Climate action literacy interventions increase commitments to more effective mitigation behaviors

Abstract

Reducing lifestyle carbon emissions is a critical component of decarbonizing society. However, people hold substantial misperceptions about the relative efficacy of different behavioral changes, such as comprehensively recycling or avoiding long flights, and these misperceptions may lead to the suboptimal allocation of resources. In a preregistered experiment in the United States, we tested the effects of two literacy interventions on correcting misperceptions and increasing commitments toward more effective individual-level climate actions. Participants (n = 3,895) were randomly assigned to one of three experimental conditions: a Prediction condition, in which they were asked to rank the relative mitigation potential of 21 climate behaviors after which they received feedback; an Information condition, in which they were passively exposed to information about the relative mitigation potential of the same behaviors; and a no-information Control condition. Both the Prediction and Information interventions led to more accurate efficacy perceptions and increased commitments to engage in higher-impact individual-level actions relative to the Control group. Greater initial misperceptions were associated with larger shifts in commitments, such that participants reduced commitments to behaviors that were overestimated and increased commitments to behaviors that were underestimated in their carbon reduction potential. However, we also found evidence for a negative spillover effect from individual to collective actions: participants in the literacy conditions decreased their commitments to collective climate actions such as voting or marching, suggesting an unintended consequence of interventions focusing solely on individual-level actions.

Keywords: behavior; climate; intervention; misperception; spillover.

Fig. 1.

A) Map of individual-level climate behaviors on two key dimensions: carbon mitigation potential as detected in a meta-analysis by Ivanova et al. (2) (x-axis) and behavioral plasticity as assessed here (y-axis). B) Prediction errors (i.e. the difference between the predicted and actual rank) associated with each of the 21 individual-level climate-relevant behaviors assessed. Error bars represent 95% CIs.

Fig. 2.

A) Real vs. perceived climate action efficacy. Steeper positive slopes correspond to more accurate predictions of action efficacy. X-axis units are absolute carbon mitigation potential per year for each behavior. B) Real climate action efficacy as a function of change in commitments from pretest to posttest, in each of the three conditions.

Fig. 3.

A) Prediction errors (difference between real and perceived relative effectiveness of climate actions) as a function of changes in commitments in corresponding actions, from pretest to posttest. Larger underestimates with correction increase commitment, while overestimates with correction reduce commitment. B) Experimental conditions (Control, Information, and Prediction split into Overestimates–Underestimates–Accurate) as a function of commitments to engage in climate actions.

Fig. 4.

Prediction errors (difference between real and perceived relative effectiveness of climate actions) as a function of changes in commitments in corresponding actions, from pretest to posttest, split by political affiliation.

Fig. 5.

A) Changes in commitments from pretest to posttest in collective actions as a function of experimental condition. B) Changes in perceptions of efficacy of collective actions as a function of experimental condition. Error bars represent SEs of the mean.

Fig. 6.

A) Pretest commitments to individual-level actions as a function of perceived behavioral plasticity and perceived efficacy. B) Pretest commitments to collective-level actions as a function of perceived behavioral plasticity and perceived efficacy.

Fig. 7.

Screenshot of the survey material depicting the true feedback.

Fig. 8.

Screenshot of the survey material depicting the ranking prediction task and true feedback.