Nature: Seeming public apathy over climate change is often attributed to a deficit in comprehension. The public knows too little science, it is claimed, to understand the evidence or avoid being misled1. Widespread limits on technical reasoning aggravate the problem by forcing citizens to use unreliable cognitive heuristics to assess risk2. We conducted a study to test this account and found no support for it. Members of the public with the highest degrees of science literacy and technical reasoning capacity were not the most concerned about climate change. Rather, they were the ones among whom cultural polarization was greatest. This result suggests that public divisions over climate change stem not from the public’s incomprehension of science but from a distinctive conflict of interest: between the personal interest individuals have in forming beliefs in line with those held by others with whom they share close ties and the collective one they all share in making use of the best available science to promote common welfare.
The study collected data on the climate-change risk perceptions of a large representative sample of US adults (N=1,540). Measures were selected to permit assessment of two competing accounts of public opinion on climate change. One, already adverted to, can be called the science comprehension thesis (SCT). As members of the public do not know what scientists know, or think the way scientists think, they predictably fail to take climate change as seriously as scientists believe they should3.
The alternative explanation can be referred to as the cultural cognition thesis (CCT). CCT posits that individuals, as a result of a complex of psychological mechanisms, tend to form perceptions of societal risks that cohere with values characteristic of groups with which they identify4, 5. Whereas SCT emphasizes a conflict between scientists and the public, CCT stresses one between different segments of the public, whose members are motivated to fit their interpretations of scientific evidence to their competing cultural philosophies6.
Explanations for the public’s perceptions of climate change risk can be tested by observational study insofar as such hypotheses imply correlations between concern over climate change and specified individual characteristics7. We instructed subjects to rate the seriousness of climate change risk on a scale of 0 (no risk) to 10 (extreme risk), a general risk-concern measure that furnishes a parsimonious focus for such testing8, 9.
SCT asserts, first, that ordinary members of the public underestimate the seriousness of climate change because of the difficulty of the scientific evidence3. If this is correct, concern over climate change should be positively correlated with science literacy—that is, concern should increase as people become more science literate.
Second, and even more important, SCT attributes low concern with climate change to limits on the ability of ordinary members of the public to engage in technical reasoning. Recent research in psychology posits two discrete forms of information processing: system 1, which involves rapid visceral judgments that manifest themselves in various decision-making heuristics; and system 2, which requires conscious reflection and calculation10. Most members of the public, according to this research, typically employ system 1 reasoning without resorting to more effortful system 2 processing. Although system 1 works well for most daily contingencies, ordinary citizens’ predominant reliance on heuristic rather than analytic modes of reasoning is viewed as leading them to underestimate climate change risks, which are remote and abstract compared with a host of more emotionally charged risks (for example, terrorism) that the public is thought to overestimate2, 3.
If this position is correct, one would also expect concern with climate change to be positively correlated with numeracy. Numeracy refers to the capacity of individuals to comprehend and make use of quantitative information11. More numerate people are more disposed to use accuracy-enhancing system 2 forms of reasoning and are less vulnerable to system 1 cognitive errors11, 12. Hence, they should, on this view, form perceptions of climate-change risk less biased towards underestimation.
These predictions were unsupported (Fig. 1). As respondents’ science-literacy scores increased, concern with climate change decreased (r=−0.05, P=0.05). There was also a negative correlation between numeracy and climate change risk (r=−0.09, P<0.01). The differences were small, but nevertheless inconsistent with SCT, which predicts effects with the opposite signs.
Figure 1: SCT prediction versus actual impact of science literacy and numeracy on climate change risk perceptions.
Contrary to SCT predictions, higher degrees of science literacy and numeracy are associated with a small decrease in the perceived seriousness of climate change risks. Derived from Supplementary Table S4, Model 1. Low and high reflect values set at −1 s.d. and +1 s.d. on the composite Science literacy/numeracy scale (see Supplementary Information). Responses on the 0–10 risk scale (M=5.7, s.d.=3.4) were converted to z-scores to promote ease of interpretation. Confidence intervals reflect the 0.95 level of confidence.
CCT also generates a testable prediction. CCT posits that people who subscribe to a hierarchical, individualistic world-view—one that ties authority to conspicuous social rankings and eschews collective interference with the decisions of individuals possessing such authority—tend to be sceptical of environmental risks. Such people intuitively perceive that widespread acceptance of such risks would license restrictions on commerce and industry, forms of behaviour that hierarchical individualists value. In contrast, people who hold an egalitarian, communitarian world-view—one favouring less regimented forms of social organization and greater collective attention to individual needs—tend to be morally suspicious of commerce and industry, to which they attribute social inequity. They therefore find it congenial to believe those forms of behaviour are dangerous and worthy of restriction4. On this view, one would expect egalitarian communitarians to be more concerned than hierarchical individualists with climate change risks.
Our data, consistent with previous studies6, supported this prediction. Hierarchical individualists (subjects who scored in the top half on both the Hierarchy and Individualism cultural-world-view scales) rated climate change risks significantly lower (M=3.15, s.e.m.=0.17) than did egalitarian communitarians (subjects whose scores placed them in the bottom half; M=7.4, s.e.m.=0.13). Even controlling for scientific literacy and numeracy (as reflected in the composite scale Science literacy/numeracy; see Supplementary Information), both Hierarchy (b=−0.46, P<0.01) and Individualism (b=−0.30, P<0.01) predicted less concern over climate change (Supplementary Table S4).
These findings were consistent, too, with previous ones showing that climate change has become highly politicized13, 14. Cultural-world-view and political-orientation measures are modestly correlated. Nevertheless, the impact that cultural world-views has on climate change risk perceptions cannot be reduced to partisanship. The mean hierarchical individualist in our sample was an Independent who leans Republican and is slightly conservative; the mean egalitarian communitarian was also an Independent, but one who leans Democrat and is slightly liberal (Supplementary Fig. S4). The difference between their respective perceptions of climate change risk, however, significantly exceeded what political-orientation measures alone would predict for individuals who identify themselves as conservative Republicans and liberal Democrats (Supplementary Fig. S5).
The finding that cultural world-views explain more variance than science literacy and numeracy, however, does not by itself demonstrate that SCT is less supportable than CCT. SCT asserts not merely that members of the public lack scientific knowledge but also that they lack the habits of mind needed to assimilate it, and are thus constrained to rely on fallible heuristic alternatives. Proponents of this bounded-rationality position treat cultural cognition—the conforming of beliefs to the ones that predominate within one’s group—as simply one of the unreliable system 1 heuristics used to compensate for the inability to assess scientific information in a dispassionate, analytical manner15.
This claim generates another testable prediction. If cultural cognition is merely a heuristic substitute for scientific knowledge and system 2 reasoning, reliance on it should be lowest among those individuals whose scientific knowledge and system 2 reasoning capacity are highest. SCT thus implies that as science literacy and numeracy increase, the scepticism over climate change associated with a hierarchical individualistic world-view should lessen and the gap between people with hierarchical individualistic world-views and those with egalitarian communitarian ones should diminish.
However, this SCT prediction, too, was unsupported. Among egalitarian communitarians, science literacy and numeracy (as reflected in the composite scale Science literacy/numeracy) showed a small positive correlation with concern about climate change risks (r=0.08, P=0.03). In contrast, among hierarchical individualists, Science literacy/numeracy is negatively correlated with concern (r=−0.12, P=0.03). Hence, polarization actually becomes larger, not smaller, as science literacy and numeracy increase (Fig. 2 and Supplementary Table S4 and Fig. S3). As the contribution that culture makes to disagreement grows as science literacy and numeracy increase, it is not plausible to view cultural cognition as a heuristic substitute for the knowledge or capacities that SCT views the public as lacking.
