Doing What Others Do

Abstract

Does rhetoric highlighting social norms or mentioning science in a communication affect individuals’ beliefs about global warming and/or willingness to take action? We draw from framing theory and collective-interest models of action to motivate hypotheses that are tested in two large web-based survey-experiments using convenience samples. Our results show that attitudes about global warming, support for policies that would reduce carbon emissions, and behavioral intentions to take voluntary action are strongly affected by norm- and science-based interventions. This has implications for information campaigns targeting voluntary efforts to promote lifestyle changes that would reduce greenhouse gas emissions.

Keywords

collective action global warming climate change social norms

Mounting scientific evidence suggests that global warming poses significant threats to humans and the wider environment. Nonetheless, opinion polls reveal a growing skepticism among the U.S. public about whether global warming is actually occurring, and if it is, whether human activities are the main cause of the observed warming trend over the past 150 years.¹ With skepticism on the rise, questions remain about what—if anything—can be done to influence public attitudes and behaviors toward global warming. Because climate represents an important common pool resource, understanding (and indeed, shaping) individuals’ attitudes and behavioral intentions regarding global warming is critical.

What factors shape citizens’ attitudes toward action on global warming? We theorize that norm-based communications can influence willingness to take action for the public good—for example, encouraging voluntary reductions in greenhouse gas emissions and support for public policies targeting the nation’s emissions. We focus on the issue of global warming because it presents a severe collective action problem: individuals alone cannot control the climate, and the recommended behaviors may have high costs and the benefits are nonexcludable (Lubell, Zahran, & Vedlitz, 2007). This article has three interrelated contributions. First, it explores the role that communications that invoke social norms play in shaping individuals’ attitudes and willingness to take action on global warming.² Second, we assess how global warming communications affect individuals’ attitudes and willingness to act. These outcomes matter both to the extent that individual attitudes and behavioral intentions with regard to global warming matter from a policy perspective and because they reflect public beliefs about science and scientific evidence, which matter in their own right (Bauer et al., 2007; Pielke, 2007; Sarewitz, 2000, 2004). Third, given the partisan divide in America on the issue of global warming, we explore the potential linkages between partisanship and these communications. That is to say, we do not examine the divide per se but instead consider under what potential conditions that divide might be mitigated by norm- and/or science-based communications.

In pursuit of the goals outlined above, we present two large survey experiments. Experiment 1 tests how norms shape attitudes and behavioral intentions toward global warming. We find that a norm-based communication that discourages action, for example, by providing information that others are not cooperating, significantly reduces perceptions of efficacy and willingness to take action. Experiment 2 further tests the effects of norm-based communications in tandem with messages about the scientific consensus surrounding global warming. We begin by discussing the connections between collective action, norms, and global warming, then discuss the design of the experiments, and discuss the results thereof.

Norms and Collective Action on Global Warming

We focus on the determinants of behavior taken to secure a public good: any good that cannot feasibly be withheld from others in a group if it is provided for any member of that group (Olson, 1965). Individuals will tend to cooperate on this collective action problem only when they believe that others are also likely to cooperate (Alpízar et al., 2008; Axelrod, 1984; Frey & Meier, 2004; Kogut & Beyth-Marom, 2008; Shang & Croson, 2009). In a number of social comparison treatments in large-scale field experiments which examine how information about the behavior of others affects energy and water conservation,³ it has been found that giving consumers feedback on their consumption, providing information on energy savings opportunities, comparing their use to their neighbors’ use, facilitating public or private goal setting, and structuring commitment devices causes households to reduce consumption between 5% and 20% (Allcott, 2010). Thus, immediate action is taken by citizens when they are convinced that others will commit; however, we know little about how responsive individuals are to behavior change communications (Allcott & Mullainathan, 2010; McKenzie-Mohr, 2000) and, in particular, how individuals respond to behavior change communications concerning climate change.

Research on the factors that promote private action for the public good indicates that social norms often play a central role, for example, providing cues about the efficacy of a collective action (Bolsen, 2013; Cialdini, 2003; Nolan, Schultz, Cialdini, Goldstein, & Griskevicius, 2008). Here, “norms” refer to how most people behave in a given social context, but they are not uniform. Social psychologists refer to information describing how most people act in a decision context as a descriptive norm, while injunctive norms refer to information about how people ought to behave in a situation, regardless of how people are actually behaving (Cialdini, 2003). Norms can be especially strong in situations in which an individual’s action causes negative effects on the lives of others and thus evolve in communities as a way to regulate social life (Biel & Thogersen, 2006; Thogersen, 2008). In these situations, norms serve to restrain egoistic impulses and induce cooperation among group members.⁴

Our research is novel in that it is one of the first attempts to shape beliefs and intentions to take action on global warming in an experimental context. To understand how exposure to communications invoking social norms, or highlighting the science related to global warming, affects opinions and willingness to take action, we develop a framework that links framing theory (Chong & Druckman, 2007) with expectancy-value models of collective action (Finkel, Muller, & Opp, 1989; Lubell, 2002; Lubell et al., 2007). First, in deciding whether or not to take a collective action, individuals evaluate the costs and benefits likely associated with the outcome from taking action (Ajzen & Fishbein, 2005). In doing so, individuals form an attitude toward a specific behavior (i.e., an evaluation toward the target) that is a function of accessible considerations. For instance, in deciding whether to voluntarily reduce one’s own personal travel as a way to combat global warming, an individual may consider the positive effects such actions might have for the collective environment or the personal costs associated with such sacrifices.

Framing theory also explains that one’s overall attitude toward any object is a function of the salience and weight attached to various considerations toward the object. Thus, exposure to rhetoric—that is, persuasive communications targeting attitude change—can affect one’s attitude toward a behavior either through altering the salience associated with a particular consideration (i.e., a framing effect, see Druckman (2001, 2004) or through persuasion via attitude change (O’Keefe, 2002). Our research goals necessitate the wedding of framing theory to collective-interest models of political and environmental action, not unlike Finkel et al. (1989), Lubell (2002), and Lubell et al. (2007). This literature explains that individuals consider not only the personal costs and benefits associated with taking a collective action, but also collective interest calculations such as the likelihood one’s own actions will influence collective outcomes as well as the likelihood of the group achieving success. Communications that highlight social norms or scientific evidence may directly influence one’s attitude toward and willingness to take collective action on the environment by altering the cost-benefit calculus at the individual level.

Based on this framework, we make the following predictions:

Hypothesis 1a (H1a): A norm-based communication promoting action to reduce the nation’s greenhouse gas emissions will increase individuals’ beliefs that taking action toward global warming is efficacious (i.e., makes a difference).

Hypothesis 1b (H1b): A norm-based communication discouraging action to reduce the nation’s greenhouse gas emissions will decrease individuals’ beliefs that taking action toward global warming is efficacious (i.e., makes a difference).

Again, we argue that norms provide a signal about the willingness of others to cooperate in securing a public good. When others are perceived as taking action, individuals will be more likely to take action for the public good because their behavior is perceived as efficacious—that is, as making a difference in terms of the collective outcome; when others are perceived as not cooperating, individuals will be less likely to take costly action for a public good.

Hypothesis 2a (H2a): A norm-based communication promoting action to reduce the nation’s greenhouse gas emissions will increase individuals’ willingness to take action for the public good (e.g., drive a smaller vehicle and support a carbon tax).

Hypothesis 2b (H2b): A norm-based communication discouraging action to reduce the nation’s greenhouse gas emissions will decrease individuals’ willingness to take action for the public good (e.g., drive a smaller vehicle and support a carbon tax).

In addition to norms, information about a scientific consensus regarding the existence of global warming should also directly affect the public’s attitudes and behavior. First, communications highlighting the views of a credible group—scientists conducting research on the phenomena of global warming, for instance—may increase the perceived strength of communications promoting action (Lupia & McCubbins, 1998; Zaller, 1992). It may also influence individuals’ attitudes about the existence and anthropogenic causes of global warming, which according to our cost-benefit model of collective action should increase the likelihood of motivating individual action. Thus, we make the following predictions:

Hypothesis 3 (H3): Invoking beliefs among scientists about the existence and causes of global warming will increase individuals’ beliefs that global warming is occurring and anthropogenic.

Hypothesis 4 (H4): Invoking beliefs among scientists about the existence and causes of global warming will increase individuals’ willingness to take action for the public good (e.g., drive a smaller vehicle and support a carbon tax).

Finally, there is a clear partisan divide in America with regard to the anthropogenic nature and potentially harmful effects of global warming (Dunlap & McCright, 2008; Klick & Smith, 2010; Krosnick, Holbrook, & Visser, 2000; Villar & Krosnick, 2010), and we therefore expect party identification will play a significant moderating role on the effects of social norms and science-related communications about global warming. The lack of public engagement on this issue may be, in part, attributable to ineffective frames in the debate over global warming (Malka, Krosnick, & Langer, 2009; Nisbet, 2009; Nisbet & Scheufele, 2009). Alternatively, it could simply be a function of one’s party affiliation, which has been found to have significant (Nisbet & Goidel, 2007) and insignificant (Evans, 2011) effects on the receipt of scientific information. We approach this at an exploratory level to the extent that we are looking not only at the effects of social norms on beliefs and willingness to act on global warming, but also at the interaction of such norms with scientific communications. To our knowledge, few examinations of potential linkages between partisanship and the receipt of these kinds of communications have been conducted. An exception is Hart and Nisbet (2011) who find that Republicans and Democrats respond differently to communications targeting action on global warming. Another exception is a recent study which finds that Republicans households in California increased their energy consumption in response to a norm-based social comparison intervention designed to promote conservation; the intervention significantly reduced consumption among Democrats (Costa & Kahn, 2012).

Exploratory Hypothesis: Republicans are more likely than Democrats to discount (or react negatively to) social norms and science-based communications promoting action on global warming.

Data/Method

We test our predictions in two large survey experiments involving 622 and 390 undergraduate students, respectively, at Georgia State University in December, 2010, and December, 2011. Participants for the two studies were recruited from a first-year political science course, which is required for all students at the university.⁵ Participants who chose to participate were compensated with extra course credit for their participation.⁶

Experiment 1, which tests Hypotheses 1 and 2, contained two manipulations: (a) the direction of a norm toward action on global warming—that is, Americans were described as either willing or unwilling to take specific actions toward global warming (e.g., drive smaller vehicles and support a tax on carbon); and, (b) the content of the communication itself—that is, either a descriptive norm communication alone or a combination of descriptive and injunctive communication components.⁷ Table 1 outlines the designs for each experiment, treatment group sample sizes, and reiterates our expectations regarding treatment effects relative to a baseline (control group) not exposed to any treatment. Students were randomly assigned to one of the treatment groups or the control group. Those in the pro norm conditions read the following paragraph of text:

A recent poll showed that over 85% of Americans believe that the world’s average temperature is rising primarily because of human activities. In addition, the vast majority of respondents said they would consider driving smaller cars, reducing travel, and supporting legislation (e.g., a tax) to reduce the nation’s emission of greenhouse gases.

Table 1.

Experimental Design and Predictions.

Experiment 1
	Pro norm	Con norm
Descriptive	Increase action relative to baseline (n = 117)	Decrease relative to baseline (n = 135)
Descriptive + injunctive	Largest increase relative to baseline (n = 153)	Largest decrease relative to baseline (n = 140)
Control group (no norm)	Baseline (n = 77)

Experiment 2
	No social norm	Social norm
No information	Baseline (n = 97)	Increase relative to baseline (n = 92)
Scientific consensus communication	Increase relative to baseline (n = 109)	Largest increase relative to baseline (n = 92)

Those in the injunctive norm condition additionally read: “Respondents said the most important reason for taking these actions is because it is “the right thing to do for all of us.”

Those in the con norm conditions similarly read that:

A recent poll showed that less than 15% of Americans believe that the world’s average temperature is rising primarily because of human activities. In addition, only a small minority of respondents said they would consider driving smaller cars, reducing travel, and supporting legislation (e.g., a tax) to reduce the nation’s emission of greenhouse gases.

Those in the con injunctive condition additionally read: “Respondents said the main reason they are unwilling to take these actions is because reducing our standards of living is the “wrong thing to do.”

Experiment 2, which further tested Hypotheses 3-5, also contained two manipulations: (a) the inclusion or exclusion of a pro social norm toward action on global warming similar to that mentioned above; and, (b) the inclusion or exclusion of the science-based information about global warming indicating that scientists have achieved a consensus that global warming is occurring (a “science” communication). Those in the social norm condition read the following paragraph of text:

A 2010 report by the Pew Research Group found that a solid majority of Americans (59%) believe the Earth is experiencing a long-term warming trend because of humans’ activities. The poll also showed that the vast majority of Americans are willing to drive a smaller car and support legislation that taxes polluters of greenhouse gases.

Those in the science-based communication group read the following:

A 2010 report by the Pew Research Group found that the vast majority of scientists (over 90%) believe the Earth is experiencing a long-term warming trend. Indeed, there is broad scientific consensus about this issue and nearly all research conclusively finds that the planet will continue to warm over time.

Finally, those in the social norm and science condition read both paragraphs.⁸ Those in the no-norm/no-science condition simply answered the outcome questions. As was the case for Experiment 1, students were randomly assigned to one of the treatment groups or the control group.

Note that neither experiment emphasizes the source of information. Instead, we present nonpartisan polling information about what Americans, scientists, or both feel about global warming’s anthropogenic nature and potential effects. We acknowledge that global warming is so politically charged (in the United States) that it may not make a difference as to the source and/or content (Mutz, 2008), but we expect that this method reduces any potential confounding effects.

In both experiments, we included two key dependent measures of individuals’ willingness to take action. The first measure taps support for a carbon emissions cap, “even if it increases costs to consumers” (1-7 scale, where 1 = strongly oppose, 7 = strongly support) and the second measure explores behavioral intentions to change one’s habits (e.g., driving habits) as a way to reduce carbon emissions (1-7 scale, where 1 = extremely unlikely, 7 = extremely likely).

For Experiment 1 only, we measured perceptions about the efficacy of action with two measures. The first asked participants to indicate the extent to which they agreed with a statement that taking personal action has an impact on the nation’s carbon emission (1-7 scale, where 1 = strongly disagree, and 7 = strongly agree).⁹ The second item measures expected behavioral reciprocity from others by asking participants to respond to a statement about whether their actions “encourage others in my community” to take actions to reduce carbon emissions (1-7 scale, where 1 = strongly disagree, and 7 = strongly agree).

Both experiments also included two belief items (a) belief about whether global warming is happening (where 1 = definitely is not happening, and 7 = definitely is happening) and (b), if it is happening, whether the trend is a result of human activities or natural changes (1-7 scale, where 1 = definitely naturally induced, and 7 = definitely human induced).

All variables were measured on seven-point scales, exact wordings for which can be found in the Appendix. For the analysis, all variables are rescaled from −1 to +1 to ease interpretability. In presenting the results, we report treatment effects as linear regression coefficients controlling for partisanship (as well the interactions between partisanship and the treatment indicators). We report linear regression results, following the advice of Angrist and Pischke (2009), and rely on bootstrapped (n = 2,000) standard errors to avoid imposing parametric assumptions on our inference. The reported results are robust to alternative functional forms, including ordered probit regression. For both experiments, we include party identification as a control.¹⁰

Results

Our hypotheses anticipated that beliefs and behaviors toward global warming would be influenced by norm- and science-based communications. The stickiness of global warming beliefs (and to a lesser extent behavioral intentions) is, however, reflected in our relatively mixed findings. We begin with results from Experiment 1, which tests our norm-based hypotheses and explores the influence of partisanship, before turning to Experiment 2, which further tests those hypotheses as well as our predictions regarding science-based communications.

Although we expected norms promoting action to influence both beliefs (Hypothesis 1a) and behavioral intentions (Hypothesis 2a), we find little support our hypotheses. Table 2 reports regression results with indicators for each treatment condition (relative to a control condition), including a control for party identification. Looking at the rows for the Pro Norm conditions, we do not see any effect on any outcome. Indeed for several outcomes (perceived personal influence and support for a carbon emissions tax), we find effects in the opposite direction of our expectations at a nonsignificant level. Norm-based communications that stress others’ actions to affect global warming appear to have little impact on individuals’ attitudes or behavioral intentions.¹¹

Table 2.

Treatment Effects (Experiment 1).

	Perceived personal influence (1)	Expected reciprocity (2)	Believe global warming is happening (3)	Believe global warming is human induced (4)	Support cap on carbon emissions (5)	Willingness to take personal action (6)
Intercept	0.38** (0.06)	0.23** (0.06)	0.37** (0.07)	0.25** (0.07)	0.28** (0.07)	0.20** (0.07)
Pro descriptive (D)	−0.04 (0.08)	0.02 (0.08)	0.05 (0.09)	−0.03 (0.09)	−0.06 (0.09)	−0.02 (0.09)
Prodescriptive + injunctive (D+I)	−0.08 (0.08)	0.07 (0.08)	0.09 (0.09)	0.06 (0.09)	−0.02 (0.09)	−0.05 (0.08)
Con D	−0.18** (0.08)	−0.06 (0.08)	−0.06 (0.09)	−0.05 (0.09)	−0.16** (0.09)	−0.17** (0.09)
Con D+I	−0.10 (0.08)	0.11* (0.07)	0.03 (0.09)	−0.06 (0.09)	−0.08 (0.09)	−0.06 (0.09)
PartyID	0.21** (0.08)	0.21** (0.08)	0.28** (0.11)	0.31** (0.09)	0.21** (0.10)	0.13* (0.09)
PartyID × pro D	−0.05 (0.11)	−0.04 (0.11)	−0.10 (0.14)	−0.04 (0.13)	−0.05 (0.13)	0.04 (0.13)
PartyID × pro D+I	−0.10 (0.11)	−0.12 (0.11)	−0.06 (0.13)	−0.13 (0.12)	−0.14 (0.13)	0.04 (0.11)
PartyID × con D	−0.10 (0.11)	−0.19** (0.11)	0.00 (0.14)	−0.10 (0.13)	0.03 (0.13)	0.11 (0.13)
PartyID × con D+I	−0.13 (0.11)	−0.21** (0.10)	0.01 (0.13)	−0.15 (0.12)	0.01 (0.13)	0.04 (0.11)
SER	0.48	0.47	0.52	0.53	0.51	0.54
n	612	612	612	611	612	612

Note: D = descriptive, I = injunctive.

Cell entries are unstandardized linear regression coefficients with bootstrapped standard errors in parentheses. Baseline is the control condition. Party identification is coded with seven categories from Republican (−1) to Independent (0) to Democrat (+1). SER is the standard error of the regression (Beck, 2010).

p < .10. **p < .05. one-tailed test.

In contrast, when communications invoke con norms, in which other Americans are characterized as unwilling to take action on global warming, there are some quite interesting patterns. The descriptive con norm treatments (as seen in Table 2) significantly reduced perceived personal influence on global warming (p < .05), consistent with our Hypothesis 1b. The descriptive con norm treatment also significantly reduced support for a carbon emissions tax (p < .05) and willingness to take personal action (p < .05), consistent with Hypothesis 2b. Effects were also in the expected direction (though not statistically significant) for some of the other outcomes. We had no clear expectations about the different effects of descriptive versus injunctive norms and the results here suggest that descriptive rather than injunctive norms are more efficacious with regard to global warming-related attitudes.

Although there are not consistent effects from each of the treatments, the results suggest that norms may have the potential to influence beliefs and behaviors, though only under particular conditions. The con norms, in particular, seem to have powerful effects. A separate analysis, pooling all pro norm conditions and the control group (which does not differ significantly from the pro norm conditions on most measures) compared to the pooled con norm conditions, shows that con norms suppressed belief that global warming was human induced (p = .02), willingness to take personal action (p = .06), support for an emissions cap (p = .11),¹² perceived personal influence (p = .00), and expected reciprocity (p = .08). Although most research (including our Experiment 2) has focused on the positive impacts of pro norms, these findings suggest con norms are particularly important. The results presented in Table 2 also suggest that party identification has a strong impact on global warming-related outcomes. Figure 1 presents the means scores by party within each treatment for belief that global warming is happening, belief that it is human induced, support for an emissions cap, and willingness to take personal action. Again, con norms without an injunctive statement tend to push Republicans further negative on these measures. We take up the impact of party identification in greater detail below.

Figure 1.

Conditional influence of partisanship on global warming belief.

Recall that Experiment 2 tested the impact of communications that invoked pro norms, which showed little impact in Experiment 1, in combination with messages that communicate scientific consensus about global warming. We find additional evidence that beliefs and behavioral intentions about global warming are quite sticky, even in the face of both pro norm and scientific communications. We again examine Hypotheses 1a and 2a regarding the effects of pro norms and provide our first tests of Hypotheses 3 and 4, regarding science-based messages. As is clear in Table 3, the effects of our manipulations are generally not statistically significant except on our measure of willingness to take action. As is clear from the large, positive, and statistically significant coefficients on party identification, much of the variation in all four outcomes is explained by partisan differences (as it was also the case in Experiment 1). Yet, at least as it relates to willing to take personal action, a pro norm significantly increases behavioral intent (p < .05). That the size of this effect is comparable to the effect of a scientific consensus message alone (p < .05) or both the norm and science message together (p < .05) suggests that efforts to influence personal behavior with regard to global warming can benefit from either approach. Yet the combination of these communications appears to have little added benefit: the strongest effects seem to come from the pro norm alone (providing support for Hypothesis 2a, but little for Hypothesis 4). This is consistent with the evidence from Experiment 1 that (con) descriptive norms alone had the largest and most significant effects on outcomes. The evidence thus lends little support to the expectations (as outlined in Hypotheses 3 and 4) that communicating scientific consensus has an impact on beliefs, attitudes, or behavioral intentions.

Table 3.

Treatment Effects (Experiment 2).

	Believe global warming is happening (1)	Believe global warming is human induced (2)	Support cap on carbon emissions (3)	Willingness to take personal action (4)
Intercept	0.40** (0.05)	0.17** (0.06)	0.38** (0.05)	0.19** (0.06)
Norm/no science	−0.13* (0.08)	0.03 (0.08)	0.05 (0.08)	0.23** (0.09)
No norm/science	0.08 (0.07)	0.07 (0.08)	0.12** (0.07)	0.17** (0.09)
Norm/science	0.05 (0.07)	0.09 (0.08)	0.07 (0.07)	0.18** (0.08)
Party ID	0.18** (0.07)	0.36** (0.07)	0.19** (0.08)	0.26** (0.09)
Party ID × norm/no science	0.05 (0.12)	−0.17* (0.11)	0.00 (0.11)	−0.17* (0.13)
Party ID × no norm/science	−0.11 (0.10)	−0.27** (0.12)	−0.15* (0.10)	−0.10 (0.13)
Party ID × norm/science	−0.03 (0.10)	−0.31** (0.11)	−0.04 (0.10)	−0.08 (0.11)
SER	0.46	0.48	0.46	0.42
N	382	382	382	382

Note: Cell entries are unstandardized linear regression coefficients with bootstrapped standard errors in parentheses. Baseline is the control (no norm/no science) treatment condition. Party identification is coded with seven categories from Republican (−1) to Independent (0) to Democrat (+1). SER is the standard error of the regression (Beck, 2010).

p < .10. **p < .05. One-tailed test.

Finally, we turn to the question of partisan moderation effects—the “boomerang effect” (Hart & Nisbet, 2011)—that we described in our exploratory hypothesis. In both experiments, party identification was consistently and significantly associated with higher scores on all outcome measures (i.e., Democrats reported higher scores on measures of beliefs, attitudes, and behavioral intentions).¹³ As an example, strong Democrats in the control condition reported, on average, a score .52 points higher on willingness to take personal action than strong Republicans, a full 25% of the response scale (we find a similar pattern of partisan effects in Experiment 1).

These results are perhaps unsurprising given prior evidence of a partisan divide on global warming. More interesting are interactions between partisanship and the treatment indicators, which report the effect of partisanship under each type of norm-based communication. Our results suggest an interesting pattern that runs somewhat contrary to Hart and Nisbet’s (2011) boomerang hypothesis. In the regression models reported in Tables 2 and 3, we can interpret each coefficient on an interaction term (between partisanship and a treatment) as follows: a coefficient of zero implies no partisan differences in the outcome (i.e., under that treatment, Democrats and Republicans do not significantly differ from one another), a positive coefficient indicates that the treatment exacerbates partisan differences (e.g., Democrats are even more willing to take action and Republicans less than they otherwise would be), whereas a negative coefficient on an interaction implies that the treatment reduces partisan differences.¹⁴ These coefficients indicate that norm communications (in Experiment 1 and Experiment 2) and scientific consensus communications (in Experiment 2) undercut partisan differences.¹⁵ In Experiment 1, partisan differences are almost entirely eliminated by norm-based communications for perceived personal influence, expected reciprocity, and belief that global warming is anthropogenic. In Experiment 2, we find a similar pattern for both willingness to take action and belief that global warming is anthropogenic. With regard to willingness to take action, a pro norm alone nearly eliminates partisan differences in behavioral intention though the other treatments have less dramatic effects. Much more dramatic are the effects of norms and science communications in reducing partisan differences in belief that global warming is happening, in belief that global warming is human induced, in support for an emissions cap, and in willingness to change, all from Experiment 2. We display this pattern in Figure 2, which shows the mean scores by party within each treatment condition. Looking at the plot whether global warming is human induced, in particular, shows that combinations of science and norm-based communications can be particularly powerful in mitigating the partisan divide about global warming. And, norms alone are nearly sufficient to bridge the divide on willingness to take personal action, whereas science alone is sufficient to bridge the (smaller) divide on support for a carbon emissions cap.

Figure 2.

Conditional influence of partisanship on global warming belief.

As either a pro norm or a science communication are added, partisan differences are significantly reduced. When both are expressed, there are almost no partisan differences in belief that global warming is anthropogenic. Though both experiments showed that pro norms seem to have minimal ability to increase the public’s aggregate beliefs and attitudes about global warming, they may undercut partisan differences on the issue, which are known to be severe. This suggests that these types of communications are effective even on those who are predisposed to not believe in global warming. As well, given that it is Republicans who are responding to the science-only treatment in most cases, we reject Hart and Nisbet’s (2011) boomerang hypothesis.

Discussion

Despite the frequently null effects, the results of our experiments do show that norm-based treatments can to some extent directly shape beliefs, policy support, intentions, and actions. Supporting previous work, perceptions of efficacy appear to be central to the process by which norms influence behavior in collective action settings (Bolsen, 2013; Finkel et al., 1989; Lubell, 2002; Lubell et al., 2007). It is especially important to read the results in the context of the particular issue under examination. Global warming is not a new issue nor is one that individuals are likely to be unfamiliar with or have no attitudes about. To the extent that individuals’ attitudes toward these types of issues are well formed and strongly held, it can be quite difficult to influence attitudes and behavioral intentions (Druckman & Leeper, 2012a). While much experimental work avoids these types of issues to focus instead on novel or obscure issues (Druckman & Leeper, 2012b), our interest in collective action necessitates a focus on a real and important issue. That we find any treatment effects at all and that some of those effects are large suggests the potential of norm-based communications to influence even well-entrenched social dilemmas. The results regarding con norms are particularly important: Experiment 1 clearly demonstrates that when others are perceived as less supportive of a carbon tax or driving less, individuals are significantly less likely to support or engage in the collective endeavor themselves. Norms can both serve to solve and exacerbate collective action problems. Experiment 2 showed that science-based communications, which are commonly referenced in discussions of global warming, are relatively less influential.

Additionally, the absence of consistently significant main effects of the treatments for both experiments reveals challenges in properly measuring beliefs and willingness to act. These hazards are well documented in the existing research, of which the preponderance of null effects here is additional evidence. It is also possible, though, that the null effects are due to the apolitical nature of the treatments (i.e., nonpartisan polls; Montpetit, 2011). Without an identifiable media source, participants in our experiments had little basis for evaluating the credibility or trustworthiness of the communications they received. This could create a problem, especially as information source plays a significant role in how citizens judge scientific evidence (Nisbet & Goidel, 2007). Particularly in light of evidence that partisanship biases the acceptance of science-based communications (Jasanoff, 2011; Nisbet, 2005), and that the politicization of science-based communications is a key confounding mechanism (Jasanoff & Wynne, 1998; Kitcher, 2001; Pielke, 2002, 2004, 2007; Sarewitz, 1996, 2004), future research should explore how the politicization of science, which necessarily involves the partisan communication of scientific information, influences public attitudes toward global warming.

Inconsistencies between the results for Experiment 1 and Experiment 2 could be driven by differences in the language of the pro-norm treatments, but we suspect that exact percentages may not matter. As Kahneman (2011) points out, the human mind does not process numbers well but is instead responsive to impressions that anyone (rather than no one) is doing something. Attempts to increase external validity are best addressed with an ongoing research program replicating these experiments across populations, experimental treatments, and time (Druckman & Kam, 2011).

This is one of the first studies to explore the causal process by which norms influence attitudes and behavior toward the issue of global warming. Though our results are mixed, we isolate some conditions where norms do influence beliefs, attitudes, and behaviors. Our results also highlight the need to focus more on the influence of con norms (from our results in Experiment 1), focus on the combination of norm- and science-based information into theorizing about collective action (from our results in Experiment 2), and the need to incorporate individuals’ partisanship in work on all of these issues. The results shed light on the process by which norms affect behavior in collective action settings, and also provide insight for policymakers regarding how communications that promote social comparisons may be used to foster voluntary behavior change (e.g., by including information about the efficacy of individual action and/or highlighting cooperation by others). Despite the stickiness of attitudes toward global warming, norms seem to have some impact on behavioral intentions—the one area where individuals are able to make an immediate impact on climate mitigation efforts.

Footnotes

Appendix. Question Wordings

Declaration of Conflicting Interests

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

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Toby Bolsen is an Assistant Professor of Political Science at Georgia State University. He studies political behavior, public opinion, media and communications, experimental methods, and U.S. energy policy.

Thomas J. Leeper is a Postdoc in the Department of Political Science and Government at Aarhus University. He studies the role of information in public opinion formation.

Matthew A. Shapiro is an Assistant Professor of Political Science at the Illinois Institute of Technology, where he focuses on information transfer problems in the context of environmental and science policies. His research looks particularly at the connections between publicly funded R&D and environmentally-related outcomes, how the public responds to the relevant science, and how the government incentivizes change in the public and private sectors.

References

Ajzen

Fishbein

(2005). The influence of attitudes on behaviors. In Albarracin

Johnson

Zanna

(Eds.), The handbook of attitudes. London, UK: Lawrence Erlbaum.

Allcott

(2010). Social norms and energy conservation. Typescript, MIT and NYU.

Allcott

Mullainathan

(2010). Behavior and energy policy. Science, 327, 1204-1205.

Alpízar

Carlsson

Johansson-Stenman

(2008). Anonymity, reciprocity, and conformity: Evidence from voluntary contributions to a national park in Costa Rica. Journal of Public Economics, 92, 1047-1062.

Angrist

J. D.

Pischke

(2009). Mostly harmless econometrics. Princeton, NJ: Princeton University Press.

Axelrod

(1984). The evolution of cooperation. New York, NY: Basic Books.

Bauer

M. W.

Allum

Miller

(2007). What can we learn from 25 years of PUS survey research? Liberating and expanding the agenda. Public Understanding of Science, 16(1), 79-95.

Beck

(2010). Making regression and related output more helpful to others. The Political Methodologist, 18(1), 4-9.

Biel

Thogersen

(2006). Activation of social norms in social dilemmas: A review of the evidence and reflections on the implications for environmental behaviour. Journal of Economic Psychology, 28, 93-112. doi:10.1016/j.joep.2006.03.003>

10.

Bolsen

(2013). A light bulb goes on: norms, rhetoric, and actions for the public good. Political Behavior, 35, 1-20.

11.

Chong

Druckman

J. N.

(2007). Framing public opinion in competitive democracies. American Political Science Review, 101, 637-655.

12.

Cialdini

R. B.

(2003). Social influence: Compliance and conformity. Annual Review of Psychology, 55, 591-621. doi:10.1146/annurev.psych.55.090902.142015

13.

Cialdini

R. B.

Reno

R. R.

Kallgren

C. A.

(2004). A focus theory of normative conduct: Recycling the concept of norms to reduce littering in public places. In Aronson

(Ed.), Readings about the Social Animal. New York, NY: Worth Publishers.

14.

Costa

D. L.

Kahn

M. E.

(2012). Energy conservation “nudges” and environmentalist ideology: Evidence from a randomized residential electricity field experiment. Manuscript. UCLA and NBER.

15.

Druckman

J. N.

(2001). The implications of framing effects for citizen competence. Political Behavior, 23, 225-256.

16.

Druckman

J. N.

(2004). Political preference formation: Competition, deliberation, and the (ir)relevance of framing effects. American Political Science Review, 98, 671-686.

17.

Druckman

J. N.

Kam

C. D.

(2011). Students as experimental participants: A defense of the “narrow data base.” In Druckman

J. N.

Green

D. P.

Kuklinski

J. H.

Lupia

(Eds.), Cambridge handbook of experimental political science. New York, NY: Cambridge University Press.

18.

Druckman

J. N.

Leeper

T. J.

(2012a). Learning more from political communication experiments: Pretreatment and its effects. American Journal of Political Science, 56, 875-896.

19.

Druckman

J. N.

Leeper

T. J.

(2012b). Is public opinion stable? Daedalus, in press. Daedalus, 141, 50-68.

20.

Dunlap

R. E.

McCright

A. M.

(2008). A widening gap: Republican and democratic views on climate change. Environment, 50(5), 26-35.

21.

Evans

M. S.

(2011). Supporting science: Reasons, restrictions, and the role of religion. Science Communication. Advance online publication. doi:10.1177/1075547011417890

22.

Finkel

S. E.

Muller

E. N.

Opp

(1989). Personal influence, collective rationality, and mass political action. The American Political Science Review, 83, 885-903.

23.

Frey

Meier

(2004). Social comparison and pro-social behavior: Testing “conditional cooperation” in a field experiment. American Economic Review, 94, 1717-1722.

24.

Gelman

Stern

(2006). The difference between “significant” and “not significant” is not itself statistically significant. The American Statistician, 60, 328-331.

25.

Green

Gerber

A. S.

(2010). Introduction to social pressure and voting: New experimental evidence. Political Behavior, 32, 331-336.

26.

Hart

P. S.

Nisbet

E. C.

(2011). Boomerang effects in science communication: How motivated reasoning and identity cues amplify opinion polarization about climate mitigation policies. Communication Research. Advance online publication. doi:10.1177/0093650211416646

27.

Jasanoff

(2011). Making the facts of life. In Jasanoff

(Ed.), Reframing rights: Bioconstitutionalism in the genetic age. Cambridge, MA: MIT Press.

28.

Jasanoff

Wynne

(1998). Science and decisionmaking. In Rayner

Malone

E. L.

(Eds.), Human choice and climate change: Vol. 1. The societal framework. Columbus, OH: Battelle Press.

29.

Kahneman

(2011). Thinking fast and slow. New York, NY: Farrar, Straus and Giroux.

30.

Kitcher

(2001). Science, truth, and democracy. New York, NY: Oxford University Press.

31.

Klick

Smith

(2010). Salvation by science: Confidence in scientists and engineers to solve global warming. Paper presented at the 2010 Annual Meeting of the Western Political Science Association, San Francisco, CA.

32.

Kogut

Beyth-Marom

(2008). Who helps more? How self-other discrepancies influence decisions in helping situations. Judgment and Decision Making, 3, 595-606.

33.

Krosnick

J. A.

Holbrook

A. L.

Visser

P. S.

(2000). The impact of the fall 1997 debate about global warming on American public opinion. Public Understanding of Science, 9, 239-260.

34.

Lubell

(2002). Environmental activism as collective action. Environment and Behavior, 28, 431-454.

35.

Lubell

Zahran

Vedlitz

(2007). Collective action and citizen responses to global warming. Political Behavior, 29, 391-413.

36.

Lupia

McCubbins

(1998). The democratic dilemma. Cambridge, MA: Cambridge University Press.

37.

Malka

Krosnick

J. A.

Langer

(2009). The association of knowledge with concern about global warming: Trusted information sources shape public thinking. Risk Analysis, 29, 633-747.

38.

McKenzie-Mohr

(2000). Fostering sustainable behavior through community-based social marketing. American Psychologist, 55, 531-537.

39.

Montpetit

(2011). Scientific credibility, disagreement, and error cost in 17 biotechnology policy subsystems. Policy Studies Journal, 39, 513-533.

40.

Mutz

(2008). How the mass media divide us. In Nivola

Brady

(Eds.), Red and blue nation? Washington, DC: Brookings Institution Press.

41.

Nisbet

M. C.

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

42.

Nisbet

M. C.

(2005). The competition for worldviews: Values, information, and public support for stem cell research. International Journal of Public Opinion Research, 17(1), 90-112.

43.

Nisbet

M. C.

Goidel

R. K.

(2007). Understanding citizen perceptions of science controversy: Bridging the ethnographic-survey research divide. Public Understanding of Science, 16, 421-440.

44.

Nisbet

Scheufele

D. A.

(2009). What’s next for science communication? American Journal of Botany, 96, 1-12.

45.

Noelle-Neumann

(1984). The spiral of silence. A theory of public opinion—our social skin. Chicago, IL: The University of Chicago Press.

46.

Nolan

Schultz

Cialdini

Goldstein

Griskevicius

(2008). Normative social influence is underdetected. Personality and Social Psychology Bulletin, 34, 913-923.

47.

O’Keefe

(2002). Persuasion. London, UK: SAGE.

48.

Olson

(1965). The logic of collective action: Public goods and the theory of groups. Cambridge, MA: Harvard University Press.

49.

Opp

K. D.

(2001). Why do people vote? Kyklos, 54, 355-378.

50.

Pielke

R. A

Jr. . (2007). The Honest broker: Making sense of science in policy and politics. London, UK: Cambridge University Press.

51.

Pielke

R. A

Jr. . (2002). Policy, politics, and perspective: The scientific community must distinguish analysis from advocacy. Nature, 4166, 367-368.

52.

Pielke

R. A

Jr. . (2004). When scientists politicize science: Making sense of controversy over. The Skeptical Environmentalist. Environmental Science & Policy, 7, 405-417.

53.

Sarewitz

(1996). Frontiers of illusion: Science, technology, and the politics of progress. Philadelphia, PA: University Temple Press.

54.

Sarewitz

(2000). Science and environmental policy: an excess of objectivity, In Frodeman

(Ed.), Earth matters: The earth sciences, philosophy, and the claims of community. Upper Saddle River, NJ: Prentice Hall, pp. 79-98.

55.

Sarewitz

(2004). How science makes environmental controversies worse. Environmental Science & Policy, 7, 385-403.

56.

Schuldt

J. P.

Konrath

S. H.

Schwartz

(2011). “Global warming” or “climate change”? Whether the planet is warm depends on question wording. Public Opinion Quarterly, 75(1), 115-124.

57.

Schultz

W. P.

Nolan

J. N.

Cialdini

R. B.

Goldstein

N. J.

Griskevicius

(2007). The constructive, destructive, and reconstructive power of social norms. Psychological Science, 18, 429-434.

58.

Shang

Croson

(2009). A field experiment in charitable contribution: The impact of social information on the voluntary provision of public goods. The Economic Journal, 119, 1422-1439.

59.

Stern

(2007). The economics of climate change. Cambridge, UK: Cambridge University Press.

60.

Thogersen

(2008). Social norms and cooperation in real-life social dilemmas. Journal of Economic Psychology, 29, 458-472.

61.

Villar

Krosnick

J. A.

(2010). Global warming vs. climate change, taxes vs. prices: Does word choice matter? Climatic Change, 105(1-2), 1-12. doi:10.1007/s10584-010-9882-x

62.

Zaller

(1992). The nature and origin of mass opinion. New York, NY: Cambridge University Press.