Who are your people? – The effect of political ideology and social identity on climate-related beliefs and risk perceptions

ABSTRACT

Persistent divides among American voters regarding climate change, especially climate skepticism among conservatives, have long been explained with reference to ideology, vested interests, and trends of political polarization. More recently, an alternative set of explanations for the opinion gap between conservatives and liberals has been gaining traction, arguing that these divisions are generated by social identities and their effects on individual beliefs and attitudes. Here, we focus on global citizenship as a specific social identity. Seeking to connect ideology and social identity approaches, we study how the interaction between a person’s ideological leanings and their social identity as a global citizen relates to beliefs and risk perceptions regarding climate change. Analyzing two kinds of survey data, we find that a global citizen identity moderates the relationship between a conservative ideology and a person’s climate-related beliefs and risk perceptions, while it does not seem to have the same effect for liberal individuals. In other words, a global citizen identity is associated with a potential decrease in the ideological divide between conservatives and liberals regarding climate change. We explore the implications of these findings for climate change communication and policy and other issue areas.

KEYWORDS:

• Climate change
• social identity
• global citizenship
• global identity
• political ideology

Disclosure statement

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

Notes

1 The “alpha” function in STATA generates the summative scale from the items (variables) specified by computing interitem correlations and Cronbach’s alpha. The items that correlated negatively with the rest of the items (2 and 3 above) were reversed.

2 For added robustness, we also created a latent variable of climate change beliefs combining the 6 survey items using a confirmatory factor analysis via structural equation modeling. Tests conducted on this latent variable confirm our main results on the effect of global citizenship and political ideology on climate change beliefs, adding confidence to our findings. Appendix 2 shows the factor loadings in the new variable and these tests results.

3 Among people who identified as global citizens, the means of climate change beliefs no longer changed in order along the ideological spectrum. Mean climate change belief scores for very liberal, somewhat liberal, somewhat conservative, and very conservative global citizens were .922 ± .038, .902 ± .024, .756 ± .042, and .970 ± .132, respectively. The global citizen identity narrowed the gap of very liberal and somewhat conservative citizens’ climate change beliefs. Very liberal global citizens had a higher mean “climate change beliefs” than somewhat conservative global citizens (.166, 95%CI: .014, .318, p = .024), a narrower gap compared to the .227 among non-global citizens. Somewhat liberal global citizens also still reported a higher mean of climate change beliefs than somewhat conservative global citizens, .146 (95%CI: .016, .275), p = .018, but this gap was smaller than between non-global citizens who were somewhat liberal and somewhat conservative (.168).

4 Among non-global citizens, climate change beliefs decreased from higher to lower levels along the ideological spectrum when moving from very liberal to very conservative. Mean “climate change beliefs” scores for very liberal, somewhat liberal, somewhat conservative, and very conservative individuals were .943 ± .057, .884 ± .033, .716 ± .038, and .530 ± .060, respectively. Even when not identifying as global citizens, liberals still overall reported higher means of climate change beliefs compared to conservatives. Specifically, very liberal individuals had a higher mean of climate change beliefs than somewhat conservative individuals, with a mean difference of .227 (95%CI: .044, .410), p = .007, and very conservative individuals, .413 (95%CI: .190, .635), p < .0005. Somewhat liberal individuals also reported a higher mean of climate change beliefs compared to somewhat conservative individuals, .168 (95%CI: .034, .302), p = .006, and very conservative individuals, .354 (95%CI: .173, .535), p < .0005.

5 The interaction term is statistically significant at the 90 percent confidence level.

6 Graphs that plot the ordered logit model in Table 3, as well as a similar model that uses a binary measure of global citizen identification for improved visibility, are included in the appendix.

7 Moderates (5 on the 1–10 ideology scale) measured in between the extreme conservatives and liberals. The stronger they indicated a global citizen identity, the less likely they were to report that global warming or the greenhouse effect were not serious at all. Throughout the range of moderates who strongly disagreed to strongly agreed (4 scale) of having a global citizen identity, the probability they reported global warming or the greenhouse effect as not serious at all decreased, from 9.7% (p = .003, 95%CI: .034, .160), 3.5% (p < .0005, 95%CI: .023, .047), 3.2% (p < .0005, 95%CI: .023, .042), and finally 2% (p < .0005, 95%CI: .012, .027), respectively.

8 Moderates were in between the extreme conservatives and liberals, but still were more likely to report global warming and the greenhouse effect as very serious the stronger their global citizen identity. Moderates who strongly disagreed, and disagreed, that they were global citizens each had a 24.3% (p < .0005, 95%CI: .114, .371) and 48.4% (p < .0005, 95%CI: .425, .544) probability of reporting that global warming and the greenhouse effect were very serious. Moderates who agreed, and strongly agreed, that they were global citizens, reported with a 50.4% (p < .0005, 95%CI: .460, .547) and 62.6% (p < .0005, 95%CI: .562, .690) probability that global warming and the greenhouse effect were very serious, respectively.