What predicts climate change activism?: An examination of how depressive symptoms, climate change distress, and social norms are associated with climate change activism

Abstract

Background:

The current and future harms caused by climate change are highly distressing. Different theoretical models suggest diverse impacts of distress on behavior. We examined how psychological distress, climate change distress, and social norms may foster or impede climate change activism.

Methods:

As part of an ongoing online longitudinal study in the US beginning in March 2020, respondents were assessed on their depressive symptoms (CES-D 10), climate change distress, climate change mitigation social norms, and six outcomes of the climate change activism behaviors of writing letters, e-mailing, or phoning government officials; voting for candidates who support measures to reduce climate change; signing petitions; volunteering with organizations; donating money to organizations; and attending protests.

Results:

Of the 775 respondents, 53% were female, 72% white, 12% Black, 7% Hispanic, and 6% Asian. Climate change social norms predicted all six climate change actions in the bivariate and multivariable cross-sectional logistic regression models. A similar finding was observed with the brief climate change distress scale (BCCDS), except it was not associated with volunteering in the multivariable model. Depressive symptoms were associated with greater odds of contacting government officials and signing petitions in the bivariate models but did not retain significance in the multivariable models. Longitudinal models indicated a weak association between depressive symptoms and climate change activism.

Conclusions:

Climate change distress and social norms are positively associated with climate change activism. Although climate change distress may not usually impede climate change activism, organizations addressing climate change should consider providing social support to members and assisting those with high levels of psychological and climate change distress. Social norms around climate change activism should be fostered.

Introduction

Climate change mitigation is an enormous undertaking. It requires moving a global economy off fossil fuels. Massive amounts of capital, political organization, and well-functioning governments are needed to limit global warming to reach the goals outlined in the Paris Agreement [1]. There are also entrenched business interests and political parties opposed to meaningful climate mitigations. Developing and implementing policies to address climate change will require extensive and sustained collective action. Collective action can provide the co-benefits of social cohesion, a sense of community, and social support, which can, in turn, foster sustainable collective action. This manuscript examines engagement in advocacy actions to address climate change and mental health and social predictors of climate change activism.

At the individual level, climate change is causing concern or worry for many Americans. The Yale Climate Change in the American Mind project documented an increase in worry about climate change in the last decade. By 2021, 70% of Americans reported that they were at least “somewhat worried,” and 35% of respondents reported that they were “very worried” about global warming [2]. The burgeoning field of mental health and climate change research has produced detailed measures of and identified different types of climate change-related psychological distress [3]. There is an emerging body of research on “eco-anxiety,” which is an experience of fear about environmental damage or disaster. A systematic scoping review in 2021 identified nine studies on eco-anxiety [4]. This research domain has linked anxiety about climate change to poor mental health [5]. The percentage of Americans who reported that climate change “probably” or “definitely” affects their mental health rose from 47% in 2019 to 68% in 2020 [6], after a season of devastating wildfires and tropical storms. Given that the impact of climate change is increasingly felt by many, more research investigating associations between eco-anxiety and mental health is needed.

Climate change mitigation and adaptation require system-level change, yet there is a critical lack of “political will” for climate action [7]. This situation necessitates the mobilization of citizens advocating for policymakers to enact meaningful climate change policies. Therefore, it is essential to know if distress about climate change impedes or facilitates behaviors to address climate change when fostering climate change activism. However, the relationship between psychological distress and engagement in climate change action is not well understood.

Some theories of behavior change suggest that psychological distress may impede action. For example, social cognitive theories of behavior change hypothesize that anxiety and depression can reduce self-efficacy and hence reduce the probability of effective behavior change [8]. Additionally, the theory of learned helplessness would suggest that experiences of helplessness, which could be caused by an individual-level inability to address climate change, may lead to lower motivation levels to engage in climate change action [9]. In a study of Canadian college students, Landrey et al. found that learned helplessness moderated the relationship between environmental concern and pro-environmental behavior [10]. Anxiety and depression may also lead to burnout [11] and reduce climate change actions. In distinction from the perspective that negative mood states may impede climate change action is the perspective that negative mood states may impel action. If people are overloaded with demands on their time, energy, and attention, they are more likely to attend to highly salient information with strong emotional valance [12]. Furthermore, evidence suggests that negative as compared to positive events are more memorable; hence, it may be that negative emotions are more likely than positive mood states to lead people to act on climate change [12]. In contrast to these hypothesized and observed relationships between psychological distress and climate change action, Clayton and Karazsia reported that climate anxiety was not correlated with individual-level pro-environmental behavior in an online sample [13]. Previous research has produced conflicting findings on the potential relationship between psychological distress and climate change action, and additional studies are needed to elucidate these associations.

While working to understand the directionality of the relationship between mental health and climate change action, it is also essential to examine the relationship between climate change-related distress and potentially pathological states, such as depression. Often negative mood states and cognitions such as depression, anxiety, worry, and feelings of helplessness are highly correlated. Yet, depression and climate-specific distress may play different roles in engagement in climate change action. For example, an Australian study found that eco-anger was negatively associated with anxiety and depression, as measured by the DASS-21 [14], and greater engagement in pro-climate activism [15]. However, the study also found that eco-anxiety was associated with higher general anxiety levels and a lower likelihood of collective climate action. In the present analysis, we assess the relationship between depressive symptoms and climate change distress as well as examine the independent and adjusted association of climate-change distress and depressive symptoms on climate change actions.

Climate change activism is also impacted by factors other than psychological well-being, and it is important to understand the social dynamics that may foster it. Social network norms may impact engagement in climate change activism. When faced with complex issues, such as climate change mitigation, social influence often plays a role in behavioral responses [16]. How others act can motivate engagement in similar behaviors. Additionally, peers can serve as heuristics or models of climate change action behaviors [16]. Doherty and Webler’s study of Vermont residents found that social norms impacted engagement in climate change political activism, including contacting government officials, voting for “green” candidates, and protesting [17]. A greater understanding of social influence processes may be helpful in designing programs to promote climate change activism. Social norm interventions have successfully promoted energy conservation and are linked to a range of pro-environmental behaviors [18]. These same social dynamics likely influence climate change activism.

This study assesses climate change activism behaviors, such as contacting elected representatives, supporting organizations working on the issue, and attending climate change rallies [19]. We examine three questions: (1) How are social norms associated with engagement in climate change actions? (2) What is the relationship between climate change distress and climate change actions? (3) What is the relationship between depressive symptoms, climate change distress, and climate change action?

Methods

Procedures and participants

Study participants were from the online longitudinal COVID-19 and Well-Being Study, which started in March 2020 [20]. This study examines individual, social, and societal-level fluctuations amid the rapidly changing landscape of the COVID-19 pandemic. Questions about climate change were included in the study, as climate change impacts health and well-being in the current era [21]. Study assessments occurred every few months and aimed to capture changes in COVID-19-related information, behaviors, and health status. Participants were recruited through Amazon’s Mechanical Turk (MTurk).

MTurk is a crowdsourcing platform for recruiting individuals from over 40 countries to work on different tasks. Researchers can generate a HIT (Human Intelligence Task) that gives MTurk users a description of the online task, compensation, and anticipated time to complete the task. Although MTurk is open to anyone age 18 or older with internet access, study-specified qualification requirements can restrict potential participants to those who are eligible. For example, qualification requirements can restrict the audience based on geographic location, number of HITs completed, quality of prior work, and age. This platform is regularly used by health, social, and behavioral researchers, as it allows for diverse samples to be collected rapidly and in a timely manner [22]. Previous research suggests that MTurk provides better-quality data in less time than other convenience samples [23]. Study populations recruited through MTurk are not nationally representative but outperform other samples on several dimensions [24]. Studies using MTurk have demonstrated good reliability [25]. Moreover, despite the COVID-19 pandemic, the demographic characteristics of MTurk respondents have been stable [26].

The study protocols followed MTurk’s best practices, including ensuring participant confidentiality, protecting study integrity, generating unique completion codes, integrating attention and validity checks throughout the survey, repeating study-specific qualification questions, and removing ineligible participants [23,27,28]. To enhance data quality, the study enrollment criteria were embedded in a screener that did not explicitly delineate the study criteria. This method reduces the bias to over-report specific attributes to enroll in the study. Individuals were paid for the screener and those who were eligible were invited to join the study. Eligibility criteria included being age 18 or older, living in the United States, being able to speak and read English, having heard of the coronavirus or COVID-19, and providing written informed consent. Additionally, to enhance reliability, eligible participants were required to pass attention and validity checks embedded in the survey [29]. These checks included survey questions with exceedingly low probabilities, such as deep-sea fishing in Alaska and having several appendages removed. We also repeated questions to ensure consistency. Finally, we examined participants’ time to complete the survey and verified data completeness.

In the current study, we utilized data from wave 1 (baseline, March 24^th-27^th, 2020), wave 4 (November 18^th-29^th, 2020), and wave 7 (November 16^th-29^th, 2021). We chose these waves to examine whether the association might differ by phase of the pandemic and to have a longitudinal analysis component. Depressive symptoms were assessed at each study wave, while climate change activism, distress, and social norms were only assessed at wave 7. Participants were compensated $2.50 for wave 1 and $4.25 for wave 4 and wave 7. This amount was equivalent to approximately $12 per hour. The study protocols were approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board.

Measures

Climate change actions

The six climate change activism behaviors were derived from Doherty and Webler [17]. These items asked about participation (yes/no) in the following actions in the last year: “Wrote letters, e-mailed, or phoned government officials to urge them to take action to reduce climate change,” “Voted for candidates who support measures to reduce climate change,” “Signed a petition to curb climate,” “Volunteered with organizations working to curb climate change,” “Donated money to organizations working to reduce climate,” and “Attended protests or rallies to reduce climate change.” These items were individually modeled as it was hypothesized that variables associated with each item might differ. The six climate activism behavior items were also added together for a composite measure of the level of climate change action [17]. Engagement in climate change activism behaviors was assessed at wave 7.

Climate change distress

The brief climate change distress scale (BCCDS) assessed depression, anxiety, worry, and feelings of helplessness and hopelessness associated with thinking about climate change. The five questions included “News about climate change tends to depress me,” “Thinking about climate change makes me anxious,” “I worry about the future because of climate change,” “I sometimes feel hopeless about the future because of climate change,” and “I feel helpless to do anything about climate change.” The response categories were “Strongly agree,” “Agree,” “Neither agree nor disagree,” “Disagree,” and “Strongly disagree.” A Principal Component Analysis revealed one factor that accounted for 70.1% of the variance and an Eigenvalue of 3.5 (Cronbach’s alpha 0.89, range 5–25). Climate change distress was assessed at wave 7.

Depressive symptoms

Depressive symptoms were assessed using the 10-item Center for Epidemiologic Studies Depression Scale (CES-D 10), which has been well validated [30, 31,32]. The questionnaire asks respondents to indicate how often they experienced 10 possible symptoms of depression during the past week. The response categories were: Rarely or none of the time (1=less than 1 day), Some or a little of the time (2=1–2 days), Occasionally or a moderate amount of the time (3=3–4 days), and All of the time (4=5–7 days) (Cronbach’s alpha 0.90, range 10–40). The 10 items were summed after reverse coding the two positively worded items. The CES-D 10 was assessed at each study wave.

Climate change mitigation social norms

Social norms were assessed at wave 7 on 5-point scales with the items “Most of my friends are trying to act in ways that reduce climate change,” “My family and friends think it is important that I take action to reduce climate change,” and “How much of an effort do your family and friends make to reduce climate change?” The three questions were summed to create a scale (Cronbach’s alpha 0.83, range 3–15) with higher scores indicating greater pro-climate change mitigation social norms.

Sociodemographics

Demographic variables were analyzed and included age and sex, assessed as biological sex at birth. The categories for race/ethnicity were White, Non-Hispanic Black, Hispanic, Asian, Mixed, and Other. Due to small sample sizes, multiracial and “other” categories were collapsed into one category. Political party affiliation was assessed with the question, “Do you consider yourself Republican, Democrat, Independent, Libertarian or Other?” Due to the small cell size, the Libertarian and the “Other” groups were collapsed into the “Other” group. Educational attainment was dichotomized as a bachelor’s degree and higher versus an associate degree or less. Family income was assessed and dichotomized, based on the median, at less than $60,000 versus $60,000 or more.

Analyses

Descriptive statistics of means, standard deviations, and percentages were first calculated, and a correlation matrix was used to examine the relationship among the five climate change distress items. Then logistic regression models were used to examine associations between sociodemographic factors, climate change mitigation social norms, level of climate change distress, depressive symptoms and the six dichotomous climate change actions outcomes. Scale items were converted to z-scores to facilitate comparisons between the scales. For the multivariable logistic regression models, all sociodemographic variables (age, sex, income, race/ethnicity, and education) were included as potential confounds, and other variables with a p-value of <0.20 in the bivariate models were also included in the adjusted model. This approach allows for both the measurement and the control of potential confounds. We also graphically assessed the relationship between the mean number of climate change actions and climate change mitigation social norms, brief climate change distress scale scores, and depressive symptoms (CES-D 10 scores) to better understand the whole distribution. In a final analysis, we examined the correlations among the waves 1, 4, and 7 depressive symptoms (CES-D 10 scores), climate change distress at wave 7, and the number of different types of climate change actions at wave 7. Climate change distress scale scores were controlled to examine the unique contribution of the level of depressive symptoms to the number of climate change actions in the prior year. The sample size was not determined for the specific analyses; instead, the goal was to obtain a sample for the longitudinal study and have sufficient power to detect a 15% difference with 500 respondents.

Results

The mean age of subjects in the sample was 41, and slightly more than half (53%) were female (Table 1). There was a substantial distribution by race/ethnicity in the sample, with 72% white, 12% Black, 7% Hispanic, and 6% Asian. There was also a distribution of political affiliation with 23% Republican, 46% Democrat, 27% Independent, and 4% other. The sample was well educated, with 60% reporting a bachelor’s degree or higher, and 44% reporting an annual household income of more than $60,000. The sample also reported a high level of climate change action participation in the prior year, with 51% voting for a candidate who supported measures to address climate change, 24% signing a petition, and 25% donating money to organizations working to reduce climate change. Other activities were less frequently reported, with 11% reporting that they volunteered with organizations working to curb climate change, 11% wrote letters, e-mailed, or phoned government officials to urge them to take action to reduce climate change, and 7% attended protests or rallies to reduce climate change. Participation in no climate change actions was reported by 42% of the sample; 24% reported one, 15% reported two, 10% reported three, 5% reported four, 2% reported five, and 3% reported all six actions. The mean score on the CES-D 10 was 17.86 (Range: 10–40) at wave 7. The five items on the brief climate change distress scale were all statistically significantly correlated (p<.01), with Spearman’s correlations ranging from 0.41 to 0.75 (Table 2).

Table 1.

Background, demographics, and climate change behaviors among respondents at wave seven (N=775).

	n (%), mean (SD)
Age, in years, mean (SD)	40.84 (11.91)
Sex, female	414 (53.42%)
Race/Ethnicity
White	557 (71.87%)
Non-Hispanic Black	96 (12.39%)
Hispanic	55 (7.10%)
Asian	43 (5.55%)
Other	24 (3.10%)
Political Affiliation
Republican	176 (22.71%)
Democrat	356 (45.94%)
Independent	211 (27.23%)
Other	32 (4.13%)
Education, undergraduate degree or higher	463 (59.74%)
Annual income, >$60,000	342 (44.13%)
Scale scores
Climate change mitigation social norms, mean (SD)	9.81(1.70)
Depressive symptoms (CES-D 10), mean (SD)	17.86 (6.50)
Brief climate change distress scale, mean (SD)	15.56 (5.43)
Six Climate Change Actions (in past year)
Wrote letters, e-mailed, or phoned government officials	88 (11.35%)
Voted	399 (51.48%)
Petition	190 (24.52%)
Volunteered	86 (11.10%)
Donated	186 (24.00)
Attended protests or rallies	55 (7.10)

SD=standard deviation, CES-D= Centers for Epidemiological Studies Depression Scale.

Table 2.

Spearman correlation matrix for five climate change distress survey items and the sum of six climate change actions (N=775).

	News about climate change tends to depress me	Thinking about climate change makes me anxious	I worry about the future because of climate change	I sometimes feel hopeless about the future because of climate change	I feel helpless to do anything about climate change	Six climate change actions added together
News about climate change tends to depress me	1.00
Thinking about climate change makes me anxious	0.75	1.00
I worry about the future because of climate change	0.63	0.71	1.00
I sometimes feel hopeless about the future because of climate change	0.62	0.72	0.70	1.00
I feel helpless to do anything about climate change	0.50	0.50	0.41	0.62	1.00
Six climate change actions added together	0.40	0.45	0.51	0.43	0.27	1.00

In the multivariable logistic regression models of the six climate change actions (Table 3), age was positively associated with voting and negatively associated with volunteering and attending rallies or protests. Sex was not associated with any of the outcomes. There was a consistent finding of Democratic political party affiliation linked to all climate change action outcomes in the bivariate models and contacting governmental officials, voting, and donating in the multivariable models. A higher level of education was also associated with all outcomes except petition signing in the bivariate models and four of the six outcomes in the multivariable models (not petition signing or donating). Lower income was associated with two outcomes (contacting governmental officials and attending protests or rallies) in the bivariate and multivariable models.

Table 3.

Unadjusted and adjusted logistic regression models for six climate change actions (N=775).

	Model 1: Wrote letters, e-mailed, or phoned government officials		Model 2: Voted		Model 3: Petition
	OR(95%CI)	aOR (95% CI)	OR (95% CI)	aOR (95% CI)	OR(95%CI)	aOR (95% CI)
Age	0.99(0.97,1.01)	1.00(0.98,1.03)	0.99 (0.98,1.01)	1.02(1.01,1.04)	0.98 (0.97, 0.99)	0.99(0.98,1.01)
Sex	1.17(0.75,1.83)	1.18(0.73,1.91)	0.98 (0.74,1.30)	0.82(0.58,1.16)	1.38 (0.99,1.93)	1.38(0.96,2.00)
Race/Ethnicity (ref: White)
Non-Hispanic Black	0.44(0.19,1.05)	0.52(0.21,1.31)	0.93 (0.61,1.44)	1.09(0.65,1.84)	0.98 (0.59,1.64)	1.35(0.76,2.41)
Hispanic	0.52(0.18,1.48)	0.47(0.16,1.43)	1.12(0.64,1.95)	1.28(0.66,2.48)	1.40 (0.76, 2.56)	1.61 (0.81,3.17)
Asian	0.50(0.15,1.65)	0.35(0.10,1.25)	0.89 (0.48,1.66)	0.51 (0.25,1.07)	0.71 (0.32,1.58)	0.57(0.24,1.35)
Other	0.60 (0.14, 2.62)	0.51 (0.11,2.38)	0.93 (0.41,2.11)	1.02(0.39,2.64)	1.29 (0.52, 3.17)	1.28(0.48,3.44)
Political Affiliation (ref: Republican)
Democrat	4.89 (2.19,10.95)	2.61 (1.11, 6.14)	5.66 (3.80, 8.44)	2.85 (1.76, 4.61)	3.26 (1.98, 5.36)	1.47 (0.84, 2.58)
Independent	2.39 (0.98, 5.82)	1.80(0.70,4.62)	2.27(1.48,3.48)	1.58(0.95,2.62)	2.06(1.19,3.57)	1.37(0.74,2.53)
Other	1.61 (0.32, 8.12)	1.49 (0.27, 8.29)	1.60 (0.73,3.52)	1.65(0.66,4.10)	1.96 (0.76, 5.07)	1.51 (0.52,4.36)
Education	1.92(1.18,3.15)	2.26(1.30,3.91)	1.63(1.22,2.18)	1.54(1.07,2.21)	0.93 (0.67,1.30)	0.88(0.60,1.29)
Income	0.62 (0.39, 0.99)	0.54 (0.32, 0.90)	1.13(0.85,1.50)	1.26(0.87,1.81)	0.75 (0.54,1.05)	0.86(0.59,1.27)
Climate change mitigation social norms scale (z-score)	2.13 (1.65, 2.74)	1.74 (1.30,2.33)	2.24 (1.90,2.66)	1.60(1.31,1.96)	2.07(1.71,2.51)	1.78(1.43,2.21)
CES-D, Depressive symptoms scale (z-score)	1.25 (1.02,1.54)	1.22(0.95,1.56)	0.99 (0.87,1.15)	0.84(0.70,1.02)	1.21 (1.03,1.42)	1.07(0.88,1.30)
Brief climate change distress scale (z-score)	2.35 (1.78,3.09)	1.74 (1.26,2.42)	3.03 (2.52, 3.65)	2.88 (2.29, 3.63)	2.55(2.07,3.14)	2.15(1.68,2.75)
	Model 4: Volunteer		Model 5: Donated		Model 6: Attended protests or rallies
	OR (95% CI)	aOR (95% CI)1	OR (95% CI)	aOR(95%CI)	OR(95%CI)	aOR (95% CI)3
Age	0.96 (0.93, 0.98)	0.95 (0.93, 0.98)	0.99(0.97,1.00)	0.99 (0.98,1.01)	0.95 (0.93, 0.98)	0.95 (0.92, 0.98)
Sex	1.11 (0.71,1.75)	1.16(0.70,1.94)	0.96(0.69,1.34)	0.92 (0.64,1.34)	0.90(0.52,1.55)	0.82(0.44,1.51)
Race/Ethnicity (ref: White)
Non-Hispanic Black	1.06 (0.55,2.05)	1.51 (0.68,3.32)	0.88(0.53,1.46)	1.06 (0.59,1.89)	0.602 (0.23,1.54)	0.69(0.24,1.99)
Hispanic	0.28 (0.07,1.18)	0.26 (0.06,1.16)	0.74(0.37,1.47)	0.76 (0.36,1.60)	0.202 (0.03,1.49)	0.15(0.02,1.15)
Asian	0.76 (0.26, 2.20)	0.33 (0.10,1.05)	0.67(0.31,1.49)	0.41 (0.17,0.98)	0.532 (0.12,2.26)	0.21 (0.04,1.02)
Other	0.68 (0.16,2.94)	0.54(0.10,2.84)	0.59(0.20,1.76)	0.55(0.16,1.88)	1.002	1.00
Political Affiliation (ref: Republican)
Democrat	2.391 (1.30,4.41)	1.08(0.52,2.24)	3.70 (2.25, 6.08)	1.98(1.12,3.49)	3.691 (1.53,8.86)	1.83(0.69,4.83)
Independent	0.641 (0.28,1.44)	0.48(0.19,1.19)	1.49(0.84,2.63)	1.16(0.62,2.17)	1.111 (0.38,3.28)	0.87(0.27, 2.83)
Other	1.001	1.00	1.00(0.32, 3.12)	1.05 (0.31,3.57)	1.001	1.00
Education	3.92 (2.17,7.09)	4.41 (2.27, 8.57)	1.52(1.08,2.15)	1.26 (0.84,1.87)	3.25 (1.61, 6.55)	4.37 (2.00, 9.55)
Income	1.11 (0.71,1.75)	0.89(0.52,1.50)	1.25(0.90,1.75)	1.33 (0.91,1.96)	0.54 (0.30, 0.98)	0.39 (0.20, 0.75)
Climate change mitigation social norms scale (z-score)	3.06 (2.29, 4.08)	2.67(1.90,3.76)	2.54(2.07,3.13)	2.00(1.60,2.51)	2.69 (1.93,3.75)	2.17(1.46,3.22)
CES-D, Depressive symptoms scale (z-score)	1.02 (0.82,1.27)	1.22 (0.93,1.62)	0.96(0.81,1.13)	0.94(0.77,1.16)	1.15(0.89,1.49)	1.20(0.87,1.66)
Brief climate change distress scale (z-score)	2.00(1.54,2.61)	1.34(0.95,1.89)	2.17(1.78,2.64)	1.77(1.38,2.26)	2.38 (1.69,3.36)	1.58(1.04,2.41)

OR= odds ratio, aOR=adjusted odds ratio, CI=confidence interval, CES-D= Centers for Epidemiological Studies Depression Scale.

¹n=743.

²n=751.

³n=721.

In all the bivariate and multivariable models, climate change mitigation social norms were associated with all six climate change actions. A similar finding was observed with climate change distress, except it was not significantly associated with volunteering in the multivariable model. The level of depressive symptoms was only associated with greater odds of contacting government officials and signing petitions in the bivariate models. These associations did not retain significance in the multivariable models.

The subsequent analyses graphically examined the association between depressive symptoms, climate change distress, climate change social norms, and the number of different types of climate change actions in the prior year. As seen in Fig. 1, there is a linear association between climate change distress and the number of different types of climate change actions. However, at the right end of the climate change distress distribution, there appears to be a drop-off in the number of different types of climate change actions. In comparison, there is a clear linear relationship between the level of climate change mitigation social norms and the number of different types of climate change actions. The relationship between the level of depressive symptoms and the number of types of climate change actions is modest and curvilinear, with those reporting very high levels of depressive symptoms less likely to be engaged in climate change actions. The measure of depressive symptoms and the climate change distress distribution (Fig. 2) also demonstrates a deviation from linearity at the right end.

Fig. 1.

Mean number of climate change actions by climate change distress, climate change social norms, and CES-D 10 scores. CES-D: Centers for Epidemiological Studies Depression Scale.

Fig. 2.

Mean climate change distress scores by CES-D 10 scores. CES-D: Centers for Epidemiological Studies Depression Scale.

The final analyses examined the relationship between prior levels of depressive symptoms (baseline and wave 4), climate change distress (wave 7), and the number of climate change actions reported (wave 7) among the 414 participants who completed all three waves (Table 4). Level of climate change distress was strongly associated with the number of climate change actions in the prior year (p<.01, r=.46). Wave 1 and wave 4 CES-D 10 scores predicted the number of climate change actions in the prior year in wave 7 (r=.17, p<.01, r=.11, p<.05). However, after adjusting for climate change distress at wave 7, neither wave 1 nor wave 4 CES-D 10 scores predicted the level of climate change action at wave 7.

Table 4.

Correlations and partial correlations of measures of depression, climate change actions, and climate change distress over three waves of data collection (N=414).

	Variables	Number of change climate actions	CES-D 10 at wave 7	CES-D 10 at wave 4	CES-D 10 at wave 1	Brief climate change distress scale
	Pearson correlations, (p-values in parentheses)
Pearson correlation coefficients	Number of climate change actions	1.000.	.059 (.228)	.111 (.023)	.168 (.001)	.449 (.000)
	CES-D 10 Wave 7	.059 (.228)	1.000.	.789 (.000)	.734 (.000)	.239 (.000)
	CES-D 10 Wave 4	.111 (.023)	.789 (.000)	1.000.	.763 (.000)	.238 (.000)
	CES-D 10 Wave 1	.168 (.001)	.734 (.000)	.763 (.000)	1.000.	.250 (.000)
	Brief climate change distress scale	.449 (.000)	.239 (.000)	.238 (.000)	.250 (.000)	1.000.
Partial correlation controlling climate change distress	Number of climate change actions	1.000.	−.055 (.261)	.005 (.915)	.065 (.189)
	CES-D 10 Wave 7	−.055 (.261)	1.000.	.777 (.000)	.717 (.000)
	CES-D 10 Wave 4	.005 (.915)	.777 (.000)	1.000.	.748 (.000)
	CES-D 10 Wave 1	.065 (.189)	.717 (.000)	.748 (.000)	1.000.

CES-D: Centers for Epidemiological Studies Depression Scale.

414 respondents completed all three waves of data collection: wave 1 (baseline, March 2020), wave 4 (November, 2020), and wave 7 (November, 2021).

Discussion

This study identified climate change distress and social norms about climate change mitigation as strong and independent factors associated with engaging in climate change activism behaviors. However, Clayton and Karazsia reported that their measure of climate change anxiety was not associated with pro-environmental behaviors [13]. Yet, their measures of behaviors were individual pro-environmental such as recycling and turning off lights, while the current study assessed climate change activism behaviors. One key question to further explore is whether depression, anxiety, and climate change distress have similar influences on individual-level pro-environmental behaviors. Future research should also examine the relationship between individual-level pro-environmental behaviors and collective actions and assess if promoting individual-level pro-environmental behaviors leads to collective actions.

The current analysis expands on Doherty and Webler’s findings that identified climate change mitigation social norms as a predictor of engagement in climate change action [17]. Doherty and Webler’s analysis included Vermont residents who were highly concerned about climate change. The current analysis generalizes these findings using a US population sample with diverse views about climate change mitigation.

Another key finding from this study is that climate change distress and depressive symptoms have a differential impact on climate change activism behaviors. The measure of climate change distress included items that assessed depression, anxiety, worry, hopelessness, and helplessness. These five items were all significantly associated, with correlations ranging from 0.41 to 0.75, with anxiety and depression having the highest correlation. These strong associations suggest that we are measuring the construct of climate change distress which, albeit correlated, is distinct from clinical depression and anxiety. One advantage of the current study was that the brief measure of climate change distress was strongly associated with several climate change activism behaviors. We encourage other researchers to use this brief measure in studies of other populations.

The relationship between depressive symptoms and climate change distress was slightly linear. However, it demonstrated a deviation from linearity at the right end, suggesting that this relationship may plateau at high levels. Despite the correlation between the level of climate change distress and depressive symptoms, depressive symptoms were not independently associated with engagement in climate change action. In contrast, scores on the brief climate change distress scale were consistently and independently associated with climate change activism behaviors. This study also identified a nonlinear relationship between climate change distress and engagement in climate change action. A drop-off was evident at the extreme right-hand tail of the plot of climate change distress and climate change action. These findings suggest that severe levels of climate change distress and depressive symptoms may impede climate change actions.

Study findings suggest that it is important to acknowledge climate change distress and promote methods of channeling it into climate change activism behaviors and ensuring that it does not lead to psychopathology or deleterious coping mechanisms. Providing social support by climate action organizations may be beneficial. This support can be accomplished by structuring climate change action activities in small groups, providing personal feedback and support to members, and providing opportunities for activists to get to know each other and support each other. It may also be beneficial to acknowledge in forums with climate change activists that climate change distress is an appropriate reaction to the current situation and that climate change distress is not pathological or irrational. Leaders of climate change action organizations should also be trained to recognize signs of clinical depression and obtain professional support for those who have high levels of anxiety and depression symptomology. Moreover, since the relationship between high levels of climate change distress and climate change actions may become negatively correlated, it is also important that climate change action organizations monitor and address high levels of climate change distress. Finally, as the study was cross-sectional, we do not know if heightening climate change distress will lead to increased actions. Consequently, future research should longitudinally examine these associations.

The empirical literature suggests that social norms interventions can promote pro-environmental behaviors [18]. These types of interventions could be expanded to include climate change actions. Potential strategies include communication campaigns that promote positive climate change mitigation social norms, mechanisms and strategies for managing climate change distress, and educating individuals to be peer educators about climate change and mitigation strategies. As climate change mitigation requires collective action, individuals who engage in climate change activism can notify their friends of these actions to heighten social norms about climate change activism. Moreover, given the association between climate change distress and climate change activism behaviors, support from friends to engage in climate change action can buffer stress while heightening climate change activism social norms.

In line with other research, Democrats were more likely than Republicans to report voting for a candidate who supported addressing climate change [33]. We also found that Democrats were more likely than Republicans to donate to organizations working to reduce climate change. Organizations that address climate change may want to consider greater outreach to Republicans. Democrats tend to perceive Republicans as less concerned about climate change than they actually are [34, 35]. This dynamic may impede outreach to Republicans, who may be an essential constituency for groups promoting climate change activism since Republicans may have more contacts with Republican legislators and, hence, well-positioned to lobby them.

Study limitations should be noted. The survey was not completed by a random or representative sample, which limited the generalizability of the findings. However, attrition was low for an online sample, and wave 7 had substantial diversity in race/ethnicity and political party affiliation. The measure of social norms has not been validated, and the cross-sectional analyses limit causal inferences. Although we used a validated climate activism measure, there is no standard definition of these behaviors. Hence key behaviors may have been omitted, and future studies should systematically delineate these activities. This longitudinal study was initiated early in the COVID-19 pandemic, and we did not assess climate change actions or climate change distress in the early waves since the pandemic was overshadowing other important issues. It is also likely that the pandemic increased levels of depression and anxiety [36, 37]. In addition, due to the strong correlation between depression and anxiety symptoms, separate measurements of the two conditions were not included as that can lead to multicollinearity if both are included in one statistical model [33, 38]. As the sample was comprised only of adults, the findings cannot be generalized to children. The sample size also was not determined a priori due to the longitudinal nature of the study and unknown effect sizes. We encourage other researchers to replicate the findings in other populations, given the study limitations. As many racial minorities and ethnic groups are at increased risk from climate change, we also encourage researchers to work with these populations to study climate change distress, mental health, and climate activism.

There are many avenues of research that warrant further attention. Future research should examine those with very low levels of climate change distress, both those engaged in climate change action and those who are not. This will allow for a greater understanding of the role of psychological distress in engaging in climate change action. Moreover, future research should validate findings with other online platforms that have documented high-quality data, such as Prolific [39]. Organizations that work towards addressing climate change should monitor and address high levels of depressive symptoms and climate change distress. However, they should not pathologize climate change distress. Future research should examine the role of both depression and climate change distress in burnout from climate change activism. It is also important to examine how people felt about their climate change actions or psychosocial factors that helped them maintain climate change action.

Climate change distress can be viewed as a normal and appropriate reaction to the current climate change emergency. Although there is a substantial level of concern among the general population in many countries about global warming, there is far too little action to ensure that the temperature of the earth’s atmosphere does not rise to a level that will cause massive destruction and adverse consequences for all life on the planet. Key questions for the community of climate change researchers and practitioners are how to support individuals who report climate change distress and psychological distress while channeling climate change distress effectively into action and how to foster pro-climate change action social norms to increase global action on climate change. It also may be beneficial to examine the types of social norms most effective at facilitating climate change activism.

Availability of data and materials

This is an ongoing longitudinal study. A deidentified data set is available from the authors.