The Politicization of Climate Science: Media Consumption, Perceptions of Science and Scientists, and Support for Policy

1

Abstract

Differential media treatment of climate change, including conservative media’s tendency to reject the anthropogenic climate change scientific consensus, has reinforced polarized perceptions of climate change. Studies have found differences in coverage patterns and in perceptions among those relying more heavily on conservative rather than liberal or moderate media. This scholarship has been limited by narrow measurements of media exposure, climate-related outcomes, and the mechanism of effects. We analyzed nationally representative US data (N = 1,181) using measures that included not only reported use of mainstream print, cable, and social media captured in past research, but also science programming, as well as far-right, alternative-health, and Christian media. On average, participants relied more heavily on centrist and liberal media, followed by Fox News and social media. The results corroborate findings associating exposure to centrist media with pro-climate attitudes, and conservative media, including Fox News with the opposite views. Use of far-right outlets was associated with the lowest levels of belief in anthropogenic climate change, perceptions of personal threat, and support for climate-friendly policy. Reliance on science media was associated with pro-climate views. Most associations were mediated via perceptions of science and scientists (using the Factors Associated with Self-Presentation of Science, FASS scale).

Climate change and political polarization have been characterized as two of the twenty-first century’s critical socio-political issues (Falkenberg et al., 2022). Critically, the two are intertwined. Most people are exposed to information about science through the media (Hwang & Southwell, 2009; Ophir, 2018). As such, the coverage of scientific topics can increase awareness (Jacobsen & Jacobsen, 2011) and knowledge (Ophir & Jamieson, 2018) about scientific topics while also shaping perceptions of risk and vulnerability to threats (Fung, Namkoong, & Brossard, 2011). However, Americans are exposed to an array of content from a wide range of media, and not all outlets adhere to the same standards and norms when covering issues related to climate change. Because the media can put scientific topics on the public agenda (Schmidt, Ivanova, & Schäfer, 2013) and media frames can affect public understanding of issues (Grundmann & Scott, 2014; Trumbo, 1996), the media on which a person relies for information about climate change may reinforce and shape one’s views, intentions, and behaviors (Bromley-Trujillo & Poe, 2020; Chinn, Hart, & Soroka, 2020; Feldman, Maibach, Roser-Renouf, & Leiserowitz, 2012; Hornsey, Harris, Bain, & Fielding, 2016).

One’s ideology affects one’s media exposure (Iyengar & Hahn, 2009), a selective exposure effect more pronounced on average among conservatives than moderates or liberals (Haenschen, 2021). In 2014, Pew researchers found that those they characterized as consistent conservatives were more reliant on a single news source (i.e. Fox News) and expressed greater distrust than trust in most other news sources than consistent liberals (Mitchell, Gottfried, Kiley, & Eva Matsa, 2014). Subsequent research supported the notion that Republicans are prone to distrust the so-called mainstream media, and rely instead on ideologically congruent conservative media (Benkler, Faris, & Roberts, 2018). Such reliance could lead to increased likelihood to accept misinformationfor example, in the area of climate change (Bolin & Hamilton, 2018). For example, during the COVID-19 pandemic, exposure to conservative media was associated with belief in conspiracy theories (Romer & Jamieson, 2021). Outside the realm of health, in the aftermath of the January 6th 2021 riot, consumption of conservative media was associated with belief that the election had been stolen from the incumbent Donald J. Trump (Levendusky et al., 2023). Both liberals and conservatives may believe some specific conspiracy theories and misinformation, depending on the context, yet studies do suggest that susceptibility to climate change denial is higher among conservatives (Enders et al., 2023).

At the same time, studies have found differences in the climate-related content of liberal and moderate media outlets when compared to conservative ones, such as Fox News, which tend to cast doubt on the scientific consensus around the existence of anthropogenic climate change (Feldman, Maibach, Roser-Renouf, & Leiserowitz, 2012). The selective exposure of conservatives to an ideologically homogeneous range and types of media content and outlets that challenge the scientific consensus about climate change has exacerbated partisan gaps in belief in climate change and support for climate policy (Hornsey, Harris, Bain, & Fielding, 2016), which in turn could increase the reliance on such congruent sources. This may result in a self-reinforcing cycle of polarization (Slater, 2007).

This paper contributes to the growing literature by a) examining self-reported exposure to media outlets largely overlooked in previous work, including far-right, Christian alternative health, and scientific outlets; b) exploring the mechanism underlying the relationship between exposure and climate perceptions, specifically how exposure to different media is associated with perceptions of science and scientists (using the FASS scale suggested by Ophir, Walter, Jamieson, & Jamieson, 2023); c) using a broader set of outcomes, including perception of personal climate-related threats, belief in anthropogenic climate change, and support for a carbon tax, and; d) examining these relationships using a large-scale nationally representative empaneled sample of US nationals.

Exposure to Different Media and Perceptions of Climate Change

Americans are politically divided over the existence and causes of global warming, the risks that such changes pose, and the actions needed to ameliorate them (Smith & Mayer, 2019; Tyson, Funk, & Kennedy, 2022). This fact is particularly concerning since the US remains a prominent carbon polluter (World Population Review, n.d.), the need to dramatically reduce carbon pollution is urgent (IPCC, n.d.), and belief in anthropogenic climate change can drive intentions to change one’s behavior and to support climate-friendly policies (Bord, O’Connor, & Fisher, 2000; Hornsey, Harris, Bain, & Fielding, 2016). Misbeliefs about climate change could hinder public attempts to reduce carbon pollution.

In this study, we examine support for carbon taxation, a step that would reduce carbon emissions by assessing fees per tons of CO2 released (Köppl & Schratzenstaller, 2022). Failing to authorize such taxes removes a strong incentive from the menu of options able to reduce carbon emissions.

Adopted by 196 Parties, including the US, at the UN Climate Change Conference (COP21) in Paris, France, in December 2015, the Paris Climate Accords had as their goal holding “the increase in the global average temperature to well below 2°C above pre-industrial levels” and pursue efforts “to limit the temperature increase to 1.5°C above pre-industrial levels” (UNCC., n.d.). Although as a signatory of to that agreement, the US pledged to cut its greenhouse gas emissions (GHG) 50–52% from their 2005 levels by the end of this decade (Detrow, 2021), the research firm Rhodium Group reported in January 2024 that US emissions are currently just 17.2% below 2005 levels (Daly, 2024).

Following decades of disinformation, misinformation and lobbying (Oreskes & Conway, 2011), conservatives have been prone to downplay the existence and extent of human-caused climate change (Kahan, 2015), perceive its risks to be lower, and resist policies addressing it (Gromet, Kunreuther, & Larrick, 2013), including taking steps toward making the US carbon-neutral (Tyson, Funk, & Kennedy, 2022). A March 2022 poll by Pew Research Center shows that only 23% of Republicans report seeing global warming is a major threat, compared to 78% among liberals (54% for all US adults)(Tyson, Funk, & Kennedy, 2023). Pew’s 2023 survey data demonstrate that Republicans remain much less likely than Democrats to prioritize climate action (Pasquini, Spencer, Tyson, & Funk, 2023). As expected, partisan differences predict differences in perceived-risk (Bromley-Trujillo & Poe, 2020) and differential support for policy actions (Ranney & Clark, 2016).

As most Americans are exposed to science through media, scholars have examined whether partisan gaps in climate perceptions could result, in part, from differences in media choices. Research on a wide range of media has documented the existence of such effects. Coverage of climate change in newspapers reflects and reinforces the increased politization and polarization of beliefs about climate change over time (Chinn, Hart, & Soroka, 2020), with political actors increasingly featured and scientific ones less often the focus and the rhetoric of Democrats and Republicans becoming markedly more distinct. Unsurprisingly, this politization and polarization is reflected in national public opinion polls (Tyson, Funk, & Kennedy, 2023). Reporting on hurricanes and wildfires has been politicized as well (Molder & Calice, 2023).

Studies have demonstrated the ability of the media to shape public understanding of climate change (Krosnick, Holbrook, Lowe, & Visser, 2006). study of news coverage of the 1995 Intergovernmental Panel on Climate Change’s (IPCC) report, the first to clearly attribute increased global temperatures to human activity, demonstrated that media exposure-shaped audience perceptions. While national network news featured the certainty of that finding; newspapers featured uncertainty by counterbalancing that conclusion with the views of scientists who questioned the finding. The researchers found that among Ohioans, television news exposure was positively associated and newspaper exposure negatively associated with belief in global warming (Krosnick, Holbrook, Lowe, & Visser, 2006). Analyses of coverage in large circulation US newspapers found concerning differences in coverage of climate change across outlets of different ideological slants (Boykoff, 2007; Feldman, Hart, & Milosevic, 2017). Also, viewing conservative media outlets such as Fox News has been associated with reduced belief in global warming and anthropogenic climate change while exposure to MSNBC, CNN, and political comedy has been positively associated with such beliefs (Feldman, Maibach, Roser-Renouf, & Leiserowitz, 2012; Hmielowski, Feldman, Myers, Leiserowitz, & Maibach, 2014; Krosnick & MacInnis, 2010; Meirick, 2022). Similar polarizing patterns exist in mainstream social media platforms, such as Twitter (Falkenberg et al., 2022). At the same time, the use of mainstream social media as a news source has been associated with a reduction in climate skepticism (Diehl, Huber, Gil de Zúñiga, & Liu, 2021), though some studies have suggested the effect of social media could be polarizing, presumably due to confirmation bias and the ability to engage with information affirming one’s prior beliefs (Meirick, 2022). Empirical evidence suggests conservative media use decreases trust in scientists, which decreases certainty that global warming is happening. Non-conservative media use is associated with the opposite effects (Hmielowski, Feldman, Myers, Leiserowitz, & Maibach, 2014). Our study builds upon these findings while significantly expanding the scope of inquiry, as detailed below.

The Current Study

Existing literature shows that one’s media exposure is associated with and could be plausibly understood to influence one’s climate change-associated perceptions, beliefs, and behaviors and that the effects in recent years have coincided with increased polarization of the issue. However, prior research was mostly limited in three ways. First, it looked only at mainstream media such as the broadcast networks and New York Times, conservative media such as Fox News and widely popular social media like Facebook and Twitter. Hence, little is known about the relationship between climate perceptions and exposure to content that falls outside the scope of mainstream media. Here we look specifically at extremist, religious, alt-health, and science outlets. Second, researchers have concentrated their attention on associations between media exposure and belief in the existence, extent, and human cause of climate change and to a lesser extent on associations between media exposure and climate-related risk which we define as the perception that climate change poses a serious threat to the respondent or their way of life during their lifetime, and to support for policies that can ameliorate such risks. Finally, while the role of trust in scientists has been probed (Hmielowski, Feldman, Myers, Leiserowitz, & Maibach, 2014), with a meta-analysis finding a positive relationship between the levels of trust, media use, and beliefs consistent with the scientific consensus on climate change (Bogert, Buczny, Harvey, & Ellers, 2023), its measurement tends to be circumscribed, often using single measurement items. However, a wide range of other science-specific perceptions, including that scientists are credible, prudent, honor scientific norms, and produce beneficial findings, that may mediate the association between media exposure and these dependent variables has not been studied. Because these factors have been associated with support for science funding (Ophir, Walter, Jamieson, & Jamieson, 2023), determining whether they affect climate beliefs is a worthwhile task. We examine the relationship between exposure to media and three dependent variables, which we describe next. We ask:

RQ1:

Research Question 1: How does self-exposure to different media outlets relate to a) belief in anthropogenic climate change; b) perceived personal climate-related threats; and c) support for a carbon-tax.

The research discussed above found important associations between exposure to different media and perceptions of climate change. However, little is known about the mechanisms underlying such relationships. Here, we argue that the association between media exposure and the set of dependent variables described above could be mediated through perceptions of science and scientists. Studies have shown that media exposure can reinforce and shape perceptions of science and scientists, and such perceptions are associated with beliefs about climate change, perceptions of threat, and support for climate policy. When it comes to perceptions and support for climate action, researchers have found that gaps between conservatives’ trust in climate scientists and in the evidence they produce could play a key role (Bugden, 2022; Douglas & Sutton, 2015; Druckman & McGrath, 2019; Merk, Pönitzsch, Kniebes, Rehdanz, & Schmidt, 2015). While important, these studies tended to rely on single items to measure perceptions of science, usually asking participants about their trust in scientists.

The Factors Assessing Science’s Self-presentation (FASS) scale proposed instead 13 items that capture various aspects of such perceptions, specifically that scientists and science are Credible, Prudent, Self-correcting, Unbiased [able to overcome biases], and that science is Beneficial. In past studies, some of these components were associated with science-related policy making in general and in the context of climate change (Chinn, Hasell, Roden, & Zichettella, 2023; Petrovic, Madrigano, & Zaval, 2014). The full rationale for the scale is beyond the scope of this discussion and can be found elsewhere (Ophir, Walter, Jamieson, & Jamieson, 2023). The specific items used can be seen in the Measurement section below. Drawing on the factors associated with the self-presentation of science items, we hypothesized that:

H1:

Hypothesis 1: The relationships between self-exposure to different types of media and the dependent variables will be mediated by the perceptions of science and scientists.

Method

Data were collected via the Annenberg Science and Public Health survey (ASAPH) conducted using an online national probability panel survey drawn from a larger SSRS national probability panel of US adults aged 18 and older. The survey was reviewed and approved by the University of Pennsylvania’s Institutional Review Board on March 23, 2021, and was deemed to meet the eligibility criteria for IRB review exemption authorized by 45 Code of Federal Regulations (CFR) 46.104, category 2. The data reported are from a large-scale panel study tracking perceptions of science, scientists, and public health and climate beliefs over. Specifically, self-reported exposure to media and perceptions of science and scientists were measured on March 2023 (n = 1,638). Perceptions related to climate were measured on November 28^th 2023 (n = 1,538). The data analyzed here consist of only respondents who participated in both surveys (N = 1,417). For our estimations of direct and indirect effects we retained only complete cases, which resulted in a final data analysis of 1,181 cases.

Measures

Media exposure was measured using the following question: “How much information do you get from each of the following sources? Use a scale from 0 to 5, where 0 means you get “No information” from these sources, and 5 means you get “A lot of information” from these sources. Participants responded to this question in regard to the following media (for each we provided a couple of examples). Importantly, we did not label sources as “centrist,” “far-right” etc. when presenting them to participants. We use these here for readers’ convenience only.

The political slant of channels was determined based on a combination of polls showing audience preference (Grieco, 2020), academic studies (Groseclose & Milyo, 2005) and evaluations by the nonpartisan source, allsides.com, that was shown in recent studies to be consistent with other sources on the specific outlets we included (Wang, Thimmanayakanapalya, & Ophir, 2024). We used Allsides.com to identify far-right sources, including Breitbart, Newsmax, and OAN. To them we added platforms that were created as hyper-conservative alternatives to mainstream social media and had become hubs for extremism, including Telegram (Rogers, 2020), Gettr (Mekacher, Falkenberg, & Baronchelli, 2023), Gateway Pundit (Haller, Holt, & de La Brosse, 2019), Truth Social, Parler (Zhang, Lukito, Suk, & McGrady, 2024), and as an additional exemplar we added Mark Levin’s shows to the list (Dicker, 2024). While there are different ways to evaluate the exact slant of an outlet (e.g., whether MSNBC is strongly liberal or somewhat liberal), different methods tend to agree on the general slant of the specific outlets we use here. Note that since the conservative bias of the Wall Street Journal is limited to its editorial sections, we asked participants about their exposure to the news pages specifically. Alternative health media was defined as “information sources that promote and amplify non-mainstream, non-standard treatments and health practices unsanctioned by such custodians of U.S. health information,” such as the CDC and the FDA (Stecula, Motta, Kuru, & Jamieson, 2022, p. 376).

1. Centrist or mainstream media: “sources such as CBS News, NBC News, ABC News, CNN, the Associated Press, the news pages of the Wall Street Journal, or the New York Times”

2. Conservative media: Fox News

3. Far-right media: “sources such as Newsmax, One America News (OAN), Gateway Pundit, Parler, Telegram, GETTR, Breitbart News, Mark Levin or Life, Liberty and Levin, and the social media site known as Truth Social.”

4. Liberal media: “sources such as MSNBC or Huffington Post”

5. Science media: “sources such as Nova, Popular Science, or Scientific American”

6. Christian media: “sources such as Christian Broadcasting Network (CBN), or Daystar Television Network.”

7. Alternative health: “social media accounts dedicated to alternative health, such as Natural News or Earth Clinic.”

8. Social media: Mainstream social media platforms, “sources such as Facebook, Instagram, Twitter, and TikTok.”

Perceptions of science and scientists (Cronbach’s alpha = .894) was measured as the average of 13 items, taken from the FASS scale proposed by Ophir, Walter, Jamieson, and Jamieson (2023), with all items measured from 1 (strongly disagree) to 5 (strongly agree). Items included: “scientists in general are competent.” “scientists in general are trustworthy,” “scientists in general share my values,” “scientists in general feel superior to others,” “science creates unintended consequences and replaces older problems with new ones,” “scientists do whatever it takes to get grants and publish, even if that means cutting corners,” “scientists provide the public with unbiased conclusions about their area of inquiry, “in general, when scientists are doing their work, they are able to overcome their human and political biases,” “Just your best guess, when fraud in research is uncovered, how often does the scientific community take action to prevent other researchers from engaging in the same type of conduct?,” “Just your best guess, when there are mistakes in research, how often are they caught by the peer-review process in which scientists evaluate each other’s work before it is published?,” “Just your best guess, when scientists make a mistake, how often do they take responsibility for it?,” “Scientific findings produced by U.S scientists in the past decade have benefited the country as a whole,” and “Scientific findings produced by U.S scientists in the past decade have benefited people like me.”

Political ideology was measured from 1 (strongly conservative) to 5 (strongly liberal).

Belief in anthropogenic climate change was measured by asking “do you believe increases in the Earth’s temperature over the last century are due more to: 1) the effects of pollution from human activities; 2) natural changes in the environment that are not due to human activities; or 3) climate change is not happening. We then created a binary variable based on respondents acceptance of the scientific consensus around anthropogenic climate change. Participants who believed there is no evidence for climate change and those believing climate changes were more due to natural changes were coded as 0. Those endorsing the statement that climate change is happening and that its main contribution is human activity were coded as 1.

Perceived threat was measured by asking participants “to what extent, if at all, do you think that climate change already poses or will pose a serious threat to you or your way of life in your lifetime,” with answers ranging from 1 (not at all) to 5 (a great deal), with an option to state that “climate change is not happening” (6). We recoded 6 to be the lowest degree of perceived-threat, thus creating a variable from 1 (climate change is not happening) to 6 (a great deal of threat).

Support for carbon tax policy was measured using the item “how much do you favor or oppose taxing corporations based on the amount of carbon emissions they produce to reduce the effects of climate change,” with answers ranging from 1 (strongly oppose) to 5 (strongly favor).

Results

To examine how, if at all, the relationship between self-exposure to different media outlets relates to a) belief in anthropogenic climate change; b) perceived climate-related threat; and c) support for a carbon-tax (RQ1), we conducted multiple regressions. To test whether these relationships were mediated via perceptions of science and scientists, we conducted an analysis of direct and indirect effects in R. Descriptive statistics for each variable, including media consumption, dependent variables, the mediator, and the covariate controlled (ideology) can be found in Table 1. In items related to media exposure, 0 indicates no information from the source and 5 indicates a lot of information. As can be seen in the table, on average, the sources respondents relied on the most were centrist media, liberal media, Fox News, and social media (in that order). Reliance on far-right, Christian, science, and alternative-health was, on average, lower.

Table 1. Measurements and descriptive statistics.

Item	M	SD
Media Consumption
Far-Right Media	0.246	0.667
Centrist Media	2.253	1.342
Social Media	0.870	1.008
Liberal Media	1.378	1.600
Fox News	1.150	1.521
Science Oriented Media	1.036	1.406
Christian Media	0.271	0.832
Alt-Health Media	0.360	0.941
DVs
Climate threat Perception	4.412	1.404
Belief in Anthro Climate Change	0.742	0.438
Support for Carbon Tax	3.900	1.335
Mediator
Perceptions of Science	3.455	0.701
Covariate
Political Ideology (5=Liberal)	3.125	1.138

First, we examined which variables predicted belief in anthropogenic climate change, controlling for political ideology and perceptions of science and scientists.SThe model (Table 2) explained 38.3% of the variance in the belief (adjusted R-squared). The analysis shows that exposure to centrist media (B = .060, p < .001) and social media (B = .028, p = .016) was associated with increased belief in anthropogenic climate change. The belief was negatively associated with exposure to Fox News (B = −.028, p < .001) and far-right media (B = −.098, p < .001). The more participants identified as liberal (B = .107, p < .001) and the more positive their perceptions of science and scientists were (B = .096, p < .001), the more they reported believing that the climate was warming due to human behavior. The causal mediation analysis (Table 3) revealed that the relationship between the belief and far-right (proportion mediated = .092), Fox News (.171) and centrist media (.211) was mediated via perceptions of science and scientists. The relationship between social media use and the belief was not mediated via these perceptions (p > .05).

Table 2. Regression predicting belief in anthropogenic climate change controlling for political ideology and perceptions of science and scientists.

	Belief in Anthropogenic Climate Change
	B	Std. Error	t	p
(intercept)	−.048	.062	−0.781	=.435
Political Ideology	.107	.011	9.613	<.001***
Perceptions of Science	.096	.018	5.307	<.001***
Far-Right Media	−.098	.019	−5.089	<.001***
Centrist Media	.060	.011	5.371	<.001***
Social Media	.028	.011	2.414	=.016*
Liberal Media	.008	.009	0.899	=.369
Fox News	−.028	.008	−3.342	<.001***
Science Media	.006	.008	0.705	=.481
Christian Media	.004	.014	0.252	=.801
Alt-Health Media	.008	.013	0.598	=.550

F-statistic: 74.36 on 10 and 1,170 DF, p-value: < .001

Adjusted R-squared: .383

* p < .05, ** p < .01, *** p < .001

Note: Adj. R-sq value for the model without political ideology was .335

Table 3. Causal mediation analysis for a relationship between media and belief in anthropogenic climate change via perceptions of science and scientists.

	Estimate
	Far-Right	Centrist	Social	Liberal	Fox News	Science	Christian	Alt-Health
ACME	−.011***	.016***	.001	.001	−.006***	.007***	−.005+	−.004
ADE	−.098***	.060***	.028***	.008	−.028***	.006	.002	.009
Total-Effect	−.108***	.076***	.029*	.008	−.034***	.013	−.003	.005
Prop Mediated	.092***	.211***	.030	.058	.171***	.463	.060	−.168

* p < .05, ** p < .01, *** p < .001

Next, we examined which variables predicted perceptions of threat associated with climate change. The model (Table 4) explained 42.9% of the variance in the belief (adjusted R-squared). Exposure to centrist media (B = .190, p < .001), science media (B = .073, p = .004), and Christian media (B = .112, p = .012) was associated with increased perception of threat. Threat perception was negatively associated with exposure to Fox News (B = −.106, p < .001) and far-right media (B = −.321, p < .001). The more participants identified as liberal (B = .360, p < .001) and the more positive their perceptions of science and scientists were (B = .337, p < .001), the more they reported perceiving themselves under threat from climate change. The causal mediation analysis (Table 5) revealed that the relationship between perceptions of threat and far-right (proportion mediated = .096), Fox News (.142), centrist (.233) and science media (.242) was mediated via perceptions of science and scientists. The relationship between Christian media use and the belief was not mediated via these perceptions (p > .05).

Table 4. Regression predicting climate change threat perceptions controlling for political ideology and perceptions of science and scientists.

	Climate-related Threat Perception
	B	Std. Error	t	p
(intercept)	1.680	.191	8.78	<.001***
Political Ideology	.360	.034	10.803	<.001***
Perceptions of Science	.337	.056	6.014	<.001***
Far-Right Media	−.321	.060	−5.374	<.001***
Centrist Media	.190	.034	5.512	<.001***
Social Media	.051	.035	1.443	=.149
Liberal Media	.007	.026	0.272	=.786
Fox News	−.106	.025	−4.218	<.001***
Science Media	.073	.025	2.911	=.004**
Christian Media	.112	.045	2.509	=.012*
Alt-Health Media	.068	.040	1.685	=.092

F-statistic: 89.55 on 10 and 1,170 DF, p-value: < 0.001

Adjusted R-squared: 0.429

* p < .05, ** p < .01, *** p < .001

Note: Adj. R-sq value for the model without political ideology was .372

Table 5. Causal mediation analysis for a relationship between media and perceived threat via perceptions of science and scientists.

	Estimate
	Far-Right	Centrist	Social media	Liberal	Fox News	Science	Christian	Alt-Health
ACME	−.036***	.058***	.004	.003	−.018***	.024***	−.020*	−.015
ADE	−.322***	.186***	.047	.008	−.109***	.074***	.113***	.073*
Total-Effect	−.358***	.243***	.051	.011	−.127***	.097***	.093*	.058+
Prop Mediated	.096***	.233***	.080	.051	.142**	.242***	−.191+	−.224

Finally, we examined which variables predicted support for a carbon tax. The model (Table 6) explained 45.4% of the variance in support for the policy (adjusted R-squared). Exposure to centrist (B = .155, p < .001), social (B = .072, p = .028), Christian (B = .143, p < .001), and alternative health (B = .088, p = .020) media was associated with increased support for the policy. Support for the policy was negatively associated with exposure to Fox News (B = −.103, p < .001) and far-right media (B = −.390, p < .001). The more participants identified as liberal (B = .364, p < .001) and the more positive their perceptions of science and scientists were (B = .400, p < .001), the more they reported perceiving themselves under threat from climate change. The causal mediation analysis (Table 7) revealed that the relationship between support for the policy and far-right (proportion mediated = .099), Fox News (.183), and centrist media (.297) was mediated via perceptions of science and scientists. The relationship between Christian and alternative health media use and support for the policy was not mediated via these perceptions (p > .05).

Table 6. Regression predicting support for carbon tax controlling for political ideology and perceptions of science and scientists.

	Support for Carbon Tax
	B	Std. Error	t	p
(intercept)	1.124	.178	6.319	<.001***
Political Ideology	.364	.032	11.452	<.001***
Perceptions of Science	.400	.052	7.662	<.001***
Far-Right Media	−.390	.055	−7.036	<.001***
Centrist Media	.155	.032	4.833	<.001***
Social Media	.072	.033	2.196	=.028*
Liberal Media	.013	.025	0.518	=.604
Fox News	−.103	.023	−4.418	<.001***
Science Media	−.026	.023	−1.1	=.272
Christian Media	.143	.041	3.456	<.001***
Alt-Health Media	.088	.038	2.329	=.020*

F-statistic: 99.25 on 10 and 1,170 DF, p-value: < 0.001

Adjusted R-squared: 0.454

* p < .05, ** p < .01, *** p < .001

Note: Adj. R-sq value for the model without political ideology was .394

Table 7. Causal mediation analysis for a relationship between media and support for carbon tax via perceptions of science and scientists.

	Estimate
	Far-Right	Centrist	Social	Liberal	Fox News	Science	Christian	Alt-Health
ACME	−.043***	.067***	.003	.003	−.023***	.028***	−.025*	−.018*
ADE	−.386***	.155***	.066*	.011	−.103***	−.028	.141***	.081*
Total-Effect	−.429***	.222***	.069*	.015	−.126***	−.0003	.116*	.062+
Prop Mediated	.099***	.297***	.053	.070	.183***	−.504	−.235+	−.266

Discussion

This study examined associations between Americans’ media exposure and climate change-related dependent variables, including reported beliefs in anthropogenic climate change, perceptions of threat, and support for a carbon tax. We build upon and extend prior research on the topic by evaluating not just mainstream media outlets, but also far-right sources, as well as Christian, alternative health, and scientific media. For each, we examined the relationship between exposure and a set of dependent variables: the belief in anthropogenic climate change, perception of the threats climate change poses to the respondent and their way of life, and support for carbon tax policy. To better understand the mechanism of such relationships between exposure and climate-related outcomes, we tested whether each was mediated via perceptions of science and scientists, using the FASS scale (Ophir, Walter, Jamieson, & Jamieson, 2023). We examined these research questions using a large, nationally representative US sample.

Our results show that on average, respondents relied the most on centrist media, liberal media, Fox News, and social media (in that order). Reliance on far-right, Christian, science, and alternative-health media was, on average, lower. The variables of interest (media exposure, ideology, perceptions of science and scientists) explained a moderately large share of the variance in each of the dependent variables (ranging from 38.3% to 45.4%). Belief in anthropogenic climate change was positively associated with exposure to centrist and social media, but negatively associated with exposure to Fox News and far-right media. All relationships but those with social media exposure were partially mediated via perceptions of science and scientists. As for threat perceptions, exposure to centrist, science or Christian media led to higher perceptions of threat, and exposure to Fox News and far-right media to lower threat perceptions. All relationships except for with Chrisitan media were partially mediated via perceptions of science and scientists. Finally, support for a carbon tax was positively associated with exposure to centrist, social, Christian, and alternative health media. It was negatively associated with exposure to Fox News and far-right media.

We postulated that the positive relationship between exposure to Christian media and perceptions of personal threat and support for the carbon tax policy may have been the result of holding constant-related variables such as ideology and perceptions of science. Therefore, we examined the bivariate correlations between all variables, finding that, as we suspected, exposure to Christian media was negatively correlated with these dependent variables, when not controlling for ideology, perceptions of science and scientists, and exposure to other media. Moreover, while exposure to liberal media was not associated with any of the dependent variables in our models, the bivariate relationships were significant. This relationship only disappeared when controlling for ideology and perceptions of science. All bivariate correlations, including between exposure to different media, can be seen in Table 8.

Table 8. Bivariate correlations between the media variables, dependent variable and controls

	Perceptions of Science	Far-Right	Centrist	Social	Liberal	Fox	Science	Christ	AltH	Threat	Belief	Ideology
Far-Right	−0.29***
Centrist	0.49***	−0.13***
Social	−0.02	0.25***	0.13***
Liberal	0.35***	−0.04	0.66***	0.10***
Fox	−0.34***	0.44***	−0.12***	0.19***	−0.10***
Science	0.26***	0.12***	0.36***	−0.04	0.35***	−0.06
Christ	−0.16***	0.46***	0.05	0.26***	0.07*	0.29***	0.15***
AltH	−0.10***	0.40***	0.12***	0.42***	0.13***	0.24***	0.17***	0.43***
Threat	0.48***	−0.29***	0.46***	0.05	0.34***	−0.36***	0.23***	−0.06*	0
Belief	0.46***	−0.31***	0.44***	0.06	0.33***	−0.34***	0.16***	−0.12***	−0.04	0.64***
Ideology	0.46***	−0.29***	0.41***	0.06	0.33***	−0.44***	0.13***	−0.14***	−0.05	0.55***	0.52***
Carbon Tax	0.50***	−0.34***	0.44***	0.07*	0.32***	−0.38***	0.14***	−0.07*	−0.01	0.63***	0.64***	0.57***

Taken together, these results suggest that media exposure plays an important role in facilitating beliefs about climate change and the actions needed to mitigate the personal threat it poses. We show that this relationship tends to be partially mediated via perceptions of scientists and science captured in the FASS scale. The FASS scale consists of perceptions of scientists’ credibility, prudence, ability to overcome bias, disposition to correct error, and production of socially and personally beneficial work. In other words, our study shows that reliance on specific media for information is associated with perceptions about the science and scientists, which in turn, are associated with belief in anthropogenic climate change, perceived personal threat, and support for policy to mitigate this threat. As expected, liberals, on average, reported higher values on all three dependent variables.

Limitations

Importantly, since the data are cross-sectional, we cannot establish directionality or a causal order between media exposure and perceptions. Relatedly, we did not assign exposure to media, and thus the results should be interpreted in light of participants’ self-selective exposure to media. Nevertheless, based on prior work we postulate that media exposure could reinforce and shape public perceptions of science and scientists, in general and in relation to climate change specifically. Prior work pointed to the possibility of self-reinforcing cycles, where selective exposure and ideology influence each other in ways that can increase polarization (Slater, 2007). In future studies, experimental designs controlling the type of media participants are exposed to, could disentangle such causal ambiguity. In addition, we did not examine the content of climate and science coverage in each of the outlets considered. Although prior research has done so for mainstream, conservative, and liberal media, less is known about discussion of climate change in far-right platforms, Christian media, and alternative health podcasts and websites. Future comparative content analysis could better evaluate the different narratives about science and climate change in such partisan media.

Conclusion and Implications

The results of this study suggest that climate science scholars and advocates should pay more attention to the complex media diets of individuals and specifically of partisans, to better understand the possible influence of messages and narratives about climate science and scientists circulating in the American media environment. A better understanding of the resistance to science in conservative media in general, and in far-right platforms in particular (Ophir et al., 2022; Walter, Ophir, Lokmanoglu, & Pruden, 2022) could guide effective messaging aimed at countering misinformed narratives while also facilitating science-consistent coverage and audience acceptance of science-supported beliefs and policy. Our results indicate a need to continue studying the evolving messages and frames used in the rising alternative and far-right media. This will allow the construction of counter-narratives that could bolster trust in and compliance with science and scientists. Our findings regarding the mediating role of perceptions of science and scientists point to a need to better communicate the values of science. Future studies should evaluate the nuanced differences between conservatives and liberals in perceptions of science and scientists and identify ways to tailor communications about science in general, and climate change, in particular, for specific audiences.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Data Availability Statement

Data and code will be made available upon request.

Additional information

Funding

The panel data on which this study relies are drawn from the Annenberg Science and Public Health survey (ASAPH) funded by the Annenberg Health and Risk Communication Institute (AHRCI) endowment of the Annenberg Public Policy Center (APPC) of the University of Pennsylvania.