Accepted Manuscript Perception of public opinion on global warming and the role of opinion deviance Adina T. Abeles, Lauren C. Howe, Jon A. Krosnick, Bo MacInnis PII: S0272-4944(18)30545-0 DOI: https://doi.org/10.1016/j.jenvp.2019.04.001 Reference: YJEVP 1293 To appear in: Journal of Environmental Psychology Received Date: 31 July 2018 Revised Date: 4 April 2019 Accepted Date: 4 April 2019 Please cite this article as: Abeles, A.T., Howe, L.C., Krosnick, J.A., MacInnis, B., Perception of public opinion on global warming and the role of opinion deviance, Journal of Environmental Psychology (2019), doi: https://doi.org/10.1016/j.jenvp.2019.04.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Title: Perception of public opinion on global warming and the role of opinion deviance Author names and affiliations: Adina T. Abeles (corresponding author)a , Lauren C. Howeb, Jon A. Krosnickc, Bo MacInnisd a Department of Communication T Building 120, Room 110 P 450 Serra Mall Stanford University I Stanford, CA 94305 R e-mail: [email protected] phone: 831-402-5132 C fax: 650-725-2472 S b Department of Business Administration University of Zurich U e-mail: [email protected] N c Departments of Communication, Political Science, and Psychology A Stanford University e-mail: [email protected] M d Department of Communication Stanford University D e-mail: [email protected] E Declaration of interest: none T Full acknowledgements (note: acknowledgements are also included in the manuscript, but the P acknowledgements in the manuscript remove info that indicates authorship ): The authors thank E Lee Ross and Jonathan Schuldt for feedback on earlier drafts. The authors thank the Stanford C Woods Institute for the Environment, the Stanford Precourt Institute for Energy, and Resources C for the Future for financing the 2018 data collection, and the National Science Foundation A (award number: 1042938) for financing the 2012 data collection. Adina Abeles is a Goldman Graduate Fellow at the Stanford Woods Institute for the Environment. Jon Krosnick is University Fellow with Resources for the Future. ACCEPTED MANUSCRIPT Abstract People tend to perceive their own opinions to be especially prevalent among others, especially among members of their in-groups. This paper proposes that this process operates differently in T groups with prominent figures who take public stances on issues (opinion cues) that suggest what P group members do or should think when some members of such groups are opinion deviant, I R meaning that their own opinions differ from their group’s opinion cues. Two studies of C Democrats’, Republicans’ and Independents’ perceptions of others’ views on global warming in S two national surveys of probability samples of American adults (N = 804 in 2012, and N = 1000 U in 2018) showed that whereas partisans whose own beliefs aligned with the opinion cues of their N political party perceived their own beliefs to be more prevalent among members of their own A party and perceived the opposite belief to be more prevalent among members of the other party, M opinion deviants (e.g., Republicans who believe that global warming has been happening) did D not. Furthermore, aligned partisans perceived a larger partisan gap than opinion deviants. This research illuminates a new way in whichE group membership affects perceptions of public opinion T on environmental issues and suggests novel strategies for climate change communication. P Key words: Public perceptiEon, public opinion, global warming, climate change, false consensus, polarization C C A 1 ACCEPTED MANUSCRIPT 1. Introduction Much research on global warming public opinion has focused on the distributions of people’s beliefs about the existence and threat of global warming, the predictors and behavioral T consequences of these beliefs, and how to engage people on the issue (e.g., Hornsey, Harris, P Bain, & Fielding, 2016; Moser, 2016). Other research has explored people’s policy preferences I R on the issue and illuminated how personal experiences and information presentation shape these C preferences (e.g., McAdam, 2017, p.192-193). To date, this work has largely ignored the S importance of perceptions of others’ beliefs about global warming as drivers of individual U attitudes, preferences, and behaviors. N Social psychologists have long recognized the influential role that others’ beliefs and A behaviors (i.e., social norms) play in shaping an individual’s judgements and behaviors (e.g., M Asch, 1951; Cialdini & Goldstein, 2004; Jacobs & Campbell, 1961; Sherif, 1936). And more D recently, researchers have documented how these norms potentiate environmental behaviors (Cialdini & Goldstein, 2004; Goldstein, ECialdini, & Griskevicius, 2008; Miller & Prentice, 2016; T Schultz, Nolan, Cialdini, Goldstein, & Griskevicius, 2007; Sparkman & Walton, 2017). But only P a small number of recent studies have begun to recognize the importance of perceptions of E others’ beliefs about global warming. These studies have illuminated instances in which C perceptions of public opinion on global warming are inaccurate. A study of a probability sample C of Australian adults found that a majority of Australians misperceived public opinion on global A warming, underestimating the number of people who think global warming is human-caused (Leviston, Walker, & Morwinski, 2012). Two other studies of non-probability samples of American adults indicated that people underestimate the percent of people who believe that 2 ACCEPTED MANUSCRIPT global warming (Mildenberger & Tingley, 2017) or climate change (Ehret, Boven, & Sherman, 2018) is happening. When misperceptions such as these occur, people may act or behave in ways that are T counter to their privately held beliefs (Krueger, 1998), such as publicly supporting policies with P which they do not agree (Todorov & Mandisodza, 2004). Supporting this idea, Geiger and Swim I R (2016) found that misperceptions of others’ global warming beliefs has consequences. C Undergraduates who incorrectly assumed that most other students do not think global warming is S happening were less willing to discuss global warming with their peers. U If people tend to underestimate the extent to which others believe that global warming is N happening, that may explain a lack of support for climate change policies, and/or a depressed A willingness to talk about the issue with others (Doherty & Webler, 2016; Leiserowitz, Maibach, M Roser-Renouf, Feinberg, & Rosenthal, 2015). For example, Mildenberger and Tingley (2017) D found that respondents who perceived fewer Chinese citizens to think global warming was occurring were less likely to support a UE.S. bilateral climate agreement with China. Thus, T intervening to correct misperceptions may be a fruitful way to galvanize support for action on P climate change (Miller & Prentice, 2016; Prentice & Miller, 1996). To correct misperceptions, E however, it is important to understand the underlying factors that influence perceptions of others’ C opinions in the first place (Walton, 2014). Yet research on perceptions of climate change public C opinion has yet to explore why misperceptions exist. A Previous research on attitude perception generally suggests that people could rely on a variety of data points to estimate others’ beliefs. One well-studied data point is a person’s own beliefs and actions (Dawes, 1989, 1990); people are thought to assimilate others’ opinions toward their own (Ross, Greene, & House, 1977). Group membership is thought to moderate this 3 ACCEPTED MANUSCRIPT process; the more similar a person is to a group, the more that person assimilates group members’ opinions toward their own (Robbins & Krueger, 2005). West and Kenny (2011) described this assumed similarity between the perceiver and target as a form of bias influencing T individuals’ judgement of others’ beliefs. More recent research indicates that dissimilarity is P also an influential force shaping perceptions of others’ beliefs, showing that individuals contrast I R disliked outgroup members’ opinions away from their own (Gromet & Van Boven, 2014). C In this paper, two shortcomings of these previous literatures are addressed. First, S Americans’ perceptions of others’ beliefs about global warming are measured in two national U probability samples of American adults. Probability samples provide more accurate N measurements of public opinion in populations than do non-probability samples (e.g., MacInnis, A Krosnick, Ho, & Cho, 2018; Yeager et al., 2011) and therefore offer sound ground on which to M reach conclusions about accuracy of perceptions of public opinion in the public. D Second, this paper explores additional factors that influence perceptions of public opinion on global warming. Specifically, we intrEoduce a new perspective on how group membership T might influence individual’s judgements of group members. Some groups, such as political P parties, have prominent figures who publicly espouse viewpoints. These figures can shape E individuals’ perceptions of what group members think. When an individual’s belief conflicts C with the beliefs espoused by the group’s prominent figures, these individuals are opinion deviant. C This paper explores perceptions of others’ beliefs held by opinion deviants, testing whether such A people rely on their own belief, the cues from prominent figures, or both, when forming perceptions of others’ beliefs about global warming in the United States. The paper makes novel contributions by discussing perceptions of public opinion on global warming, how opinion 4 ACCEPTED MANUSCRIPT deviance impacts these perceptions, how these findings advance our understanding of perceptions of group members, and their implications for climate change communication. T 2. A new theory of group membership and perceptions of public opinion P 2.1. Background: Group membership and perception I R A great deal of research has found a strong positive correlation between a person’s own C belief and perceptions of others’ beliefs on the same issue (e.g., Marks & Miller, 1987; Ross, S Greene, & House, 1977). However, this early research did not take the role of group membership U into account. More recent research began to investigate how ingroup and outgroup membership N moderates this relationship. Researchers repeatedly demonstrated that a person’s beliefs A correlate more strongly with their perceptions of in-group members’ beliefs than with M perceptions of the beliefs of members of other groups, a relationship often referred to as in-group D projection or assumed similarity in the literature (Ames, 2004; Billig & Tajfel, 1973; Clement & Krueger, 2002; Robbins & Krueger, 200E5; Tajfel, Billig, Bundy, & Flament, 1971; West & T Kenny, 2011). In contrast, evidence is mixed regarding whether people perceive the opposite P belief to be more prevalent among out-group members than in-group members (i.e., assumed E out-group dissimilarity). One meta-analysis did not find evidence of assumed out-group C dissimilarity (Robbins & Krueger, 2005). C However, past research has largely focused on social groups that were not formed around A socially or politically contentious issues (e.g., academic affiliation, nationality, ethnicity), and Robbins and Krueger (2005) suggested that particular circumstances may be necessary to elicit these effects. Political partisans (i.e., Americans who identify as Democrats or Republicans) may have the characteristics necessary to solicit strong in-group assumed similarity and out-group 5 ACCEPTED MANUSCRIPT assumed dissimilarity. In-group and out-group animus among political partisans in the United States has increased since the 1960’s (Alford, Hatemi, Hibbing, Martin, & Eaves, 2011; Iyengar & Krupenkin, 2018; Iyengar, Sood, & Lelkes, 2012), and bias based on political party has T outpaced biases based on race, religion, and gender (Iyengar & Westwood, 2015). The desire to P differentiate oneself from the out-group might lead people to believe that out-groups endorse the I R opposing view compared to in-groups. Offering initial evidence for this view, Gromet and Van C Boven (2014) observed assumed out-group dissimilarity (what they called oppositional S projection) in the context of Democrats and Republicans’ perceptions of beliefs about cap trade U policies to address global warming. N 2.2. Groups with opinion cues A Exploration of these issues is enriched by the recognition that some groups have M spokespeople who send signals to others about how group members should or do think. When D groups have prominent public figures, these figures’ statements provide additional data points that may influence perceptions of groupE members’ beliefs. Prominent figures can be officially or T unofficially associated with a group and may publicly pronounce a purported majority group P opinion. For example, officials speaking for political parties include politicians and E representatives from national, state, and local political party committees, and these people often C make public statements endorsing a certain stance on an issue. Unofficial figures include people C who serve as exemplars, such as people whom activists or media pundits identify and highlight A as demonstrating how excellent or typical group members behave. Thus, for some groups, another data point that might influence perception of group members’ beliefs are opinion cues: the publicly espoused opinions of prominent figures on behalf of a group. 6 ACCEPTED MANUSCRIPT Opinion cues are related to, but different from, elite or party cues as discussed in the political science literature. Elite cues have been described as a shortcut to help individuals form their own opinions on complex topics (Popkin, 1994; Sniderman, Brody, & Tetlock, 1991; T Zaller, 1992). Public statements by elected officials, activist organizations, interest groups, and P political parties can be used by citizens as an information shortcut to decide how they should I R think about a specific issue, particularly when taking the time to understand the issue is not C worth the effort or not possible (Kuklinski & Quirk, 2000). In addition, political commentators S on television and talk radio can similarly signal listeners what conservative or liberal individuals U should or do think about an issue (Feldman, Maibach, & Leiserowitz, 2012). While party cues N provide a heuristic for individuals when forming an individual opinion, opinion cues, we posit A here, might provide a heuristic for individuals when considering what group members think. M Some groups, such as “college students,” “parents,” or “unions” are well-defined and can D be important to a person’s social identity, but these types of groups do not always send clear signals about what group members shouEld or do think. For example, some groups may not have T associated prominent figures, leaders, or institutions that make public statements about issues P (e.g., there is no one spokesperson for all parents). Or groups may have prominent figures, but E because these groups are formed for reasons other than shared opinions, people may not assume C that these leaders represent group members’ opinions or beliefs. C However, if groups are formed based on shared values or beliefs, such as activist groups A or political parties, people may be more likely to believe that prominent figures’ opinions reflect the opinions shared by a majority of group members. People may have different lay theories about why there is a relationship between prominent figures and group members’ opinions. First, prominent figures might represent the group’s prior opinions and have been elevated to a 7 ACCEPTED MANUSCRIPT position of power on that basis. Or, people may think that leaders persuade group members to adopt an opinion and thus come to represent the general opinions of the group. Regardless of the theory as to why the relationship exists, opinion cues from these groups likely act as a heuristic T upon which people base perceptions of group members’ opinions. This assumption makes sense, P since people are, in reality, particularly responsive to party endorsements when forming their I R own opinions (Cohen, 2003; Slothuus & De Vreese, 2010). C Critically, however, for groups in which opinion cues are present, individuals may S sometimes exhibit opinion deviance, when they hold an opinion in direct conflict with opinion U cues of their group. For instance, some Democrats who are pro-life, some members of the Green N Party support the use of nuclear energy, and some Republicans are concerned about climate A change. Although opinion deviance characterizes many groups in the public discourse, its M influence on perceptions of in-groups and out-groups is not yet known. D 2.3. Opinion deviance as a moderator Opinion deviance raises a puzzleE for the study of perceptions. When a person is not an T opinion deviant, there is no reason to doubt that people will assume more similarity between P their own opinion and in-group members’ opinions (reinforced by the opinion cues) and to E assume more dissimilarity from the opinions of outgroup members (reinforced by the opinion C cues of the outgroup). Previous research suggests that opinion deviants’ beliefs may be positively C correlated with perceptions of in-group opinions, despite any conflicting opinion cues A However, our theory proposes that opinion cues may alter perceptions of in-groups and out-groups. Perceivers may use the data points of opinion cues and their own beliefs additively or interactively. In an additive model, a perceiver weighs the data points of the self and the prominent figures and sums the two weighted data points to result in a perception of group 8