Psychological Reactance From Reading Basic Facts on Climate Change: The Role of Prior Views and Political Identification

ABSTRACT

Scholars continue to search for solutions to shift climate change skeptics’ views on climate science and policy. However, research has shown that certain audiences are resistant to change regarding environmental issues. To explore this issue further, we examine the presence of reactance among different audiences in response to simple, yet prominently used, climate change messages. Our results show that emphasizing the scientific consensus of climate change produces reactance, but only among people who question the existence of climate change. Moreover, adding political identification to the model as an additional moderating variable shows the increases in reactance occur among Republicans who question the existence of climate change. Finally, our results show that reactance to climate change messaging may lead to backfiring effects on important outcomes tied to climate change such as risk perceptions, climate change beliefs, and support for mitigation policies.

KEYWORDS:

• Climate change communication
• motivated reasoning
• experimental design

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Yanni Ma http://orcid.org/0000-0003-2144-4076

Notes

1. We ran these models separately with the PROCESS macro, using model 7. Indices for moderated-mediation were significant for all three of our outcome variables: support for mitigation policy (B = .06 95% CI[.019, .13]), post climate change belief (B = .15 95% CI[.04, .26]), and climate change risk perceptions (B = .07 95% CI[.019, .14]). We also ran one model in Mplus where all three outcomes were included at the same time. The results of this model were generally the same as the results from using the PROCESS macro in SPSS.

2. We ran three separate models using model 11 with the PROCESS macro. The indices of moderated moderated mediation were statistically significant for support for mitigation policies (B = .02, 95% CI[.001, .052]) and climate change risk perceptions (B = .02, 95% CI[.0001, .53]). This index was not significant for post climate change belief (B = .05, 95% CI[−.01, .1]). Moreover, we once again ran the whole model in Mplus with the three outcome variables entered at the same time and found similar results as the ones reported in the manuscript using the PROCESS macro in SPSS (see online Appendix Figure A).