Effects of system-sanctioned framing on climate awareness and environmental action in the United States and beyond

Significance

The study investigates the impact of an experimental intervention that frames climate action as patriotic and system-sanctioned on environmental attitudes and behaviors in 63 countries. In the United States, this intervention significantly increased liberal-leftists’ and conservative-rightists’ belief in climate change; support for proenvironmental policies; and willingness to share information about climate change on social media. Tailored messages were also effective in an aggregated analysis involving 63 countries. More granular analyses at the country level revealed that the intervention was successful in some non-U.S. countries (e.g., Brazil, France, Israel), but it backfired in others (Germany, Belgium, Russia). Using system justification theory to design communication strategies may improve climate awareness and action in the general public.

Abstract

Despite growing scientific alarm about anthropogenic climate change, the world is not on track to solve the crisis. Inaction may be partially explained by skepticism about climate change and resistance to proenvironmental policies from people who are motivated to maintain the status quo (i.e., conservative-rightists). Therefore, practical interventions are needed. In the present research program, we tested an experimental manipulation derived from system justification theory in which proenvironmental initiatives were framed as patriotic and necessary to maintain the American “way of life.” In a large, nationally representative U.S. sample, we found that the system-sanctioned change intervention successfully increased liberal-leftists’ as well as conservative-rightists’ belief in climate change; support for proenvironmental policies; and willingness to share climate information on social media. Similar messages were effective in an aggregated analysis involving 63 countries, although the overall effect sizes were small. More granular exploratory analyses at the country level revealed that while the intervention was moderately successful in some countries (e.g., Brazil, France, Israel), it backfired in others (Germany, Belgium, Russia). Across the three outcome variables, the effects of the intervention were consistent and pronounced in the United States, in support of the hypothesis that system justification motivation can be harnessed on behalf of social change. Potential explanations for divergent country-level effects are discussed. The system-sanctioned change intervention holds considerable promise for policymakers and communicators seeking to increase climate awareness and action.

Although the scientific community continues to warn that anthropogenic climate change poses a major threat to the global order (1), the world is not on track to solve the problem (2). It is clear that numerous institutional and behavioral factors contribute to climate inaction (3–6). In the present research, we focus on the ideological motivation to defend, bolster, and justify the societal status quo (7), which helps to explain conservative-rightist skepticism about climate change and reluctance to support proenvironmental policies (8–10). In a large, nationally representative sample of Americans and smaller convenience samples from 62 other countries (11, 12), we tested an experimental intervention derived from system justification theory to overcome and harness system-defensive motives that otherwise undermine climate awareness and action.

Research on public opinion has identified a substantial ideological divide concerning environmental beliefs and attitudes, such that conservative-rightists report more skepticism about anthropogenic climate change and opposition to proenvironmental policies, compared to liberal-leftists (4, 13–17). Ideological cleavages have been observed for well over a decade in the United States, but are increasingly common around the world (5, 18). Some scholars emphasize the role of elite-driven political communication in climate polarization, whereas others focus on politically motivated reasoning processes among ordinary citizens (19–23). Most likely both factors are important, insofar as supply-side and demand-side dynamics operate reciprocally in the market for belief systems (24).

Most theories of motivated reasoning focus on self-serving and group-serving needs to confirm one’s own prior beliefs and protect one’s partisan identity (22, 25). According to system justification theory, people are also motivated—to varying degrees, depending on situational and dispositional factors—to defend and maintain the societal status quo, whether they are primary beneficiaries or not (7). This system-serving motivation, in turn, leads people to overlook or downplay the shortcomings of the social, economic, and political systems on which they depend and to resist or oppose transformative social change (26–30).

Throughout Europe and the Americas, conservative-rightists typically (but not always) exhibit stronger system justification motivation than liberal-leftists (7). However, from the standpoint of system justification theory, even the latter are at times motivated to explain and justify the societal status quo in ways that resolve uncertainty, threat, and social discord (31–35). Thus, a system justification perspective, which suggests that most people want to reach desirable conclusions about the overarching social system, differs from a “worldview defense” perspective, which suggests that most people want to maintain their own belief system—whether it is critical or approving of the social system (36).

Consistent with system justification theory, an experimental study conducted with predominantly liberal college students revealed that making students feel dependent upon the economic system led them to misremember scientific information about climate change in ways that downplayed the threat, even though they were presumably high in climate awareness (9). Thus, increasing system justification motivation appeared to bias recall, sustaining skepticism even in a very liberal sample.

There are other studies that highlight the persuasive benefits of framing environmental protection as patriotic (37, 38). One experiment, which inspired the present research program, demonstrated that framing the need for environmental action as patriotic and congruent with the preservation of the status quo made high system-justifying individuals more willing to sign a petition on behalf of the environment, compared to a control condition (8). Although this was a promising result, one methodological critic aptly noted that a small sample of undergraduate students “will not support valid inferences about ‘message frames’ likely to offset politically motivated reasoning among individuals who are skeptical of climate change” (22).

In the present research program, we investigated—in several large, cross-national samples previously collected as part of an international collaboration (11, 12)—whether it is possible to devise an effective messaging campaign based on system justification theory. System justification motivation often leads people to regard the status quo favorably and to resist social change. Generally speaking, this leads people to downplay the problem of climate change and avoid proenvironmental action (7–9). However, based on past work, we hypothesized that framing climate action as congruent with the preservation of the status quo would disrupt this pattern by reducing the sense that such change is system-challenging. In particular, we tested the idea that making people feel more dependent upon the overarching social system and communicating that proenvironmental action is patriotic and necessary to preserve longstanding traditions and the country’s “way of life” would strengthen belief in climate change and support for proenvironmental action. To this end, we created a longer and more psychologically involving set of experimental stimuli, which included country-specific content, than has been attempted in previous studies of this kind (8, 39).

We made use of an extensive cross-national dataset (11, 12) in which participants were randomly assigned to either a control condition that made no mention of climate change, or to one of 11 experimental interventions. One of these interventions was the system-sanctioned change intervention, which framed the need to protect the natural environment as patriotic and system-serving. For the purposes of this article, we will focus in detail on the results of the system justification intervention. We gauged belief in climate change, support for environmental policies, and willingness to share information about climate change on social media. In light of previous research based on a much smaller convenience sample (8), we hypothesized that the intervention would be effective among conservative-rightists, who are more likely to endorse system-justifying values and more skeptical about climate change, without triggering backlash among liberal-leftists. We first examined the effects of the intervention in the United States because this was the context in which the general paradigm was initially developed. Next, we explored the effectiveness of the intervention in aggregated analyses involving the United States and 62 additional countries (11, 12). Finally, we conducted more focused comparisons, which also considered the political orientations of respondents in 17 countries for which samples were large enough to permit well-powered hypothesis testing.^*

Results in the United States

We first examined the effect of the system justification intervention in a large, nationally representative sample of U.S. adults. This sample consisted of 1,393 participants (mean age = 46.15 y, SD = 16.20; 50% female, 49% male, 1% nonbinary/other). Of these participants, 669 were assigned to the control condition, and 724 were assigned to the intervention condition.

Belief in Climate Change.

We estimated a linear mixed model with belief in climate change as the dependent variable, political ideology, and experimental condition as fixed effects, and by-item (four belief items), by-participant, random effects. As expected, the analysis yielded a significant main effect of ideology, such that conservatism was negatively associated with belief in climate change (b = −0.42, SE = 0.02, β = −0.35, t = −26.03, P < 0.001, d = −0.74). We also observed a significant main effect of experimental condition (b = 5.82, SE = 0.85, β = 0.18, t = 6.85, P < 0.001, d = 0.37). As hypothesized, participants assigned to the system-sanctioned framing condition expressed greater belief in climate change (or less skepticism) than those assigned to the control condition. The interaction between ideology and condition was nonsignificant, indicating that the effect of the system-sanctioned frame was unaffected by political orientation (b = −0.02, SE = 0.03, β = −0.02, t = −0.59, P = 0.558, d = −0.03; Fig. 1A).

Fig. 1.

Effect of political ideology and experimental condition on belief in climate change (Panel A), support for proenvironmental policies (Panel B), and social media sharing intentions (Panel C) among U.S. participants (n = 1,393). Higher scores on political ideology represent more conservative/right-wing. Error bars denote 95% CIs. P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***).

Support for Environmental Policies.

We estimated a linear mixed model with support for proenvironmental policies as the dependent variable, ideology and condition as fixed effects, and by-item (nine policy items), by-participant, random effects. The analysis again yielded a main effect of ideology, such that conservatism was negatively associated with support for environmental policies (b = −0.28, SE = 0.01, β = −0.23, t = −26.42, P < 0.001, d = −0.48), and a main effect of experimental condition (b = 3.98, SE = 0.57, β = 0.12, t = 7.04, P < 0.001, d = 0.25). As hypothesized, participants assigned to the system justification intervention condition were more supportive of proenvironmental policies than those assigned to the control condition. The interaction between ideology and condition was again nonsignificant, indicating that the effect of the intervention was unaffected by political orientation (b = −0.03, SE = 0.02, β = −0.02, t = −1.39, P = 0.164, d = −0.05; Fig. 1B).

Willingness to Share Climate Change Information on Social Media.

We estimated a binary logistic regression, with willingness to share climate change information on social media as the dependent variable and ideology and condition as fixed effects. No significant effect of ideology was observed in this case (β = −0.0005, SE = 0.0007, z = −0.73, P = 0.463, d = −0.001). The analysis did, however, yield a significant main effect of condition, such that participants assigned to the system-sanctioned framing condition were indeed more willing to share information on social media than those assigned to the control condition (β = 0.13, SE = 0.15, z = 3.21, P = 0.001, d = 0.26). The two-way interaction was nonsignificant, indicating that the effect of the intervention was unaffected by political orientation (β = 0.00, SE = 0.01, z = 0.26, P = 0.797, d = 0.00; Fig. 1C).

Discussion

U.S. adults assigned to the system-sanctioned framing condition expressed more belief in climate change, support for proenvironmental policies, and willingness to share climate change information on social media, compared to those assigned to the control condition (Fig. 1). Moreover, the intervention was similarly effective for liberals and conservatives, even though conservatives were more skeptical about climate change and environmental policies overall.

A previous experiment, which involved a much smaller convenience sample of U.S. college students, found that system-sanctioned framing increased support for climate action among high system-justifiers (who are relatively conservative) but not low system-justifiers (who are relatively liberal) (8). By contrast, the new intervention, which was designed to be much more psychologically engaging, was found to be effective in increasing climate awareness and environmental action regardless of political orientation. Next, we considered whether this new system-sanctioned framing intervention would also be effective in other countries, although it was originally designed to operate in the U.S. context.

Results of an Aggregated Analysis Involving 63 Countries

Modified versions of the system-sanctioned framing manipulation were administered to adult samples in 62 countries in addition to the United States (n = 59,440) (11, 12). To examine the effectiveness of the manipulation in the worldwide sample, we treated country as a random effect variable in all analyses.

Belief in Climate Change.

We estimated a linear mixed model with belief in climate change as the dependent variable, ideology and experimental condition as fixed effects, and by-country, by-item (four belief items), by-participant, random effects. The analysis yielded a significant effect of ideology, such that right-wing conservatism was negatively associated with belief in climate change in the aggregate analysis involving 63 countries (b = −0.19, SE = 0.01, β = −0.16, t = −29.10, P < 0.001, d = −0.33). The analysis also yielded a significant main effect of condition (b = 1.74, SE = 0.29, β = 0.07, t = 6.11, P < 0.001, d = 0.13). As hypothesized, participants assigned to the system-sanctioned framing condition expressed greater belief in climate change than those assigned to the control condition. No significant interaction between ideology and condition was observed, indicating that the effect of the intervention was unaffected by political orientation (b = −0.01, SE = 0.01, β = −0.01, t = −1.11, P = 0.266, d = −0.02; Fig. 2A).

Fig. 2.

Effects of political ideology and experimental condition on belief in climate change (Panel A), support for proenvironmental policies (Panel B), and social media sharing intentions (Panel C) aggregated across 63 countries (n = 59,440). Error bars denote 95% CIs.

Support for Proenvironmental Policies.

We estimated a linear mixed model with support for proenvironmental policies as the dependent variable, ideology and condition as fixed effects, and by-country, by-item (nine policy items), by-participant, random effects. The analysis again yielded a significant effect of ideology, such that right-wing conservatism was negatively associated with support for proenvironmental policies overall (b = −0.12, SE = 0.00, β = −0.10, t = −26.79, P < 0.001, d = −0.19) as well as a significant effect of condition (b = 1.03, SE = 0.21, β = 0.03, t = 5.01, P < 0.001, d = 0.07). Although the effect size was small, participants assigned to the system-sanctioned framing condition were significantly more supportive of proenvironmental policies than those assigned to the control condition. No significant interaction between ideology and condition was observed (b = −0.02, SE = 0.01, β = −0.01, t = −1.89, P = 0.059, d = −0.03; Fig. 2B).

Willingness to Share Climate Change Information on Social Media.

We estimated a mixed binary logistic regression with willingness to share climate information on social media as the dependent variable, ideology and condition as fixed effects, and by-country random effects. The analysis yielded a very small but statistically significant effect of ideology (β = 0.002, SE = 0.00, z = 4.84, P < 0.001, d = 0.003) such that right-wing conservatism was associated with greater intentions to share climate information on social media. This was an unanticipated result that could be attributable to ideological differences in social media access and/or usage around the world. The analysis also yielded a significant effect of condition, such that participants assigned to the system-sanctioned framing condition reported more willingness to share climate information than participants assigned to the control condition (β = 0.07, SE = 0.06, z = 4.44, P < 0.001, d = 0.14). No significant interaction between ideology and condition was observed (β = 0.00, SE = 0.00, z = −0.49, P = 0.627, d = 0.00; Fig. 2C).

Discussion

Although our original intervention was tailored to the U.S. context, an aggregated analysis of data from 63 countries revealed that it was also successful on a global scale. Participants from 63 countries assigned to the system-sanctioned framing intervention condition expressed more belief in climate change, support for proenvironmental policies, and willingness to share climate change information on social media, compared to participants who were assigned to the control condition (Fig. 2). Moreover, the intervention was similarly effective at reaching leftists and rightists, even though rightists were more skeptical about climate change and environmental policies in general, and leftists were more reticent about sharing climate information on social media. In a final set of analyses, we conducted a more focused set of comparisons to identify those countries in which the intervention was more vs. less effective. This also enabled us to explore the role of ideology in countries that differed considerably in terms of sociopolitical contexts.

Results of Cross-Country Comparisons Involving 17 Countries

Consistent with prior research (40), we included countries with at least 180 participants, which provided adequate power to detect at least moderate effect sizes (Table 1).^† The final analytic sample for cross-country comparisons consisted of 5,698 participants from 17 countries (mean age = 41.72 y, SD = 16.20; 52% female, 47% male, less than 1% nonbinary/other). Of these participants, 2,772 were assigned to the control condition, and 2,926 were assigned to the intervention condition.

Table 1.

The sample size for each country and corresponding statistical power

Country	N	Power
Belgium	180	0.80
Brazil	211	0.86
Canada	196	0.83
Chile	239	0.97
France	255	0.91
Germany	263	0.92
Israel	223	0.88
Italy	271	0.93
Japan	253	0.91
Netherlands	320	0.96
Nigeria	266	0.93
Poland	407	0.98
Russia	240	0.90
Sweden	387	0.99
Switzerland	186	0.81
The United Kingdom	318	0.96
The United States	1,393	1.00

Power analyses were conducted using a t test power calculator. Based on each country’s sample size, we examined the power to detect a moderate-to-large Cohen’s d effect size for the difference between the control condition and the system-sanctioned framing condition.

Belief in Climate Change.

We estimated a linear mixed model with belief in climate change as the dependent variable, ideology, condition, and country as fixed effects, and by-item (four belief items), by-participant, random effects. The analysis yielded a significant main effect of ideology, such that right-wing conservatism was negatively associated with belief in climate change, collapsing across all 17 countries (b = −0.22, SE = 0.01, β = −0.19, t = −18.85, P < 0.001, d = −0.38). However, in this analysis, we did not observe an overall main effect of condition (b = 0.37, SE = 0.49, β = 0.01, t = 0.75, P = 0.451, d = 0.03).^‡

To investigate whether the effect of condition was contingent upon different levels of political orientation in various countries, we inspected the three-way interaction involving ideology, condition, and country on belief in climate change. Specifically, we performed a type III sums of squares test on the fitted mixed-effects model to obtain omnibus effects for fixed variables and their interactions (Table 2A). A significant three-way interaction was observed, F(16, 17,304) = 1.85, P = 0.020, prompting us to decompose the interaction.

Table 2.

Omnibus tests for all three outcome measures

A. Belief in climate change
Source	Type III sum of squares	dfNumerator	dfDenominator	Mean Square	F	P
Condition	369.89	1	17,304	369.89	0.57	0.451
Ideology	231,651.05	1	17,304	231,165.05	335.45	<0.001
Country	401,055.15	16	17,304	25,065.95	38.46	<0.001
Sex	80,958.36	1	17,304	80,958.36	124.22	<0.001
Age	21,734.81	1	17,304	21,734.81	33.35	<0.001
Education	9,169.45	1	17,304	9,169.45	14.07	<0.001
Income	18,718.85	1	17,304	18,718.85	28.72	<0.001
Condition*Ideology	1,523.76	1	17,304	1,523.76	2.34	0.126
Condition*Country	59,542.56	16	17,304	3,721.41	5.71	<0.001
Ideology*Country	228,997.40	16	17,304	14,312.34	21.96	<0.001
Condition*Ideology*Country	19,323.71	16	17,304	1,207.73	1.85	0.020
B. Support for proenvironmental policies
Source	Type III sum of squares	dfNumerator	dfDenominator	Mean Square	F	P
Condition	255.89	1	38,076	255.89	0.35	0.552
Ideology	332,111.93	1	38,076	332,111.93	458.00	<0.001
Country	481,408.99	16	38,076	30,088.06	41.49	<0.001
Sex	16,567.77	1	38,076	16,567.77	22.85	<0.001
Age	3,494.80	1	38,076	3,494.80	4.82	0.028
Education	65,878.31	1	38,076	65,878.31	90.85	<0.001
Income	28,837.78	1	38,076	28,837.78	39.77	<0.001
Condition*Ideology	2,743.90	1	38,076	2,743.90	3.78	0.052
Condition*Country	64,481.77	16	38,076	4,030.11	5.56	<0.001
Ideology*Country	289,464.17	16	38,076	18,091.51	24.95	<0.001
Condition*Ideology*Country	20,855.80	16	38,076	1,303.49	1.80	0.026
C. Social media sharing intentions
Source	df		Wald Chi-Square		P
Condition	1		6.29		0.012
Ideology	1		6.55		0.011
Country	16		328.79		<0.001
Sex	1		14.21		<0.001
Age	1		35.11		<0.001
Education	1		12.58		<0.001
Income	1		0.57		0.450
Condition*Ideology	1		0.33		0.564
Condition*Country	16		13.07		0.668
Ideology*Country	16		36.05		0.003
Condition*Ideology*Country	16		10.97		0.812

Type III tests of fixed effects for belief in climate change are reported in Panel A, and those for support for proenvironmental policies are reported in Panel B. A Wald Chi-Square test of fixed effects for social media sharing intentions is reported in Panel C.

We observed that the United States was the only country in which the intervention significantly increased belief in climate change for liberals and conservatives alike. At the same time, the manipulation was successful in increasing belief in climate change among conservative-rightists in the Netherlands and Israel, and among liberal-leftists in Chile. However, the manipulation backfired in Germany and Russia, where conservative-rightists assigned to the system-sanctioned condition expressed less belief in climate change compared to conservative-rightists in the control condition (Fig. 3A).

Fig. 3.

Standardized coefficient estimates and 95% CIs of the effects of the intervention on belief in climate change (Panel A), and support for proenvironmental policies (Panel B) by political ideology and country (n = 5,698).

Support for Proenvironmental Policies.

To examine the effects of political ideology, experimental condition, and country, we estimated a linear mixed model with support for proenvironmental policies as the dependent variable, ideology, condition, and country as fixed effects, and by-item (nine policy items), by-participant, random effects. The analysis yielded a significant effect of ideology, such that right-wing conservatism was negatively associated with support for proenvironmental policies, collapsing across all 17 countries (b = −0.17, SE = 0.01, β = −0.14, t = −21.40, P < 0.001, d = −0.28). However, no significant effect of condition emerged, b = 0.21, SE = 0.35, β = 0.01, t = 0.59, P = 0.553, d = 0.01).

To investigate whether the effect of condition was contingent upon different levels of political orientation in various countries, we inspected the three-way interaction involving ideology, condition, and country. Again, we performed a type III sums of squares test on the fitted mixed-effects model to obtain omnibus effects for fixed variables and their interactions (Table 2B). A significant three-way interaction was again observed, F(16, 38,076) = 1.80, P = 0.026, prompting us to decompose the interaction.

The United States and Brazil were the only countries in which the intervention significantly improved support for proenvironmental policies for both liberal-leftists and conservative-rightists. The intervention was also successful among rightists in Israel and leftists in France. It backfired in Germany and Belgium, where rightists assigned to the system-sanctioned condition expressed less support for proenvironmental policies compared to rightists assigned to the control condition (Fig. 3B).

Willingness to Share Climate Change Information on Social Media.

To examine the effects of political ideology, experimental condition, and country, we estimated a binary logistic regression with willingness to share climate change information on social media as the dependent variable and ideology, condition, and country as fixed effects. The analysis yielded a significant but very small main effect of ideology (β = 0.002, SE = 0.00, z = 2.56, P = 0.011, d = 0.004), such that right-wing conservatism was associated with stronger intentions to share information on social media. As noted above, this result may be attributable to ideological differences in social media access and/or usage around the world. The analysis also yielded a significant main effect of condition, such that participants assigned to the system-sanctioned framing condition expressed more willingness to share climate change information on social media than those assigned to the control condition (β = 0.07, SE = 0.11, z = 2.51, P = 0.012, d = 0.15).^§

To investigate the possibility that the effect of condition was contingent upon different levels of political orientation in various countries, we inspected the three-way interaction involving ideology, condition, and country. Specifically, we performed a Wald Chi-Square test on the fitted binary logistic regression model to obtain omnibus effects for fixed variables and their interactions (Table 2C). No three-way interaction was observed.

Despite the lack of a significant interaction between condition and country, for the sake of consistency and completeness we explored the effects of the condition on sharing intentions within each country, plotting condition estimates from the binary logistic regression for each country. The intervention was successful in increasing sharing intentions only in the United States and Brazil (Fig. 4).

Fig. 4.

Standardized coefficient estimates and 95% CIs of the effects of the intervention on social media sharing intentions by country (n = 5,698).

General Discussion

Climate inaction is attributable, in part, to climate skepticism and resistance to proenvironmental policies, especially on the political right (4, 13, 16, 17). Aggregating across respondents from 63 countries, we observed a strong negative association between right-wing conservatism, on the one hand, and belief in climate change and support for proenvironmental policies, on the other. These results support the notion that a global expansion of left–right climate polarization is underway (5, 16, 18, 41, 42).

Ideological differences in climate skepticism and resistance to action may be explained, in part, by stronger system justification motives, that is preferences for maintaining the societal status quo, among conservative-rightists compared to liberal-leftists (7). Moreover, acknowledging the problem of anthropogenic climate change implicates the entire economic system of industrial capitalism and threatens conservative values such as support for the free market and opposition to governmental regulation (7–9, 43, 44). Although system justification motivation often leads people to downplay social problems and resist efforts to address them, previous studies hinted that the system justification motivation can also be harnessed in the service of social change (8, 45, 46).

The present research program represents the most ambitious and comprehensive attempt thus far to apply the technique of system-sanctioned framing as a way of overcoming resistance to social change in 63 countries. In support of hypotheses derived from system justification theory, we observed both in the United States and the aggregated sample that participants who were exposed to the system-sanctioned message frame expressed more belief in climate change, more support for proenvironmental policies, and stronger intentions to share climate information on social media, compared to participants assigned to the control condition. Importantly, the intervention was effective for liberal-leftists as well as conservative-rightists in the United States. This suggests that patriotic, system-supporting language appealed to conservative-rightists, as expected, without triggering reactance among liberal-leftists (8).

In country-specific analyses across 17 samples, we observed that results varied by region and ideology. The intervention was generally quite effective for liberal-leftists and conservative-rightists in the United States and Brazil. Because both countries are highly polarized along left–right lines, the fact that the system-justifying appeal was well received across the ideological spectrum was especially encouraging. We also observed that the intervention was at least partially effective among liberal-leftists in France and Chile and among conservative-rightists in Israel and the Netherlands. However, it appeared to backfire for conservative-rightists in Germany, Belgium, and Russia.

Although any attempt to explain country-specific patterns is somewhat speculative, we raise a few possibilities here. It may be pertinent that in France—unlike all other European countries examined thus far—liberal-leftists tend to score higher than conservative-rightists on general system justification (47). This could explain why the system-sanctioned framing manipulation was more impactful for leftists than rightists in the French context.

It is also conceivable that people are more receptive to system-justifying messages when their preferred political party is in power. For example, liberal-leftists might be more open to system-sanctioned framing when the government is left-leaning (vs. right-leaning), and vice versa for conservative-rightists. Since 2021, the government of Chile has been led by a left-wing coalition (“Approve Dignity”); this could explain why leftists were more responsive to system-justifying messages in the Chilean context. By contrast, the governments in Israel and the Netherlands have been led by rightists since 2009 and 2021, respectively. These facts could help to explain why rightists would be responsive to the call to maintain the status quo in each of these countries. However, these interpretations should be treated cautiously, because our formal analysis of the role of the party-in-power failed to produce consistent conclusions that applied to all 17 countries under investigation (SI Appendix, Tables S28–S30).

The situation in Germany is somewhat more complicated. For one thing, leftists in Germany are suspicious of patriotic language because of its association with the Holocaust; this may help to explain why the system-sanctioned frame was ineffective for them (47). It may also be pertinent that the German government was led from 2005 to 2021 by a center-right Chancellor, Angela Merkel, who expressed real concern about the problem of climate change. According to a public opinion survey conducted in 2021 by the European Investment Bank, 77% of Germans believe that climate change is the most serious problem facing humanity. However, far-rightists are much less likely than other respondents to agree with this assessment, with one-third denying that climate change is attributable to human activity. Whereas 69% of leftists in Germany are highly supportive of renewable energy initiatives, only 42% of rightists are supportive (48). Climate skepticism is also highly pronounced among far-right supporters of the Vlaams Belang Party in Belgium (49). Thus, baseline ideological differences could help to explain why our manipulation appeared to trigger backlash among rightists (but not leftists) in Germany and Belgium.

According to a recent analysis of public opinion, climate attitudes are not clearly polarized along left–right ideological lines in Russia (50). However, concern about climate change is generally lower in Russia than in other European countries, and climate skepticism appears to be driven, at least in part, by investment in the fossil fuel industry and the desire to preserve Russia’s status as a “Great Power” under Vladimir Putin (51), who has increasingly appealed to the authoritarian right (52). Thus, the current political situation in Russia may help to explain why the system-sanctioned intervention backfired among rightists. Future research based on nationally representative samples is needed to assess the validity of all of the above interpretations.

In any case, there are probably several reasons why the experimental intervention was more effective in the United States than in other countries. First, the system dependence and system-sanctioned change manipulations were originally developed and applied to environmental attitudes in the American context (8, 39). Second, we observed that U.S. participants were more skeptical about climate change and resistant to proenvironmental initiatives than participants from other countries. These differences mirror the results of public opinion surveys based on nationally representative samples (4, 16, 53). Thus, it is possible that, compared to other countries, U.S. participants had more room to improve in terms of climate awareness and action. Last, it is well known that politicians and organizations in the United States regularly promote misconceptions about the supposed incompatibility of environmental protection and economic vitality (54–57), exacerbating concerns about environmental-economic tradeoffs in the American public (17, 41, 58). In such a context, it may be especially important to overcome economic system justification motivation and harness it on behalf of environmental conservation (39).

We obtained the largest treatment effect sizes for the system-sanctioned change manipulation in the United States (with d’s ranging from 0.25 to 0.37), where the original manipulation was developed (8). According to current statistical conventions, these effects would be classified as “small” to “medium” in size (58). In other countries, the effect sizes tended to be smaller than in the United States. However, it is important to bear in mind that small effects are not psychologically insignificant, because they often exert highly meaningful cumulative effects over time (58). For example, framing effects in the communication literature are often quite small in the short term, but they can have major downstream consequences if they are repeated and adopted by communicators with large audiences (59). Even a short, 5-min intervention that produces small effect sizes holds great promise if it is adopted on a large scale by political and media elites. Future research should investigate the extent to which the effects of the intervention dissipate or diminish over time and whether repeated exposure to system-sanctioned messaging can enhance its effectiveness in promoting proenvironmental behavior.

Despite the overall success of the system-sanctioned frame intervention, there are several limitations that future work should address. For one thing, the control condition did not mention climate change, so it is possible that simply emphasizing the problem would affect climate awareness and action. However, a global megastudy comparing the effects of 11 different interventions, all of which referred explicitly to climate change, shows that this is not the case. Moreover, this study confirmed that the system-sanctioned change intervention was more effective in boosting proenvironmental outcomes than several other interventions that also mentioned climate change (11, 12).

In the 63-country analysis, for example, the system justification intervention significantly outperformed four others in boosting belief in climate change, and it was outperformed by only one. In the U.S. analysis, we found that the system justification intervention significantly outperformed two other interventions and did not differ from the top-performing intervention. With respect to support for proenvironmental policies, the system justification intervention outperformed two others but was outperformed by four others in the 63-country analysis. In the U.S. analysis, it performed significantly better than the control condition and no worse than any other intervention except for the top performer. In summary, then, the system-sanctioned change intervention performed well in comparison with 10 other experimental interventions derived from other psychological theories.^¶

Another limitation of the present research program is that we are unable to pinpoint which specific aspects of the message were most responsible for the improvement in climate awareness and action. The system-sanctioned-change intervention tested here was inspired by a previous experiment designed to harness system justification motivation in the service of constructive social change. However, the original intervention was sparse, relying on a single paragraph emphasizing the patriotic nature of climate action (8). To make it more psychologically vivid and engaging, this earlier version was embellished with photos and language tapping into other aspects of system justification theory. Specifically, the passage we administered a) primed feelings of dependence on the system, b) linked the present “way of life” to longstanding cultural traditions, and c) highlighted potential risks of climate change and conveyed that climate action is patriotic and necessary for maintaining the societal status quo. Future work would do well to determine which specific elements of this overall message are most important or effective, or whether a combination of them is necessary for improving environmental attitudes and behavior.

We recognize that numerous institutional and other barriers, including organized resistance, can hinder the effectiveness of climate change framing interventions in real-world contexts. To maximize their effectiveness, shifting narratives should be accompanied by institutional changes that address political polarization, counterframes, and climate misinformation (60). Because message sources are critical when it comes to social influence (61), the most effective campaign would feature individuals who are widely trusted across the ideological spectrum, if such individuals can be found.

Concluding Remarks

This research program has implications for messaging strategies aiming at improving climate awareness and action. From the perspective of system justification theory, it is crucial to avoid triggering epistemic and existential threats, insofar as these can motivate system-defensiveness and backlash against proenvironmental initiatives (7). A better strategy, according to our research, would be to emphasize how acknowledging and addressing climate change serves patriotic, system-preserving functions, allowing people to celebrate and perpetuate—rather than relinquish and replace—familiar ways of life and cherished cultural traditions.

Materials and Methods

Participants.

Data were collected as part of a global mega-study that assessed the effectiveness of 11 different experimental interventions that had been found previously to improve proenvironmental outcomes (11, 12). In collaboration with researchers from dozens of countries, participants were identified through snowball sampling methods and recruited through online data collection platforms (e.g., Bovitz in the United States and Market Research Institute in most other countries). Whenever possible, we obtained nationally representative samples. The U.S. and Israeli samples were the most representative, matching the national census in terms of age, gender, region, and ethnicity. In cases where representative sampling was not feasible, we relied on convenience samples, which have been found to be adequate for estimating treatment effects (62).

The full dataset was composed of 80,573 individuals who participated between July 2022 and May 2023. After excluding participants who failed at least one of two attention checks or failed to complete a tree-planting task successfully, 59,440 participants from 63 countries remained (11, 12). Overall, there was an attrition rate of 36% from the initial sample to the final sample. This falls within the typical range for online studies, which is between 30% and 50% (63). Participants were randomly assigned to either the control condition or one of 11 experimental interventions. In the research reported here, we focused on those participants assigned to the control condition (n = 5,086) and the system-sanctioned change condition (n = 5,173). The final analytic sample was 52% female, 47% male, with less than 1% nonbinary/other (mean age = 39.14 y, SD = 15.78).

Procedure.

Participants were informed they would be asked to read some information, report on their beliefs and behaviors, and write a small paragraph. To ensure a common understanding of climate change, they were initially presented with the following definition: “Climate change is the phenomenon describing the fact that the world’s average temperature has been increasing over the past 150 y and will likely be increasing more in the future.” Participants were then randomly assigned to either the control condition or to the experimental condition and asked to answer three sets of questions (administered in random order) pertaining to belief in climate change, support for proenvironmental policies, and willingness to share climate information on social media.^#

Experimental Conditions.

Participants assigned to the control condition read a literary passage from Charles Dickens’ (1861) novel Great Expectations, which made no references to the climate (11). In the system-sanctioned framing condition, participants instead read a passage that was intended, first, to increase feelings of dependence upon the social system (39), second, to link the nation’s “way of life” to natural and cultural traditions, and third, to communicate that climate change poses a threat that requires patriotic, proenvironmental action to “protect and preserve” the national heritage “as it should be” (8). The passage was accompanied by photographs of cities, natural resources, people, and flags.

The initial system-sanctioned message, which focused on the Slovakian context, was written by Ondrej Buchel, Michael Tyrala, and Andrej Findor, who nominated system justification theory for inclusion in the many-labs collaborative tournament. These researchers also drafted the U.S. version of the manipulation. It was reviewed and edited by the many-labs organizational team and also by John T. Jost, an expert in system justification theory and the use of system-sanctioned change messages (see SI Appendix for the final version). Modified, country-specific versions of the passage and accompanying images were administered to participants in 63 countries in their native language. Prior to data collection, the mega-study was reviewed by all collaborators and piloted to ensure the Qualtrics survey was correctly programmed and that all interventions took less than 5 min to complete (12).

Measures.

Belief in climate change.

To measure belief in climate change, participants rated how accurate each of the following statements were on a scale ranging from 0 (“Not at all accurate”) to 100 (“Extremely accurate)”: 1) “Taking action to fight climate change is necessary to avoid a global catastrophe,” 2) “Human activities are causing climate change,” 3) “Climate change poses a serious threat to humanity,” and 4) “Climate change is a global emergency” (α = 0.93; M = 78.67, SD = 24.04).

Support for proenvironmental policies.

To measure support for proenvironmental policies, participants were asked how which they agreed with each of the following statements on a scale ranging from 0 (“Not at all”) to 100 (“Very much so”): “I support…” 1) “raising carbon taxes on gas/fossil fuels/coal,” 2) “significantly expanding infrastructure for public transportation,” 3) “increasing the number of charging stations for electric vehicles,” 4) “increasing the use of sustainable energy such as wind and solar energy,” 5) “increasing taxes on airline companies to offset carbon emissions,” 6) “protecting forested and land areas,” 7) “investing more in green jobs and businesses,” 8) “increasing taxes on carbon-intensive foods (for example, meat and dairy),” and 9) “introducing laws to keep waterways and oceans clean” (α = 0.87; M = 69.24, SD = 19.38).

Information sharing intentions.

To measure willingness to share information about climate change on social media, participants were first asked, “Did you know that removing meat and dairy for only two out of three meals per day could decrease food-related carbon emissions by 60%? It is an easy way to fight #climatechange.” Next, they were asked whether they would be willing to share this information on social media (Yes/No). Participants who reported abstaining from social media (25.9% of the analytic sample) were excluded from analyses pertaining to this variable.

Demographics.

Last, participants answered sociodemographic questions about sex, age, education, and income. They were also asked to answer two questions about their political orientation on a scale from 0 (“Extremely liberal/Left-wing”) to 100 (“Extremely conservative/Right-wing”): 1) “What is your political orientation for economic issues (e.g., taxes),” and 2) “What is your political orientation for social issues (e.g., health care, education, etc.).” Because responses to these items were highly intercorrelated (r = 0.71, P < 0.001), we averaged them to estimate participants’ political ideology, as in previous research (64, 65).

Materials, Data, and Code.

All study materials and measures can be found in SI Appendix, Part I: Study Materials, and all data and code can be accessed via GitHub: https://github.com/katherineannemason/System-Sanctioned-Change. For the sake of transparency, the authors note that results pertaining to the general effectiveness of the 11 interventions included in the megastudy (including the system-sanctioned change intervention) were previously reported by Vlasceanu et al. (11) and Berkebile et al. (42). However, our work includes several analyses and interpretations that were not reported or discussed in any previously published articles.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Anonymized data have been deposited in GitHub (https://github.com/katherineannemason/System-Sanctioned-Change). Previously published data were used for this work (11, 12). All other data are included in the manuscript and/or SI Appendix.