Neutralizing the Effect of Political Worldviews by Communicating Scientific Agreement: A Thought-Listing Study

Abstract

Messages emphasizing scientific agreement are increasingly used to communicate politically polarizing issues. Proponents argue that these messages neutralize the effect of people’s political worldviews due to the neutral scientific character of the message. Yet this argument has not undergone extensive testing. Addressing this, we measured participants’ thoughts on scientists featured in messages emphasizing scientific agreement on politically dissonant issues. Our results show that readers often produce less favorable thoughts and moral judgments when scientists agree on a politically dissonant issue. As a result, messages emphasizing scientific agreement on politicized issues might not always neutralize the effect of people’s political worldviews.

Keywords

science communication environmental communication climate change nuclear power

Attempts to improve science communication are increasingly challenged by skeptical audiences. For example, in polls of American Association for the Advancement of Science (AAAS) scientists and American adults, Pew Research has documented wide belief gaps between the lay public and scientists for a variety of issues. Political ideology can widen these gaps such that conservatives and liberals diverge from the scientific community for topics like climate change and nuclear power, respectively (Funk & Rainie, 2015a).

To help close the belief gaps, scholars have explored the importance of communicating a scientific consensus. Indeed, cross-sectional research on climate change (Ding, Maibach, Zhao, Roser-Renouf, & Leiserowitz, 2011; Hornsey, Harris, Bain, & Fielding, 2016; Lewandowsky, Gignac, & Vaughan, 2013; McCright, Dunlap, & Xiao, 2013) and vaccination (Dixon & Clarke, 2013) suggests that one’s perception of scientific agreement acts as a significant predictor of one’s belief in a scientific topic. As a result, messaging that emphasizes scientific agreement has been offered as a tool for improving views on contentious and politicized scientific topics, such as climate change (Cook, 2016; Lewandowsky, Gignac, & Vaughan, 2013; van der Linden, Leiserowitz, Feinberg, & Maibach, 2015) and vaccination (van der Linden, Clarke, & Maibach, 2015). Critical to the success of these messages is that they neutralize the effect of readers’ political worldviews such that the scientists forming the consensus are viewed in neutral rather than unfavorable terms (van der Linden, Leiserowitz, et al., 2015). While scholars have made this argument (see Cook, 2016; Lewandowsky Gignac, & Vaughan, 2013; van der Linden, Leiserowitz, et al., 2015; van der Linden, Leiserowitz, & Maibach, 2018), no study has specifically examined participant thought listings toward scientists featured in consensus messages. This article fills this gap by testing the neutralizing impact of communicating simple scientific agreement on two topics—one that is politically dissonant to conservatives (i.e., climate change) and another that is politically dissonant to liberals (i.e., nuclear power). Doing so can provide greater insight on whether, and to what extent, communicating scientific agreement can neutralize the effect of political worldviews for issues dissonant to those on the political left and right.

Dissonant Science and Motivated Reasoning

When confronted with disconfirming evidence for a strongly held belief, people rarely evaluate the information in a rational, objective manner. Instead, people are motivated to preserve their prior beliefs by processing dissonant information through a biased lens via processes like selective exposure, psychological reactance, and counterarguing (Kunda, 1990; Lodge & Taber, 2000; Nisbet, Cooper, & Garrett, 2015). Described as directional motivated reasoning, this form of biased processing often renders persuasive science messages ineffective, as well as results in boomerang effects whereby people strengthen their prior views (Byrne & Hart, 2009; Bolsen & Druckman, 2015; Hart & Nisbet, 2012; Kunda, 1990; Nyhan & Reifler, 2010; Zhou, 2016). Directional motivated reasoning is of particular concern in the context of politicized science, where issues like climate change and nuclear energy continue to remain polarized along ideological lines (Funk & Rainie, 2015a).

Emphasizing scientific agreement has been offered as a way to reduce directional motivated reasoning for polarizing topics. Indeed, scientists’ and the general public’s views on a variety of science, health, and environmental issues often diverge significantly (Funk & Rainie, 2015b). Many of these gaps are further accentuated by political ideology, where conservatives are less likely than liberals to believe in anthropogenic climate change and liberals are less likely than conservatives to support nuclear power (Funk & Rainie, 2015a). For both climate change and nuclear power, scientific support is strong among scientists, with a majority of scientists endorsing anthropogenic climate change, as well as supporting building more nuclear power plants in the United States (Anderegg, Prall, Harold, & Schneider, 2010; Cook, 2016; Rainie & Funk, 2015). Highlighting this scientific agreement could reduce the likelihood of directional motivated reasoning, resulting in depolarization of the public’s scientific views.

Cross-sectional research has found that people’s perception of what scientists believe often correlates with their personal beliefs on a variety of topics, including climate change (Ding et al., 2011; Hornsey et al., 2016; Lewandowsky, Gignac, & Vaughan, 2013; McCright, Dunlap, & Xiao, 2013), environmental policy (Aklin & Urpelainen, 2014) and vaccination (Dixon & Clarke, 2013). In situations where individuals underestimate the amount of scientific agreement for a topic, emphasizing scientific agreement to decrease the gap between perceived and actual scientific support might act as leverage in shifting views to be more in line with scientists. In fact, highlighting a scientific consensus surrounding climate change has been documented to improve scientific acceptance even for audiences most likely to resist persuasive attempts, such as free market supporters and political conservatives (Cook, 2016; Lewandowsky, Gignac, & Vaughan, 2013; van der Linden, Leiserowitz, et al., 2015). As a result, it has been argued that consensus messaging can neutralize the effect of political worldviews due to the “neutral scientific character of the [consensus] message” (van der Linden, Leiserowitz, et al., 2015, p. 6).

However, consensus messaging research has been met with criticism on methodological and theoretical grounds (Kahan, 2015; Pearce et al., 2017). One of most prominent critiques comes from Kahan’s cultural cognition thesis, which argues that a scientific consensus inconsistent with a cultural group’s position will likely be met with selective dismissal of evidence by that cultural group, as well as unfavorable credibility assessments of the scientists forming the consensus (Kahan, 2015; Kahan, Jenkins-Smith, & Braman, 2011). Cultural cognition draws on the cultural theory of risk and the psychometric paradigm to explore how cultural values (and their interaction with the psychological mechanisms of the psychometric paradigm) produce opposing risk perceptions in different people (Douglas & Wildavsky, 1983, Kahan et al., 2011; Slovic, 1987). For example, individuals might more readily recall experts who hold views that are consistent with their cultural value, while dismissing experts holding inconsistent views. This cultural availability effect can be caused by overrepresenting value-consistent experts in one’s mental inventory through selective exposure of the information environment, and through biased assimilation whereby judgments of scientific expertise are processed in ways that reinforce prior views (Kahan et al., 2011). As a result, a person’s assessment of a scientific consensus is likely processed through a biased lens, which proposes that consensus messaging will do little at reducing polarization for highly politicized topics.

Recent studies suggest the possibility of this perspective, finding consensus messages are ineffective in shifting conservatives’ views on climate change (Deryugina & Shurchkov, 2016; Dixon, Hmielowski, & Ma, 2017) or shifting nonsupporters’ views on genetically modified foods (Dixon, 2016). Consensus messaging could also act as a catalyst for increased polarization. For instance, messages emphasizing scientific agreement on anthropogenic climate change were found to backfire for individuals possessing strong free market ideology (Cook & Lewandowsky, 2016), as well as for Republicans with high knowledge in politics, science, and energy (Bolsen & Druckman, 2016), thus leading to greater polarization surrounding climate change belief.

Cultural cognition can provide insight into these inconsistent findings on consensus messaging. First, Cook and Lewandowsky’s (2016) finding shows a difference in the effect of free market values in Australian and American samples. Simple consensus messaging produced strong effects for Australian free market believers, similar to Lewandowsky, Gignac, and Vaughan’s (2013) finding. However, their reported backfiring effects among American free market holders suggest a potentially unique politicization of climate change in the United States that makes American free market believers more resistant to climate change messaging than those in other countries. This could be due to American free market believers viewing climate scientists as less credible because of their perceived incompatible worldviews. Second, political organizations have been noted to campaign against consensus statements and characterize scientists as biased individuals. Indeed, as consensus messaging is increasingly being used to persuade climate change skeptics, conservative politicians, think tanks, and activists have attempted to discredit its validity, arguing that there is no scientific consensus on climate change (Bast & Spencer, 2014). Instead of being portrayed as neutral actors, climate scientists are painted as partisan advocates. The same pattern has occurred for nuclear power, with liberal media outlets painting scientists who support nuclear power as shills for the nuclear power industry (Greenworld, 2016). The prevalence of these refutations provides greater counterarguing ammunition for those whose views diverge from the scientific consensus. As a result, these refutations might inoculate certain audiences against consensus messaging by prompting them to form unfavorable views of the scientists who form the consensus.

This article addresses this idea further by exploring how dissonant consensus messages are processed by individuals. In doing so, we test an important facet of consensus messaging research on whether scientists forming a consensus are viewed in neutral or biased terms.

Study 1

Study 1 examines participants’ thoughts on messages emphasizing scientific agreement on a liberal-dissonant and conservative-dissonant topic. While past research has shown that dissonant science messaging prompts resistance (e.g., Nisbet et al., 2015) and motivated interpretation (Washburn & Skitka, 2017) among conservatives and liberals, such work has not explored how this resistance might manifest as biased thoughts about scientists, such as those featured in messages emphasizing scientific agreement. As noted in previous consensus messaging research, for this type of messaging to be effective, audiences would need to evaluate the featured scientists similar to how they would view a scientist in a neutral context (Cook, 2016; Lewandowsky, Gignac, & Vaughan, 2013, van der Linden, Leiserowitz, et al., 2015). With conflicting findings, questions remain on whether communicating a simple message of scientific agreement for a politically dissonant topic neutralizes the effect of political worldviews. Therefore, we investigate the following research question:

Research Question 1: Do thoughts about scientists become more negative when these scientists agree on a politically dissonant topic?

Method

Sample

A diverse sample of 310 adult American participants was recruited by Survey Sampling International.¹ This Survey Sampling International panel roughly matched the demographic distribution of American adults as detailed by the 2010 U.S. census: mean age = 43.05 (SD = 16.7); White = 71%; and gender = 54.2% female. This study was conducted online from March 8 to 10, 2017.

Procedure

After first answering a series of questions measuring demographics and political ideology, participants were randomly assigned to one of three message conditions—conservative-dissonant message, liberal-dissonant message, and a neutral message.

The conservative-dissonant message condition consisted of a short message emphasizing scientific agreement on human-caused climate change (“Did you know? A majority of scientists believe that climate change is occurring and human-caused”). Conservatives are much less likely than moderates and liberals to endorse anthropogenic climate change, making it an ideal conservative-dissonant message (Funk & Rainie, 2015a). Furthermore, there exists strong agreement among scientists that human-caused climate change is happening (Anderegg et al., 2010; Cook, 2016; Rainie & Funk, 2015).

The liberal-dissonant message consisted of the same type of message but emphasized scientific agreement on nuclear power (“Did you know? A majority of scientists favor building more nuclear power plants in the U.S.”). Research has consistently found liberals are less supportive of nuclear power than moderates and conservatives (Funk, 2017; Funk & Rainie, 2015a). Furthermore, research shows a majority of AAAS members, working PhD scientists, and active research scientists, polled by Pew Research, favor building more nuclear power plants (Rainie & Funk, 2015). Thus, liberals’ views on nuclear power diverge from the scientific community, whereas conservatives tend to hold more consistent views.

A neutral message was used to assess people’s thoughts of scientists in a nonpoliticized context. To do so, the message used a slightly altered version of Dictionary.com’s definition of “scientists,” stating, “Did you know? A majority of scientists are people who have expert knowledge in the natural or physical sciences.” This message condition served as the reference category in the analyses.

Measures

Political ideology

Using a 7-point scale (1 = very conservative to 7 = very liberal), participants responded to the following statement, “In general, I would describe my political views as . . . ,” M = 4.03, SD = 1.71.

Thought listing

Immediately after reading the message to which they were randomly assigned, participants completed a thought-listing procedure that assessed their thoughts about the scientists depicted in their message. Using the exact instructions outlined by Cacioppo and Petty (1981), participants were told to write down their thoughts about the scientists in their message, with each thought placed in a text box. Ten text boxes were provided, and participants were told they did not have to worry about completing all 10 boxes or be concerned with spelling or grammar. After completing their thought listing, participants were presented with their recorded thoughts and asked to evaluate whether each thought was favorable, unfavorable, or neutral/irrelevant to the scientists. This “subject rating” procedure is described as an acceptable method of judging thoughts (see Cacioppo & Petty, 1981). An index of favorability of scientist-related thoughts was created by subtracting the number of unfavorable thoughts from the number of favorable thoughts and dividing this difference by the total number of favorable and unfavorable thoughts. This method allows one to control for individual differences in total number of thoughts reported, making it an ideal measure of thought favorability (see Brinol, Petty, & Tormala, 2004; Cacioppo & Petty, 1981), M = .45, SD = .73.

Results

The liberal-dissonant message reported a significant main effect but not the conservative-dissonant message (see Table 1, Model 1). These findings indicate that those in the liberal-dissonant message, M = .26, SD = .78, thought of the scientists more unfavorably than those in the neutral message, M = .58, SD = .65, while those in the conservative-dissonant message, M = .51, SD = .73, responded similarly to how participants responded to the neutral message. However, the main analyses examined the role of political ideology as a moderator.

Table 1.

OLS Regression With Scientist Thought Favorability as the Dependent Variable.

	Model 1, B (SE)	Model 2, B (SE)
Constant	.58 (.06)***	.43 (.18)*
Conservative-dissonant message	−.08 (.1)	−.35 (.27)
Liberal-dissonant message	−.33 (.1)**	.16 (.27)
Political ideology		.04 (.04)
Conservative-dissonant * political ideology		.06 (.06)
Liberal-dissonant*political ideology		−.12 (.06)*
R ²	.037	.073
F	(2, 295) = 5.46**	(5, 292) = 4.64***

Note: OLS = ordinary least square; Message conditions were dummy coded with the neutral message as the reference category. Heteroscedasticity-consistent standard errors (HC0) enabled via the RLM macro (Darlington & Hayes, 2017).

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

The results show a nonsignificant interaction effect between the conservative dissonant message and political ideology on scientist thought favorability. Conservative thought listing scores were not significantly different than those in the neutral condition (see Table 2 and Figure 1). However, the results indicate a significant interaction effect between the liberal-dissonant message and political ideology on scientist thought favorability, b = −.12, p = .045. Probing this interaction revealed that while no significant effects existed among conservatives, backfiring effects occurred among liberals (see Table 2 and Figure 1). Specifically, exposure to the liberal-dissonant message resulted in liberals producing thoughts about the scientists that were significantly less favorable than those in the neutral message condition and conservative-dissonant condition.

Table 2.

Effect of the Liberal-Dissonant Message (Neutral Message as Reference) on Scientist Thought Favorability by Political Ideology.

Moderator value	Effect on scientist thought favorability, B (SE)
Very conservative	.04 (.21)
Conservative	−.09 (.16)
Slightly conservative	−.21 (.12)
Moderate	−.33 (.1)***
Slightly liberal	−.46 (.11)***
Liberal	−.58 (.15)***
Very liberal	−.70 (.20)***

Note: Heteroscedasticity-consistent standard errors enabled (HC0). Probed from Model 2 in Table 1 using PROCESS macro (Hayes, 2013).

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

Figure 1.

Plots of estimated means of scientist thought favorability by political ideology.

Study 2

With the liberal-dissonant condition resulting in more negative thoughts among liberals, Study 2 seeks to replicate and extend the findings from Study 1. In extending Study 1’s results, Study 2 explores how biased information processing among liberals might be attenuated when communicating scientist agreement on a dissonant topic. One way to do so is by contextualizing the scientists in terms that resonate with liberals.

First, research has found that liberals tend to support obedience to authorities with perceived ideologically consistent identities. Indeed, liberals are more likely to obey authorities and trust scientists engaged in science that highlights the human impact on the environment and human health (Frimer, Gaucher, & Schaefer, 2014; McCright, Dentzman, Charters, & Dietz, 2013). Second, liberals are more likely than their conservative counterparts to view the environment in moral terms, placing greater emphasis on morals related to harm and care toward the environment (Feinberg & Willer, 2013). That is, liberals place a greater emphasis on morals relating to “impact science” that investigates the negative human impacts on the environment and the necessity of caring and protecting the environment (see McCright, Dentzman, et al., 2013).

For scientists who agree on a liberal-dissonant topic, briefly situating these scientists as “impact scientists” (i.e., scientists engaged in research highlighting the human impact on the environment and human health) whose morals center on care toward the environment might reduce liberals’ tendency to think of them in unfavorable ways (McCright, Dentzman, et al., 2013). Indeed, Feinberg and Willer (2013) present evidence that such an approach can reduce the effects of directional motivated reasoning for politically polarizing science issues, demonstrating that conservatives report greater belief in climate change when exposed to messages emphasizing morals that resonate with them, such as purity and sanctity. However, it has not been investigated if applying this moral framework to liberal-dissonant topics works in a similar manner.

To extend this prior work and the results from Study 1, we explore how liberal-dissonant messaging impacts liberals’ moral judgments about scientists, and whether situating the scientists as holding ideologically consistent morals (i.e., by describing them as impact scientists and emphasizing harm/care morals) attenuates biased information processing.

First, we offer a replication and extension of Study 1. Drawing on Study 1’s findings, it is first hypothesized that exposure to a message emphasizing scientific agreement on a dissonant topic will not only result in liberals producing less favorable thoughts about the scientists but will also influence liberals to judge the scientists as less moral than those exposed to a neutral message. Therefore, a moderation effect is hypothesized:

Hypothesis 1a: The effect of the liberal-dissonant message (vs. neutral message) on scientist thought favorability will be moderated by political ideology.

Hypothesis 1b: A negative effect representing less favorable thoughts (vs. neutral message) will be strongest for liberals exposed to the liberal-dissonant message.

Hypothesis 2a: The effect of the liberal-dissonant message (vs. neutral message) on moral judgments of the scientists will be moderated by political ideology.

Hypothesis 2b: A negative effect representing lower moral judgments of the scientists (vs. neutral message) will be strongest for liberals exposed to the liberal-dissonant message.

Finally, it is hypothesized that describing the scientists as impact scientists will attenuate biases in scientist thought favorability and moral judgments for liberals:

Hypothesis 3: Describing the scientists as impact scientists in the liberal-dissonant message attenuates the biased effects exhibited by liberals.

Method

Sample

A diverse sample of 299 adult American participants was recruited via Mechanical Turk (MTurk).² MTurk is a crowdsourcing site that allows researchers to obtain adult research participants in exchange for payment. The MTurk sample was slightly younger than the SSI sample used in Study 1 and included a greater proportion of males, Mean age = 34.6, SD = 11.68; White = 72.2%; Gender = 44.1% female. This study was conducted online from March 15 to 16, 2017.

Procedure

The study procedure was identical to that of Study 1 with a few alterations. Participants first answered the political ideology and demographic measures and then were randomly assigned to one of three conditions: neutral, liberal-dissonant, and liberal-dissonant with impact science. Both the neutral and liberal-dissonant messages were identical to those used in Study 1. The third message was identical to the liberal-dissonant message but included a statement emphasizing that the scientists cared about protecting the environment and that their support for nuclear power was due to their belief that it can reduce carbon emissions. This manipulation draws on the harm/care morals that liberals value (Feinberg & Willer, 2013), and describes the scientists as impact scientists (McCright, Dentzman, et al., 2013). To ensure greater ecological validity, the manipulation was adapted from an open letter produced by prominent climate scientists who propose nuclear power as a way of reducing harm caused by climate change due to its lower carbon emissions (Patterson, 2013).

After reading their message, participants then performed the thought-listing procedure, evaluated their thoughts, and then were asked to rate the scientists’ morals.

Measures

Political ideology

The political ideology measure involved the same scale used in Study 1, M = 4.60, SD = 1.75.

Thought listing

The thought-listing procedure used the same procedure in study 1, M = .25, SD = .83.

Morals

Using the same measure as Feinberg and Willer (2013), participants were asked with a 6-point scale (1 = not moral at all; 6 = extremely moral), “How moral do you perceive the scientists in your message to be?” M = 4.05, SD = 1.47.

Results

Replication

The results indicate a successful replication of Study 1’s finding on scientist thought favorability. First, the liberal-dissonant message reported a significant main effect, b = −.72, p < .001, indicating that those in the liberal-dissonant message, M = .02, SD = .88, thought of the scientists more unfavorably than those in the neutral message, M = .74, SD = .52 (see Table 3, Model 1).

Table 3.

OLS Regression Models With Scientist Thought Favorability and Scientist Morals as Dependent Variables.

	Thought favorability		Scientist morals
	Model 1, B (SE)	Model 2, B (SE)	Model 1, B (SE)	Model 2, B (SE)
Constant	0.74 (0.05)***	0.34 (.2)	4.75 (0.11)***	3.86 (0.38)***
Liberal-dissonant message	−0.72 (0.1)***	−.04 (0.31)	−1.2 (0.19)***	−0.23 (0.6)
Liberal-dissonant impact science message	−0.69 (0.1)***	−0.51 (0.3)	−0.86 (0.19)***	−0.28 (0.6)
Political ideology		0.09 (0.04)*		0.19 (0.07)**
Liberal-dissonant*political ideology		−0.15 (0.06)*		−.21 (0.11)^†
Liberal-dissonant impact science*political ideology		−0.04 (0.06)		−0.13 (0.11)
R ²	.15	.17	.11	.13
F	(2, 294) 38.94***	(5, 291) 24.35***	(2, 296) 23.5***	(5, 293) 13.62***

Note: OLS = ordinary least square; SE = standard error. Message conditions were dummy coded with the neutral message as the reference category. Heteroscedasticity-consistent standard errors (HC0) enabled via the RLM macro (Darlington & Hayes, 2017).

†

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

We also note a significant interaction effect between the liberal-dissonant message and political ideology on scientist thought favorability, b = −.15, p = .016. Probing this interaction revealed that while no significant effects existed among conservatives, backfiring effects occurred among liberals (see Table 3 and Figure 2). Consistent with Study 1, this result shows that exposure to the liberal-dissonant message resulted in liberals thinking less favorably about the scientists than liberals in the neutral message condition.

Figure 2.

Plots of estimated means of scientist thought favorability by political ideology.

Impact on Moral Judgments

In addition to thoughts about the scientists, we examined how the liberal-dissonant message influences how people evaluate the scientists’ morals. Overall, there was a significant main effect for those in the liberal-dissonant condition, b = −1.2, p < .001, indicating that compared to the neutral message, M = 4.76, SD = 1.1, those in the liberal-dissonant message evaluated the scientists as less moral, M = 3.56, SD = 1.49. However, there is a close to significant interaction effect with political ideology, b = −.21, p = .06, that, when probed, suggests that moral judgments of the scientists become less favorable as political ideology moves further to the political left (see Tables 3 and 4; Figure 3). Therefore, not only does a dissonant message on scientific agreement result in less favorable thoughts about scientists, liberals also view these scientists as less moral.

Table 4.

Effect of the Liberal-Dissonant Message (Neutral Message as Reference) by Political Ideology.

Moderator value	Effect on scientist thought favorability, B (SE)	Effect on scientist morals, B (SE)
Very conservative	−0.18 (0.26)	−0.44 (0.47)
Conservative	−0.33 (0.2)	−0.65 (0.37)
Slightly conservative	−0.48 (0.15)**	−0.86 (0.28)**
Moderate	−0.63 (0.11)***	−1.1 (0.21)***
Slightly liberal	−0.77 (0.1)***	−1.27 (0.18)***
Liberal	−0.92 (0.12)***	−1.48 (0.22)***
Very liberal	−1.1 (0.16)***	−1.69 (0.3)***

Note: Heteroscedasticity-consistent standard errors enabled (HC0). Probed from Model 2 in Table 3 using PROCESS macro (Hayes, 2013).

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

Figure 3.

Plots of estimated means of scientist moral judgments by political ideology.

Impact Science

Will placing the liberal-dissonant message in the context of impact science attenuate biased information processing among liberals? Overall, describing the scientists as impact scientists did not seem to affect liberals in terms of thought favorability and moral judgments (see Table 3). For thought favorability, we report a significant main effect for those in the impact science message condition, b = −.69, p < .001, indicating that compared to the neutral message, M = 0.74, SD = 0.52, those in the impact science message condition had less favorable thoughts, M = 0.04, SD = 0.83. For moral judgments, we report a significant main effect for those in the impact science condition, b = −.86, p < .001, who evaluated the scientists as less moral, M = 3.9, SD = 1.52, than those in the neutral condition, M = 4.76, SD = 1.10. There was no significant interaction with political ideology in terms of moral judgments, b = −.13, p = .24, or scientist thought favorability, b = −.04, p = .50. Thus, irrespective of political ideology, exposure to the impact science message condition resulted in more unfavorable thoughts about the scientists as well as lower moral judgments, when compared with the neutral message condition. However, it is noted that polarization of thought favorability and moral judgments between the neutral and dissonant message did not increase when including the impact science information (see Figures 2 and 3).

Discussion

Findings from this article shed light on how people think about scientists who agree on politically dissonant issues. Importantly, we provide evidence that highlighting scientific agreement might not always neutralize the effect of political worldviews. First, we show that conservatives’ thoughts about scientists agreeing on a conservative-dissonant issue were no more or less favorable than in a neutral context. Though this provides greater credence to the view that statements of scientific agreement can neutralize the effect of political worldviews, it is important to note that conservatives’ thoughts of scientists in the neutral condition were generally less favorable than those of liberals, which corresponds with research showing conservatives as having lower levels of trust in scientists than liberals (Gauchat, 2012). As a result, presenting conservatives with a dissonant science message on climate change might not affect their thoughts of scientists to differ all that much from their views of scientists in a more neutral context.

On the other hand, when scientists were placed in a context that liberals typically do not support, liberals experienced more negative thoughts about the scientists and viewed them as less moral. Consistent with the cultural cognition thesis, these results show that a person’s assessment of a scientific consensus can be processed through a biased lens. Indeed, Study 2 provides further evidence that liberals engage in biased information processing of dissonant messages even when these messages are placed in more ideologically consistent contexts, such as emphasizing scientists’ care for the environment and the role of nuclear power at reducing the harmful effects of climate change. However, it is important to note the lack of significant interaction effects between the impact science condition and political ideology, indicating that polarization of thought favorability and moral judgments between the neutral and dissonant message did not increase when including the impact science information. This finding might allude to the possibility that ideologically consistent scientists are evaluated more favorably even when they support a dissonant issue. In the present study, the experimental manipulation might not have been enough to elicit a perception that the scientists were ideologically consistent. The longer text length could have also stimulated information processing constraints. Indeed, situating the scientists as impact scientists with more morally consistent views of liberals did not significantly affect moral judgment scores. Future research could explore other methods of making scientists as more ideologically consistent, such as message targeting (see Dixon et al., 2017).

Furthermore, the inclusion of impact science not only produced continued resistance among liberals but also appeared to influence conservatives to resist the message. While conservatives did not experience backfiring effects when exposed to the liberal dissonant message on its own, when that message included impact science, they experienced significant negative effects on their thoughts about scientists. This finding corresponds with McCright, Dentzman, et al.’s (2013) article, suggesting that conservatives respond negatively when scientists are identified as being engaged in impact science.

Last, Studies 1 and 2 provide evidence countering the intrinsic thesis, which suggests that liberals are inherently more proscience and less susceptible to motivated reasoning of politically dissonant science messaging (Nisbet et al., 2015). The intrinsic thesis draws from (1) historical analysis of scientific denialist movements that are more typically aligned with political conservativism (Mooney, 2012; Oreskes & Conway, 2010), (2) survey research showing trust in science has been greater among liberals since the mid-1980s (Gauchat, 2012), and (3) evidence that conservatives are psychologically predisposed to resisting dissonant science due to having greater dogmatism and need for closure than liberals (Jost, Glaser, Kruglanski, & Sulloway, 2003; Kruglanski, 2004; Nisbet et al., 2015). Furthermore, studies have found that liberals are more likely to support new scientific technology (i.e., gene editing technology; Weisberg, Badgio, & Chatterjee, 2017), climate change, and vaccinations (Hamilton, Hartter, & Saito, 2015) as compared to conservatives. However, our findings conform to an emerging body of evidence showing that liberals are just as susceptible as conservatives to message resistance, motivated reasoning, and selective exposure when confronted with dissonant science information (Funk, 2017; Kahan et al., 2012; Nisbet et al., 2015; Washburn & Skitka, 2017). Indeed, our studies show that liberals viewed the scientists in the dissonant messages less favorably and as less moral than those exposed to a neutral message.

Limitations

There are some important limitations to note. First, the results of this article are represented by two topics—climate change and nuclear power—which limits the generalizability of its findings to other topics. We also recognize that these topics are not fully equivalent, in that one is dealing with scientific consensus on belief about evidence; the other is a consensus on support for an issue. Furthermore, the level of support for the two topics differs among AAAS scientists, with 87% believing that climate change is human-caused and 65% supporting the use of nuclear power (Rainie & Funk, 2015). However, our main goal was to use issues that were dissonant to conservatives and liberals where there was also strong scientific agreement. In that respect, the topics are equivalent. While there are other scientific topics where liberals and the scientific consensus are not in agreement (i.e., the science on genetically modified [GM] foods), there is an important distinction with that of nuclear power. For example, despite assumptions that liberals are more likely to reject GM science than conservatives, research shows that opposition to GM science is not associated with worldview constructs, such as political ideology (Lewandowsky, Gignac, & Oberauer, 2013), and public opinion polls typically find that liberals and conservatives view GM science in roughly the same manner (Khan, 2013; Mooney, 2014). Nuclear power remains the most visible topic where liberals are generally at odds with the scientific community, while conservatives are not. By the same token, climate change represents one of the most visible issues where conservatives’ views are at odds with the scientific community, while liberals are not. Furthermore, previous studies have also used nuclear power as a liberal-dissonant topic to study topics such as trust in the scientific community (Nisbet et al., 2015). Therefore, though not equivalent in all respects, we believe climate change and nuclear power serve as the best topics for our study.

Second, it is important to recognize that our message stimuli communicated scientific agreement without percentages or visual displays. Instead, we opted to keep the messaging as consistent as possible by using the term majority of scientists rather than spelling out the different percentage of scientific support for nuclear power and climate change. However, we recognize that research shows that communicating scientific agreement via pie charts and percentages is more effective than stating general scientific agreement (Myers, Maibach, Peters, & Leiserowitz, 2015; van der Linden, Leiserowitz, Feinberg, & Maibach, 2014). Future research can then explore how additional message features, such as pie charts or percentages, could aid consensus messaging in reducing biased information processing for politically polarizing scientific issues.

In addition, we recognize our study focuses solely on thoughts related to the scientists, rather than attitudes toward the issue in question or overarching trust in science. Furthermore, we focused only on thought listings because doing so gets at the heart of the question of whether scientists featured in consensus messages are perceived as neutral actors. However, we believe our findings can pave the way for future research exploring how unfavorable thoughts of scientists prompted by a consensus message can affect distrust in scientific institutions or beliefs about established scientific facts. Indeed, previous research finding that consensus messaging to be ineffective or backfire could be due to audiences perceiving the scientists unfavorably. More research in this area is needed.

Last, Study 2 used MTurk to obtain its sample. Although some studies show threats to validity when using MTurk samples for cognitively taxing studies (Krupnikov & Levine, 2014), recent reviews of MTurk do show validity of data to be consistent or even better than other samples typically used by social scientists (Chandler & Shapiro, 2016). Research generally suggests that studies with little “buy-in” are not as threatened by MTurk validity issues (Krupnikov & Levine, 2014). Since Studies 1 and 2 consisted of reading a short message and answering a few questions, it can be argued that the two studies involved little cognitive buy in.

Conclusion

In sum, this article provides additional perspective on the neutralizing impact of messages communicating scientific agreement. Available evidence cautions against assuming scientists are always viewed in neutral terms when they agree on a politically dissonant issue. Liberals, in particular, viewed scientists who agreed on a dissonant issue unfavorably and as less moral. Future research can explore how situating scientists as ideologically consistent might attenuate unfavorable thoughts produced by resistant audiences, which in turn might make consensus messaging most effective.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

Author Biographies

Graham Dixon (PhD, Cornell University) is an assistant professor in the School of Communication at The Ohio State University. His research interests include science and risk communication.

Austin Hubner (BA, Washington State University) is a master’s student in the School of Communication at The Ohio State University. Her research interests involve the intersection of science and politics.

References

Aklin

Urpelainen

(2014). Perceptions of scientific dissent undermine public support for environmental policy. Environmental Science & Policy, 38, 173-177.

Anderegg

W. R.

Prall

J. W.

Harold

Schneider

S. H.

(2010). Expert credibility in climate change. Proceedings of the National Academy of Sciences, 107, 12107-12109. doi:10.1073/pnas.1003187107

Bast

Spencer

(2014, May 26). The myth of the climate change “97%.” The Wall Street Journal. Retrieved from http://www.wsj.com/articles/SB10001424052702303480304579578462813553136

Bolsen

Druckman

J. N.

(2015). Counteracting the politicization of science. Journal of Communication, 65, 745-769.

Bolsen

Druckman

J. N.

(2016). Partisan group identity and belief in human-caused climate change (Working paper). Retrieved from http://www.ipr.northwestern.edu/publications/docs/workingpapers/2016/WP-16-21.pdf

Brinol

Petty

R. E.

Tormala

Z. L.

(2004). Self-validation of cognitive responses to advertisements. Journal of Consumer Research, 30, 559-573. doi:10.1086/380289

Byrne

Hart

P. S.

(2009). The boomerang effect: A synthesis of findings and a preliminary theoretical framework. Annals of the International Communication Association, 33, 3-37. doi:10.1080/23808985.2009.11679083

Cacioppo

J. T.

Petty

R. E.

(1981). Social-psychological procedures for cognitive response assessment: The thought listing technique. In Merluzzi

Glass

Genest

(Eds.), Cognitive assessment (pp. 309-342). New York, NY: Guilford.

Chandler

Shapiro

(2016). Conducting clinical research using crowdsourced convenience samples. Annual Review of Clinical Psychology, 12, 53-81. doi:10.1146/annurev-clinpsy-021815-093623

10.

Cook

(2016). Countering climate science denial and communicating scientific consensus. Oxford Research Encyclopedia on Climate Science. Retrieved from http://climatescience.oxfordre.com/view/10.1093/acrefore/9780190228620.001.0001/acrefore-9780190228620-e-314

11.

Cook

Lewandowsky

(2016). Rational irrationality: Modeling climate change belief polarization using Bayesian networks. Topics in Cognitive Science, 8, 160-179. doi:10.1111/tops.12186

12.

Darlington

R. B.

Hayes

A. F

. (2017). Regression analysis and linear models: Concepts, application, and implementation. New York, NY: Guilford Press.

13.

Deryugina

Shurchkov

(2016). The effect of information provision on public consensus about climate change. PloS One, 11, e0151469.

14.

Ding

Maibach

E. W.

Zhao

Roser-Renouf

Leiserowitz

(2011). Support for climate policy and societal action are linked to perceptions about scientific agreement. Nature Climate Change, 1, 462-466.

15.

Dixon

(2016). Applying the gateway belief model to genetically modified food perceptions: New insights and additional questions. Journal of Communication, 66, 888-908. doi:10.1111/jcom.12260

16.

Dixon

Clarke

(2013). Heightening uncertainty around certain science: Media coverage, false balance, and the autism-vaccine controversy. Science Communication, 35, 358-382. doi:10.1177/1075547012458290

17.

Dixon

Hmielowski

(2017). Improving climate change acceptance among U.S. conservatives through value-based message targeting. Science Communication, 8, 520-534. doi:10.1177/1075547017715473

18.

Douglas

Wildavsky

(1983). Risk and culture: An essay on the selection of technological and environmental dangers. Berkeley: University of California Press.

19.

Feinberg

Willer

(2013). The moral roots of environmental attitudes. Psychological Science, 24, 56-62. doi:10.1177/0956797612449177

20.

Frimer

J. A.

Gaucher

Schaefer

N. K.

(2014). Political conservatives’ affinity for obedience to authority is loyal, not blind. Personality and Social Psychology Bulletin, 40, 1205-1214. doi:10.1177/0146167214538672

21.

Funk

(2017). How much does science knowledge influence people’s views on climate change and energy issues? Washington, DC: Pew Research Center. Retrieved from http://www.pewresearch.org/fact-tank/2017/03/22/how-much-does-science-knowledge-influence-peoples-views-on-climate-change-and-energy-issues/

22.

Funk

Rainie

(2015a). Climate change and energy issues. Washington, DC: Pew Research Center. Retrieved from http://www.pewinternet.org/2015/07/01/chapter-2-climate-change-and-energy-issues/

23.

Funk

Rainie

(2015b). Public and scientists’ views on science and society. Washington, DC: Pew Research Center. Retrieved from http://www.pewinternet.org/2015/01/29/public-and-scientists-views-on-science-and-society/

24.

Gauchat

(2012). Politicization of science in the public sphere: A study of public trust in the United States, 1974 to 2010. American Sociological Review, 77, 167-187. doi:10.1177/0003122412438225

25.

Greenworld. (2016). How low can they go? Hansen, Shellenberger now shilling for Exelon. Retrieved from https://safeenergy.org/2016/04/06/how-low-can-they-go/

26.

Hamilton

L. C.

Hartter

Saito

(2015). Trust in scientists on climate change and vaccines. Sage Open, 5, 1-13.

27.

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

28.

Hayes

A. F.

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

29.

Hornsey

M. J.

Harris

E. A.

Bain

P. G.

Fielding

K. S.

(2016). Meta-analyses of the determinants and outcomes of belief in climate change. Nature Climate Change. doi:10.1038/nclimate2943

30.

Jost

J. T.

Glaser

Kruglanski

A. W.

Sulloway

F. J.

(2003). Political conservatism as motivated social cognition. Psychological Bulletin, 129, 339-375. doi:10.1037/00332909.129.3.339

31.

Kahan

D. M.

(2015). Climate-science communication and the measurement problem. Political Psychology, 36(Suppl. 1), 1-43.

32.

Kahan

D. M.

Jenkins-Smith

Braman

(2011). Cultural cognition of scientific consensus. Journal of Risk Research, 14, 147-174.

33.

Kahan

D. M.

Peters

Wittlin

Slovic

Ouellette

L. L.

Braman

Mandel

(2012). The polarizing impact of science literacy and numeracy on perceived climate change risks. Nature Climate Change, 2, 732-735. doi:10.1038/nclimate1547

34.

Khan

(2013). Do liberals oppose genetically modified organisms more than conservatives? Discover. Retrieved from http://blogs.discovermagazine.com/gnxp/2013/06/do-liberalsoppose-genetically-modified-organisms-more-than-conservatives/#.WNbg-G8rK70

35.

Kruglanski

(2004). The psychology of closed mindedness. New York, NY: Psychology Press.

36.

Krupnikov

Levine

A. S.

(2014). Cross-sample comparisons and external validity. Journal of Experimental Political Science, 1, 59-80. doi:10.1017/xps.2014.7

37.

Kunda

(1990). The case for motivated reasoning. Psychological Bulletin, 108, 480-498.

38.

Lewandowsky

Gignac

G. E.

Oberauer

(2013). The role of conspiracist ideation and worldviews in predicting rejection of science. PloS One, 8, e75637.

39.

Lewandowsky

Gignac

G. E.

Vaughan

(2013). The pivotal role of perceived scientific consensus in acceptance of science. Nature Climate Change, 3, 399-404. doi:10.1038/nclimate1720

40.

Lodge

Taber

(2000). Three steps toward a theory of motivated political reasoning. In Lupia

McCubbins

M. D.

Popkin

S. L.

(Eds.), Elements of reason: Cognition, choice, and the bounds of rationality (pp. 183-213). New York, NY: Cambridge University Press.

41.

McCright

A. M.

Dentzman

Charters

Dietz

(2013). The influence of political ideology on trust in science. Environmental Research Letters, 8, 044029. doi:10.1088/1748-9326/8/4/044029

42.

McCright

A. M.

Dunlap

R. E.

Xiao

(2013). Perceived scientific agreement and support for government action on climate change in the USA. Climatic Change, 119, 511-518.

43.

Mooney

(2012). The Republican brain: The science of why they deny science and reality. Hoboken, NJ: Wiley.

44.

Mooney

(2014). Stop pretending that liberals are just as anti-science as conservatives: No GMOs and vaccines are not the liberal equivalent of climate denial. Retrieved from https://www.motherjones.com/environment/2014/09/left-science-gmo-vaccines/

45.

Myers

T. A.

Maibach

Peters

Leiserowitz

(2015). Simple messages help set the record straight about scientific agreement on human-caused climate change: The results of two experiments. PloS One, 10, e0120985.

46.

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, 36-66. doi:10.1177/0002716214555474

47.

Nyhan

Reifler

(2010). When corrections fail: The persistence of political misperceptions. Political Behavior, 32, 303-330.

48.

Oreskes

Conway

(2010). Merchants of doubt: How a handful of scientists obscured the truth on issues from tobacco smoke to global warming. New York, NY: Bloomsbury Press.

49.

Patterson

(2013, April 14). Climate change warriors: It’s time to go nuclear. CNN. Retrieved from https://edition.cnn.com/2013/11/03/world/nuclear-energy-climate-change-scientists/index.html

50.

Pearce

Grundmann

Hulme

Raman

Hadley Kershaw

Tsouvalis

(2017). Beyond counting climate consensus. Environmental Communication, 11, 723-730.

51.

Rainie

Funk

(2015). An elaboration of AAAS scientists’ views. Washington, DC: Pew Research Center. Retrieved from http://www.pewinternet.org/2015/07/23/an-elaboration-of-aaas-scientists-views/

52.

Slovic

(1987). Perception of risk. Science, 236, 280-285.

53.

van der Linden

S. L.

Clarke

C. E.

Maibach

E. W

. (2015). Highlighting consensus among medical scientists increases public support for vaccines: Evidence from a randomized experiment. BMC Public Health, 15(1), 1207. doi:10.1186/s12889-015-2541-4

54.

van der Linden

S. L.

Leiserowitz

A. A.

Feinberg

G. D.

Maibach

E. W

. (2014). How to communicate the scientific consensus on climate change: Plain facts, pie charts or metaphors? Climatic Change, 126, 255-262.

55.

van der Linden

S. L.

Leiserowitz

A. A.

Feinberg

G. D.

Maibach

E. W

. (2015). The scientific consensus on climate change as a gateway belief: Experimental evidence. PloS One, 10, e0118489.

56.

van der Linden

Leiserowitz

Maibach

. (2018). Scientific agreement can neutralize politicization of facts. Nature Human Behaviour, 2(1), 2-3.

57.

Washburn

Skitka

(2017). Science denial across the political divide: Liberals and conservatives are similarly motivated to deny attitude-inconsistent science. Social Psychological and Personality Science. Advance online publication. doi:10.1177/1948550617731500

58.

Weisberg

S. M.

Badgio

Chatterjee

(2017). A CRISPR new world: Attitudes in the public towards innovations in genetic modification. Frontiers in Public Health, 5, 117-126.

59.

Zhou

(2016). Boomerangs versus Javelins: How polarization constrains communication on climate change. Environmental Politics, 25, 788-811. doi:10.1080/09644016.2016.1166602