Determinants for Accepting Climate Change Mitigation Policies: A Meta-Analysis

Public acceptance is a precondition for implementing climate change mitigation policies. What, then, determines acceptance of these policies? Based on 76 datasets from 34 countries, generating a total sample of 146,817 participants, we report a series of meta-analyses assessing the importance of 15 determinants for accepting climate change mitigation policies. Results show the following: (a) Among policy-specic beliefs, perceived fairness is the most important factor. (b) Among climate change beliefs, knowledge about climate change is weakly related to acceptance. Climate change beliefs, environmental concern, and perceived risk of and problems associated with climate change are all related to acceptance. (c) Among psychological factors, trust is most important. (d) Finally, demographic variables show weak or no relationship with acceptance. These results inform climate policy researchers as to which determinants of acceptance to include in future analyses and provide advice to policymakers about which sentiments they should consider when introducing and communicating intended climate policies.


Introduction
Policy intervention is a cornerstone in the efforts to mitigate climate change. Given the alarming threat of climate change (1,2), it is urgent to increase the understanding of determinants of public acceptance of climate change mitigation policies. Policies such as fuel taxation have shown to e ciently reduce carbon emissions (3). Yet, public acceptance is crucial for successful policy implementations. If politicians perceive risks of public resistance, they will be reluctant to implement the policy because of both potential social unrest and policy evasion.
Research on determinants for accepting climate mitigation policies has generated numerous studies across various academic disciplines (4)(5)(6)(7)(8). This research has applied models proposing that policyspeci c beliefs, climate change beliefs, and psychological factors determine acceptance (9,10). Studies report that acceptance is linked to determinants such as perceived fairness, environmental concerns, and biospheric values (11)(12)(13).
This line of research is of utmost importance for policymakers in the process of designing and implementing climate policies. The particular question the research has to answer is how much the identi ed determinants matter. To date, no study has meta-analytically summarised the degree to which the determinants are related to people's propensity to accept climate change policy instruments. This study aims at lling this knowledge gap.
In a meta-analysis, we synthesise the predictive strength of determinants for accepting climate change mitigation policies. Based on 76 eligible datasets, we investigate the importance of 15

Method Eligibility criteria
For inclusion in the meta-analysis, studies had to meet the following eligibility criteria: 1. Studies were included if reporting demographic variables, psychological variables, policy-speci c variables, or climate change beliefs that can be correlated with accepting climate change policies.
Studies were excluded if they did not report such variables or did not ful ll statistical criteria for performing correlational analyses.
2. Studies were included if reporting a measure of accepting a climate change mitigation policy. We excluded studies that measured (a) unspeci c policies (for example, pro-environmental legislations or a new policy focusing on climate change research), (b) behaviors rather than policies (e.g., farming methods), (c) policies with ambiguous environmental bene ts (e.g., land use regulations), and (d) support for a technique rather than a policy (e.g., carbon capture and storage, nuclear power, or support for renewable energy).
3. We included studies measuring attitudes or behavioral intentions toward accepting climate change policies (i.e., acceptance, acceptability, support, willingness to support, or voting intention). We excluded studies measuring willingness to pay and studies measuring pro-environmental behaviors (e.g., willingness to act and self-reported pro-environmental behaviors). 4. To assess dispersion of the determinants included in our random-effects meta-analysis, we set the cutoff for inclusion of a speci c determinant to be represented by a minimum of six studies (14).

5.
We excluded studies using between-group designs to assess in uence on the dependent variable (e.g., testing whether framing effects, such as global warming versus climate change, in uence climate change beliefs). We included studies using within-group design but analysed the data for only one measurement.
6. Included studies must be reported in English.
7. We excluded studies assessing acceptance of environmental policies by speci c professional groups (e.g., farmers, policymakers).
8. Reports of studies (or authors) must provide unique data enabling us to retrieve at least one correlation coe cient.

Conceptual de nitions
Survey methodology. We de ne survey methods as studies using questionnaires administered online, through the mail, face-to-face, or in telephone interviews that generated quantitative data.
Determinants. We de ne antecedents as measures of demographic factors or psychological characteristics ful lling statistical criteria for performing a correlational analysis (e.g., age, education, income, problem awareness, fairness/justice, trust, values, perceived effectiveness of policy).
Policy acceptance. We de ne acceptance as measures of attitudes (negative versus positive, bad versus good intentions) or direct measures of accepting climate mitigation policies (do not accept/do not support versus accept/support).
Climate mitigation policies. We de ne a climate mitigation policy as a real or ctive policy with an explicit purpose of mitigating climate change. Hence, we included policies such as road pricing, carbon tax, and fuel tax.

Search strategies
First, we searched for studies in databases by using two search strategies to assess studies published in political science, psychology, economics, and environmental science (see Table 1). Second, we searched the reference lists of four review articles (6)(7)(8)15 Fig. 1 illustrates the inclusion or exclusion of studies based on these searches.

Search strategy 1
Survey AND ("road pricing" OR "congestion charge" OR "environmental tax*" OR "fuel tax*" OR "carbon tax*" OR "green deal" OR "water scheme" OR "carbon price*" OR "climate policy" OR "climate change" OR "global warming" OR "CO 2 emissions reductions" OR "regulations" OR "mitigation policy" OR "adaptation policy" OR "climate change policy") AND ("support" OR "accept*" OR "public support" OR "policy support" OR "public accept*" OR "policy accept*" OR "public preferences" OR "policy preferences" OR "public attitudes" OR "policy attitudes" OR "public opinion" OR "willingness to act" OR "voting intention" OR "citizen support") Search strategy 2 Survey AND ("road pricing" OR "congestion charge" OR "environmental tax*" OR "fuel tax*" OR "carbon tax*" OR "green deal" OR "carbon price*" OR "climate policy" OR "climate change" OR "global warming" OR "CO 2 emissions reductions" OR "regulations" OR "mitigation policy" OR "adaptation policy" OR "climate change policy" OR "greenhouse gas tax" OR "greenhouse gas" OR "CO 2 -tax" OR "carbon dioxide tax") AND ("support" OR "accept*" OR "public support" OR "policy support" OR "public accept*" OR "policy accept*" OR "public preferences" OR "policy preferences" OR "public attitudes" OR "policy attitudes" OR "public opinion" OR "willingness to act" OR "voting intention" OR "citizen support") For each study, we coded the following: author(s), year of publications, number of participants, assessed policy, and the Pearson product-moment correlation coe cients of all determinants ful lling the eligibility criteria. For studies lacking su cient statistical information, we contacted the authors(s) asking for statistics or raw data. All included studies and how variables were measured are described in Appendix A.

Assessing publication bias and dispersion
To assess publication bias, we rst conducted fail-safe N and trim and ll analyses. Results from the failsafe N report the number of missing studies with no effect needed to result in a trivial effect, de ned as r < +/-.05. The trim and ll analysis reports the estimated number of missing studies imputed. Results showed that imputation was performed for one determinant, self-enhancement values. Imputing two estimated missing studies resulted in a minor adjustment of effect size r from r = -.09 to r = -.07. In sum, analyses for assessing publication bias indicates that publication bias was not problematic. Finally, prediction intervals indicated substantial dispersion for some determinants; we followed this up by performing subgroup analyses when possible.

Results
Final sample The nal sample includes 61 articles and 76 datasets, from 34 countries, based on 146,817 participants. Analyses We ran separate random-effects meta-analyses for each determinant using comprehensive metaanalysis, resulting in 15 meta-analyses, summarised in Table 2. Note: 95% Con dence interval = the mean effect size with a 95% con dence interval, 95% Prediction interval = the true effect size in 95% of all comparable populations, p = probability value: the probability of giving these or more extreme results given that the null hypothesis is true, n = number of participants; k = number of studies, Fail-safe N = Orwin´s fail-safe N: the number of missing studies reporting no effect needed to result in a trivial effect size de ned as r < +/-.05, Trim and ll = the number of estimated missing studies imputed. Note: 95% Con dence interval = the mean effect size with a 95% con dence interval, 95% Prediction interval = the true effect size in 95% of all comparable populations, p = probability value: the probability of giving these or more extreme results given that the null hypothesis is true, n = number of participants; k = number of studies, Fail-safe N = Orwin´s fail-safe N: the number of missing studies reporting no effect needed to result in a trivial effect size de ned as r < +/-.05, Trim and ll = the number of estimated missing studies imputed. Effectiveness. In essence, effectiveness measures people's beliefs that a policy can ful ll a speci c aim (6). As an overall measure, we found effectiveness to be the second strongest determinant for acceptance (r = .50, 95% CI [.41, .57], p < .001, k = 24). Two components of effectiveness were identi ed in the primary studies: (1) effectiveness in changing behaviors (11,16); and (2)  In sum, the most important aspect of fairness was not personal fairness (i.e., if a policy is "fair for me"), but whether the policy is perceived as overall fair and fairly distributed. Second, policies are more acceptable when perceived to effectively mitigate climate change than when perceived to effectively change people's behaviors. Climate change beliefs Knowledge. Making informed pro-environmental choices is di cult if one has incorrect or no knowledge (20). Although knowledge can be regarded as a necessary but not su cient precondition, past research con rms the importance of knowledge for environmental concern and pro-environmental behavior (21).
One crucial limitation of measuring knowledge is the incongruence between what people think is true and the actual evidence for the speci c issue (22). Consequently, past research has reported that objective knowledge is positively related to environmental issues such as climate change belief and environmental risk perceptions, while smaller or even negative relationships have been reported for subjective knowledge (15,23 Climate change belief, problem awareness, concern, and risk perception. We identi ed four types of climate change beliefs. First, in an overall measure, climate change belief assesses people's belief in anthropogenic climate change (15). Second, problem awareness assesses to what extent climate change is perceived as problematic (11). Third, environmental concern is a more general construct, measured as an evaluation or attitude toward the environment (24). Fourth, risk perception is a measure of personal or societal threats to people's well-being as a consequence of climate change (23).
When it comes to belief in climate change, denial has been reported in various cultures and is being fostered by doubt-mongering (25). It should, however, be noted that in 2020, 73% of Americans believed in climate change, while only 10% denied climate change (26). Our meta-analyses found a clear positive relationship between acceptance and climate change belief (r = .43, 95% CI [.29, .55], p = < .001, k = 11).
In sum, knowledge about climate change was weakly related to acceptance. Positive medium-sized relationships were found between acceptance and climate change beliefs, environmental concern, and perceived risk of and perceived problems associated with climate change.

Psychological factors
Trust. Trust can be de ned as a "psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another" (27). Public acceptance has been linked to both high and low levels of trust (28,29), suggesting effects of trust as well as mistrust.
Political a liation. Meta-analytic results showed that identifying as politically left or liberal (versus right or conservative) was positively associated with acceptance of climate change mitigation policies (r = .28, 95% CI [.20, .36], p < .001, k = 18).
Values. Values are de ned as desirable goals serving as guiding principles in people's lives (33,34).
Values have been shown to be antecedents for proximal determinants of pro-environmental engagement such as beliefs, personal norms (10,12), and pro-environmental identity (35). Most primary studies measured self-transcendent values as the more speci c biospheric values (describing a sense of connectedness with nature). Results showed a positive relationship with acceptance (r = .27, 95% CI [.21, .33], p < .001, k = 14). Results from self-enhancement values, measuring the extent to which people prioritise authority, social power, wealth, ambition, and in uence, showed a weak negative relationship with acceptance (r = -.09, 95% CI [-.14, -.04], p < .001, k = 10). These results suggest, somewhat counterintuitively, that self-enhancement values are not a strong barrier to accepting climate mitigation policies.
In sum, trusting institutions, identifying as political left or liberal, and valuing nature were all positively associated with accepting climate mitigation policies. Interestingly, self-enhancement values showed only a weak negative relationship with acceptance, indicating that holding egocentric values is not a strong barrier to accepting mitigation policies.
In sum, being educated, younger, and female was weakly associated with accepting climate mitigation policies. Income was not related to acceptance.

Discussion
In a meta-analysis, we analysed 15 determinants for accepting climate change mitigation policies, extracted from 76 datasets comprising a total of 146,817 participants from 34 countries. The results show that the demographic determinants age, gender, education, and income are weakly or not signi cantly related to accepting climate change mitigation policies, while data from political a liation showed a medium-sized relationship. Among all determinants, perceived fairness and effectiveness were strongly positively associated with acceptance, while the weakest relationships were found for knowledge and self-enhancement values.
Past studies have reported mixed ndings for determinants such as age, gender, income, and trust (6).
Here, we provide highly powered estimates, reporting a nonsigni cant and close to zero correlation for income, weak yet signi cant relationships for age and gender, and a medium-sized positive relationship between trust and climate change mitigation policy acceptance.
The meta-analysis did, however, report substantial dispertion for a number of determinants, calling for further analyses. Consequently, we conducted a number of subgroup analyses, examining how subcomponents of fairness, effectiveness, and concern, for example, were related to acceptance.
Studies have reported the in uence of a number of determinants not examined in the meta-analysis, such as infringing on personal freedom of choice (11,13), emotions (36,37), and pro-environmental identity (32,38). Our inclusion criteria were based on methodological considerations (14). Still, it should be noted that we did not analyse the full list of determinants for primary studies. We encourage future research to consider these determinants.
The present meta-analysis tests proximal determinants of policy acceptance. This is important to consider, as determinants showing weaker direct correlations might be indirectly related to acceptance. For example, studies have found that trust is indirectly linked to acceptance via risk perception (39), and that problem awareness is indirectly linked to via fairness and effectiveness (18).
It should also be noted that most of the included studies used data from the global north (40). There is an empirical gap in the multidisciplinary research eld of climate policy acceptance, and we still do not know whether, or to what extent, the results are valid for the global south. For example, the results may differ in contexts where the division between left and right is not the central political con ict line.

Conclusion
Based on 76 datasets from 34 countries generating a total sample of 146,817 participants, we report on a series of meta-analyses assessing the importance of 15 demographic and psychological determinants for accepting climate change mitigation policies. Policy-speci c beliefs were most important for acceptance. Speci cally, perceiving mitigation policies as overall fair or fairly distributed was strongly related to acceptance. For climate change beliefs, knowledge about climate change was weakly related to acceptance. Positive medium-sized relationships were found between acceptance and (a) beliefs in climate change, (b) environmental concern, (c) perceived risks, and (d) climate change problem awareness. Regarding psychological factors, trusting institutions, identifying as political left or liberal, and valuing nature were all positively associated with accepting climate mitigation policies. Selfenhancement values were only weakly negatively related to acceptance. Finally, weak associations were found for the demographic variables education, age, and gender. Income was not related to acceptance. At a general level, the variables assessing the policy measure show higher correlations with acceptance than variables unrelated to the policy. These variables are often referred to as policy-speci c beliefs and evaluate the attributes of the policy itself, such as whether it can be perceived as fair or e cient. These results inform scholars of climate policy acceptance as to which determinants to include in future analyses, and they also advise policymakers as to which sentiments they should consider when introducing and communicating intended climate policies. Figure 1 Flow chart for inclusion/exclusion of studies

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. AppendixA.pdf