about news media and journalism will be greater in 2018 than in 2016 (H6).

Method

The data for this study are part of the Texas Media & Society' (T-MASS) survey conducted by the Annette Strauss Institute for Civic Life at the University of Texas at Austin. The survey, fielded by the GfK Group (formerly Knowledge Networks) in both English and Spanish, utilized a probability' sample by recruiting panelists to participate in online surveys. Those individuals without internet access or a device were provided a web-enabled device to participate. Emails were sent on Days 3, 7, and 10 of the field periods to remind nonresponders to participate. Participants were 18 years or older and responses were weighted to be representative of the U.S. general population.

For this study, the 2016 and 2018 waves of the survey are used with a total of 584 participants from the United States. Response rates were 54% and 78.7% for the 2016 and 2018 wave, respectively. Data were collected between May 24 and June 14, 2016, and May 16 and May 31, 2018. The partisan breakdown for the sample used here is 48.6% liberal and 42.8% conservative. It is important to establish context for each of the survey periods with regard to both political climate and discussion of misinformation. During the 2016 data collection, party nominees had not yet been chosen but both Donald Trump and Hillary Clinton were the presumptive nominees. The phrase “fake news” had yet to be popularized by either party. Contrastingly, the 2018 data collection period marked the second year of the Trump administration during which the phrase “fake news” was used with regularity by journalistic and political elites.

Individuals responded to a variety of questions measuring their attitudes and concerns regarding journalism and news. Journalistic criticism measured participants’ agreement with four statements regarding their opinions about journalists and news media. These items included statements such as, “The news media are manipulated by elected officials who want to get media coverage.” Items were measured on a 5-point Likert scale from strongly agree (1) to strongly disagree (5). Items were summed and divided by four to create an ordinal measure ofjournalis- tic criticism with 0 being more critical and 5 being less critical. Factual importance asked participants to report their agreement with three statements concerning the presence of fact and opinion in news. Items included statements like, “The news media should just present the facts, without any analysis” and were measured on a 5-point Likert scale from strongly agree (1) to strongly disagree (5). Items were summed and divided by three to create an ordinal measure of importance of facts with 0 being more importance and 5 being less importance. Concern for news accuracy was measured by asking participants, “How often do you check the accuracy of the news you get?” Responses were measured on a 5-point scale with 1 being “all of the time” and 5 being “none of the time.”

Individuals responded to questions assessing their behaviors regarding media use as well. Social media diet was measured by asking participants to identify from a list of six options which social networking sites they use. Responses included Facebook, Twitter, Snapchat, Instagram, YouTube, and Reddit. Yes/ no response options were summed to create a scale measure of social media usership. Media diet was measured according to where individuals reported getting their news in the two weeks prior to the survey. Response options included network nightly news, local television news, local newspaper, Wall Street Journal, and social media (Facebook, Twitter, Reddit, Snapchat, Instagram, YouTube). Several media outlets were combined to create diets of liberal media (CNN, Huffington Post, MSNBC, network nightly news, NewsHour on PBS, New York Times, NPR, Reddit, and Washington Post) and conservative media (Breitbart, Drudge Report, Fox News Cable Channel, and Rush Lim- baugh) based on the values reflecting significant differences among partisans who report using each of those outlets for news consumption. Using a %² statistic, p < .05 for each outlet to be categorized as liberal or conservative. This method is consistent with previous research utilizing T-MASS Survey data (Stroud & Collier, 2018).

Finally, perceptions of media bias were measured by asking participants whether they perceived news media as a whole to have a liberal or conservative bias or neither. Responses were dummy coded for perceived bias (1) or no perceived bias (0). Participants also reported their party identification, which was recoded with leaners included in the partisan categories such that Republican includes participants who indicated identifying as either strong or leaning Republican and Democrat includes participants who identified as strong or leaning Democrat. Independents were excluded from the analyses where party identification was included.

Tables 12.1 and 12.2 offer information on this sample.

TABLE 12.1 Percentages of Democrats and Republicans Using News Outlets

Note. * p < 0.05 using yj statistic. Question wording for media use “From which sources did you get news IN THE PAST 14 DAYS, that is from (INSERT DAY OF THE WEEK) two weeks ago through today. If you are unsure, please DO NOT select it.” Partisanship assessed from 1 “strong Republican” to 7 “strong Democrat.” Strong, weak, and leaning Democrats and Republicans are combined. Independents were excluded from these analyses.

TABLE 12.2 Percentages of Democrats and Republicans using Social Outlets

Note, p < 0.05 using yj statistic. Question wording for media use “From which sources did you get news IN THE PAST 14 DAYS, that is from [INSERT DAY OF THE WEEK] two weeks ago through today. If you are unsure, please DO NOT select it.” Partisanship assessed from 1 “strong Republican” to 7 “strong Democrat.” Strong, weak, and leaning Democrats and Republicans are combined. Independents were excluded from these analyses.

Results

First, I evaluated whether conservatives and liberals differ on their critiques of journalism (e.g., what it should do/what it does too much) and the extent to which facts and opinion should be present in news coverage. Analyses showed that Democrats in 2016 (M = 2.53, SD = 0.74) were significantly less likely to agree that facts were important than Republicans in 2016 (M = 1.98, SD = 0.68), f(532) = 8.91, p < .001 (see Figure 12.1). This finding remained the same in 2018 with Democrats (M = 2.53, SD = 0.73) again placing less importance on facts than Republicans (M= 1.91, SD = 0.75), t(532) = 9.70, p < .001. These results supported HI a, which predicted that conservatives would report greater agreement than liberals that news should be factual with no opinion or analysis.

Hlb predicted that conservatives would have greater agreement than liberals that news focuses on scandal and is manipulated by elected officials. In 2016, there were no significant differences in reported levels ofjournalistic criticism between Democrats (M = 1.97, SD = 0.72) and Republicans (M = 1.98, SD = 0.70) (see Figure 12.2). In 2018, Republicans (M= 1.83, SD = 0.69) reported significantly greater levels of criticism of journalism than Democrats (M = 2.15, SD = 0.75), f(532) = 5.03, p < .001. Hlb is supported.

To test the influence of media diets on individuals’ concern for news accuracy (H2), each news outlet was first categorized as liberal, conservative, or neither, based on chi-square tests indicating significant differences in partisan use of the outlets for the 2018 survey. Nonpartisan media sources included Facebook, Ins- tagram, local newspaper, local television news, Snapchat, Twitter, Wall Street Journal, and YouTube. Liberal news media included CNN, Hujfington Post, MSNBC, network nightly news, New York Times, NewsHour on PBS, NPR, Reddit, and Washington Post. Conservative media included Breitbart, Drudge Report, Fox News, and Rush Limbaugh (see Tables 12.1 and 12.2). Of the 584 respondents,

FIGURE 12.1 Means for “Factual Importance” by Partisanship, pre- and post-Trump Note: Lower values indicate greater importance.

FIGURE 12.2 Means for “Journalistic Criticism” by Partisanship, pre- and post-Trump Note: Lower values indicate higher levels of criticism.

61.8% reported a liberal media diet while 23.8% reported a conservative media diet. I ran an OLS regression with conservative media diet, liberal media diet, and party identification as the categorical predictors and concern for news accuracy as the outcome variable. There is a main effect of conservative media diet (B = -0.17, SE = 0.13, p < .001) on concern for news accuracy when holding party identification constant. Party identification and liberal media diet do not have main effects on concern for news accuracy, supporting H2 which predicted that conservatives would have greater concern for news accuracy.

Next, I turned to examine how social media use might influence perceptions of bias and criticisms of journalism. To test how impactful social media use is on reporting bias in the news media as a whole, I used logistic regression with reported Facebook, Instagram, Reddit, Snapchat, Twitter, and YouTube use included in the model (“other” was used as the reference group). There were no significant effects of using one particular type of social media on perceptions of the news media being generally biased (H3). Additionally, there were no significant differences in journalistic criticism for social media users versus nonusers (H4). Additionally, I predicted that Facebook users would report greater levels of criticism of journalism than individuals who are nonusers of social media. This hypothesis was not supported (H5).

Lastly, to test differences across election cycles in individuals’journalistic criticism, I ran an OLS regression with the journalistic criticism measure at T_{ (2016) as a predictor variable, party identification as a control, and the journalistic criticism measure at T, (2018) as the outcome variable. Journalistic criticism at 1

(В = -0.22, SE = 0.03, p < .001) has a main effect on journalistic criticism at T, when holding party identification constant. Party identification also has a main effect (B = 0.39, SE = 0.04, p < .001) on journalistic criticism in the model. I also ran an OLS regression with the factual importance measure at T_f as a predictor variable, party identification as a control, and the factual importance measure at T, as the outcome variable. Factual importance at T₍ (B = 0.40, SE = 0.04, p < .001) has a main effect on factual importance at T, when holding party identification constant. Party identification also has a main effect (B = -0.24, SE = 0.03, p < .001) on factual importance in the model. H6 is supported.

Conclusion

Understanding individual-level factors that influence the starting point at which individuals seek out news is important for conceptualizing differences in consumption as well as why and how misinformation persists. This chapter examined how electoral cycle (2016 versus 2018), party identification, and social media usage in addition to broader news usage contribute to individuals’ attitudes regarding journalism, news accuracy, and the valuation of facts compared to opinion. There are three important findings to discuss from this chapter, which are that with regard to criticisms of news and importance placed on factual information: (1) partisanship matters in this relationship, (2) social media does not matter at the aggregate level, and (3) these effects appear to be increasing across the 2016 to 2018 election cycles.

The findings in this chapter suggest that asymmetric differences exist in the importance that partisan groups place on facts over opinion in news as well as the level of criticism that they report for journalism as a whole. The result points to Republicans who place importance on fact versus opinion at a significantly higher level than Democrats report. This finding is consistent in both 2016 and 2018, which suggests a fundamental difference in how opposing partisans value the inclusion of fact and opinion in news coverage and is consistent with previous research that highlights a difference in how Republicans and Democrats interrogate news (Tripodi, 2018). The finding is in line with expectations that Republicans might value factual information more and thus seek out nonmainstream alternatives in search of those facts. Ironically, the search for factual information might be a fruitless endeavor if it leads to greater possibility for consumption of misinformation.

Additionally, partisanship matters in how individuals critique journalism. In May 2016, there were no differences in the levels of criticism of journalism reported between Democrats and Republicans. However, in May 2018, these results were much different with Republicans reporting significantly higher levels of criticism of journalism than Democrats. While the findings support previous research that conservative audiences might be more critical of media than liberals (Jamieson & Cappella, 2008), the shift from 2016 to 2018 is important as it signals a dissatisfaction with media coverage that is particularly experienced by Republicans and has changed significantly in a relatively short period of time.

In addition to direct influences of partisanship, partisan media diets have significant effects as well. Previous research has found that individuals who are concerned with news accuracy, or have higher levels of media skepticism, seek out more nonmainstream news (Tsfati & Cappella, 2003). Though levels of media skepticism are not measured in this survey, the concept might explain why conservatives seek out alternative media (Tripodi, 2018). The findings in this chapter reveal that conservative media diets are significantly associated with greater concern for news accuracy while similar effects do not exist for liberal media diets or even for partisan identification overall. Rather, it appears that consumption of conservative media is uniquely associated with individuals’ concern for news accuracy in a way that is not explained solely by partisanship.

Additionally, categorization of the news sources based on partisanship reflected important shifts in the position of media outlets and increases in reported use between 2016 and 2018. For example, the Washington Post entered the space of liberal media with significantly more Democrats reporting using the outlet than Republicans compared to 2016 when it could be classified as nonpartisan. Reported media usage of each site increased dramatically in the two-year survey window as well with net increases in reported consumption above 7% for both MSNBC and the Washington Post. Although several media sources moved into polarized spaces, increases in media usage across outlets is an encouraging and unexpected finding.

While distinctly partisan media exert some influence on individual attitudes and self-report measures, there is no evidence for differences in basic criticisms of news or levels of importance placed on fact based on the type of social media use or the quantity of accounts reported. For social media, it is important to caveat the chapter’s findings by first noting that these effects are measured in the aggregate. Social media may matter more when individual-level factors are associated with posts and comments beyond broad questions that strictly interrogate whether individuals own a social media account and do not dig more deeply into how endorsement heuristics and other social cues might affect interactions and attitudes.

First, no significant differences exist between the type of social media individuals use and their perceptions of bias in news media, which presents encouraging findings that no particular outlet has an audience of users who might subscribe more highly to the belief of media bias—or rather, not at a rate that exceeds the level that users of any other outlet subscribe to a belief of media bias. Social media users are also not very different in their levels of journalistic criticism. Regarding social media more broadly, the number of platforms individuals used also had no effects on any attitudes toward news media, factual importance, or perception of bias measured in this survey. While there are mixed findings on the importance of partisanship and the role of social media in each of these relationships, an examination of the change in scores from 2016 to 2018 reveals that there are temporal differences suggesting journalistic criticism is increasing across these elections.

The findings from this chapter suggest that partisans are different in the ways in which they evaluate news and what is important within news, a difference that appears to be increasing over time. These findings point to social media use as perhaps not having the influence that academics and industry professionals assume it to have. Also, the results highlight the importance of consumption of partisan media over and above reported partisanship. These results have important implications for researchers interested in studying misinformation, particularly, as they point to a discrepancy between liberals and conservatives about what journalism should and should not be as well as an asymmetry in how each party values facts versus opinions. To test susceptibilities to misinformation, a common understanding of what is fact is necessary, but individuals also need to share respect for facts that the present study suggests might differ between parties. This finding presents the possibility for a fundamental difference between parties that researchers should consider when designing experiments and interventions aimed at further understanding or combating misinformation. One-size-fits-all solutions may not be possible if the targeted audiences differ on a basic level.

Despite these implications, there are several shortcomings of this study. There is a significant body of research that explores how social media is associated with exposure to political information (Settle, 2018), decreased partisan selective exposure (Messing & Westwood, 2014), and greater likelihood of exposure to false information (Guess et al., 2018). This study reduced media behaviors to dichotomous, self-report measures, which may not accurately reflect actual behavior.

Future research needs to test the implications of these findings. Most importantly, there is a difference between valuing factual information and actually heeding it. Self-report measures of valuation can admittedly be the result of social desirability or other influences. Future research needs to test whether an individuals’ valuation of facts in general influences whether they choose to believe a piece of information. An answer to this question could help to explain some of the differences observed between liberals’ and conservatives’ processing of information, or misinformation.

Discussion Questions

• 1. Flow do liberals and conservatives differ in their concerns about news?
• 2. What impact(s) might media diets have on how people think about and consume news?
• 3. What are the potential issues with liberals and conservatives having different concerns about news? Are these differences cause for concern?

References

Arceneaux, K., & Johnson, M. (2013). Changing minds or changing channels? Partisan news in an age of choice. Chicago, IL: University of Chicago Press.

Bakshy E., Eckles, D., Yan, R., & Rosenn, I. (2012). Social influence in social advertising: Evidence from field experiments. Proceedings of the 13th ACM Conference on Electronic Commerce, 146-161. doi:10.1145/2229012.2229027

Bakshy, E., Messing, S., Sc Adamic, L. A. (2015). Exposure to ideologically diverse news and opinion on Facebook. Science, 348(6239), 1130-1132. doi:10.1126/science. aaallOO

Barbera, P. (2015). Birds of the same feather tweet together: Bayesian ideal point estimation using Twitter data. Political Analysis, 23(1), 76-91. doi:10.1093/pan/mpu011

Barnidge, M. (2017). Exposure to political disagreement in social media versus face- to-face and anonymous online settings. Political Communication, 34(2), 302—321. doi: 1 0.1080/10584609.2016.1235639

Cook, J.. & Lewandowsky, S. (2016). Rational irrationality: Modeling climate change belief polarization using Bayesian networks. Topics in Cognitive Science, S(l), 160-179. doi:10.1111/tops.12186

Cramer, K. J. (2016). The politics of resentment: Rural consciousness in Wisconsin and the rise of Scott Walker. Chicago, IL: University of Chicago Press.

Garrett, R. K., Weeks, В. E., & Neo, R. L. (2016). Driving a wedge between evidence and beliefs: How online ideological news exposure promotes political misperceptions. Journal of Computer-Mediated Communication, 2/(5), 331-348. doi: 10.1111/ jcc4.12164

Guess, A., Nyhan, B., Sc Reifler, J. (2018). Selective exposure to misinformation: Evidence from the consumption of fake news during the 2016 US presidential campaign. European Research Council. Retrieved from www.dartmouth.edu/~nyhan/fake-news- 2016.pdf

Guo, L., & Vargo, C. (2018). “Fake news” and emerging online media ecosystem: An integrated intermedia agenda-setting analysis of the 2016 US presidential election. Communication Research, 47(2), 178-200. doi: 10.1177/0093650218777177.

Hart, P. S., Sc 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(b), 701-723. doi:10.1177/0093650211416646

Hochschild, A. R. (2018). Strangers in their own land: Anger and mourning on the American right. New York, NY: The New Press.

Jamieson, К. H. (2018). Cybenvar: How Russian hackers and trolls helped elect a President what we don’t, can’t, and do know. New York, NY: Oxford University Press.

Jamieson, К. H., Sc Cappella, J. N. (2008). Echo chamber: Rush Utnbaugh and the conservative media establishment. New York, NY: Oxford University Press.

Kim, Y. M., Hsu, J., Neiman, D., Kou, C., Bankston, L., Kim, S. Y., Heinrich, R., Barag- wanath, R., Sc Raskutti, G. (2018). The stealth media? Groups and targets behind divisive issue campaigns on Facebook. Political Communication, 33(4), 515-541. doi:10.

1080/10584609.2018.1476425

Krosnick,). A., Sc Maclnnis, B. (2010). Frequent viewers of fox news are less likely to accept scientists’ views of global warming. The Woods Institute for the Environment. Retrieved from http://woods.stanford.edu/docs/surveys/Global-Warming-Fox-News.pdf

Kull, S., Ramsay, C., Sc Lewis, E. (2003). Misperceptions, the media, and the Iraq war. Political Science Quarterly, 118(4), 569-598. doi:30035697

Marwick, A. (2018). Why do people share fake news? A sociotechnical model of media effects. Georgetown Law Technology Review, 2, 474.

Messing, S., Sc Westwood, S. J. (2014). Selective exposure in the age of social media: Endorsements trump partisan source affiliation when selecting news online. Communication Research, 4/(8), 1042-1063. doi: 10.1177/0093650212466406

Nielsen, R. K., Sc Graves, L. (2017). ‘News you don’t believe’: Audience perspectives on fake news. Reuters Institute for the Study ofJournalism Report. Retrieved from https:// reutersinstitute. politics, ox. ac. uk/sites/default/files/2017-10/Nielsen% 26Graves_ factsheet_1710v3_FINAL_ download, pdf

Nyhan, B., & Reifler, J. (2010). When corrections fail: The persistence of political misperceptions. Political Behavior, 32(2), 303-330. doi:10.1007/sll 109-010-9112-2 Nyhan, B., & Reifler, ). (2015). Estimating fact-checkings effects. American Press Institute. Retrieved from www.americanpressinstitute.org/wp-content/uploads/2015/04/ Estimating-Fact-Checkings-Effect.pdf

O’Connor, L. (2017, June 29). Here’s the man behind Donald Trump’s bizarre social media strategy. Huffington Post. Retrieved from www.huffingtonpost.com/entry/ dan-scavino-profile_us_595520e4e4b0da2c7321c758 Pasek, )., Sood, G., & Krosnick, J. A. (2015). Misinformed about the affordable care act? Leveraging certainty to assess the prevalence of misperceptions. Journal of Communication, 65(4), 660-673. doi:10.1111/jcom. 12165 Polletta, F., Sc Callahan, ). (2017). Deep stories, nostalgia narratives, and fake news: Storytelling in the Trump era. American Journal oj' Cultural Sociology, 5(3), 392-408. doi: 10.1007/978-3-319-95945-0_4

Settle, J. E. (2018). Frenemies: How social media polarises America. Cambridge, UK: Cambridge University Press.

Shin, J., Sc Thorson, K. (2017). Partisan selective sharing: The biased diffusion of fact- checking messages on social media. Journal of Communication, 67, 233-255. doi: 10.1111/ jeom. 12284

Stroud, N. J., Sc Collier, J. R. (2018). Selective exposure and homophily during the 2016 presidential campaign. In B. Warner, D. G. Bystrom, M. S. McKinney, Sc M. C. Ban- wart, (Eds.), An unprecedented election: Media, communication, and the electorate in the 2016 campaign (pp. 21-39). Santa Barbara, CA: Praeger.

Sulflow, M., Schafer, S., Sc Winter, S. (2018). Selective attention in the news feed: An eyetracking study on the perception and selection of political news posts on Facebook. New Media & Society, 21(1), 168-190. doi:1461444818791520 Tesler, M. (2018). Elite domination of public doubts about climate change (not evolution). Political Communication, 55(2), 306-326. doi:10.1080/10584609.2017.1380092 Tripodi, F. (2018). Searching for alternative facts: Analyzing scriptural inference in conservative news practices. Data & Society. Retrieved from https://datasociety.net/wp-content/ uploads/2018/05/Data_Society_Searching-for- Alternative-Facts.pdf Tsfati, Y., Sc Cappella, J. N. (2003). Do people watch what they do not trust? Exploring the association between news media skepticism and exposure. Communication Research, 30(5), 504-529. doi: 10.1177/0093650203253371 Van Duyn, E., Sc Collier, |. (2018). Priming and fake news: The effects of elite discourse on evaluations of news media. Mass Communication and Society, 22(1), 29-48. doi: 10.1 080/15205436.2018.1511807

Vosoughi, S., Roy, D., Sc Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 114/>—1151. doi: 10.1126/science.aap9559