Some Slice of Climate Anxiety … Is Good: A Cross-Sectional Survey Exploring the Relationship Between College Students Media Exposure and Perceptions About Climate Change

1

Abstract

Climate change anxiety among young people of college age has become a prevalent topic, with debate on whether climate change anxiety is maladaptive or can motivate climate change action. Using a cross-sectional survey of 440 college students, the study investigated the relationships between college students’ climate anxiety, climate change media exposure, efficacy beliefs, and pro-environmental intentions. The findings revealed among other things, that climate anxiety had a significant curvilinear relationship with pro-environmental intentions with moderate anxiety predicting positive intentions and higher levels of anxiety were associated with negative intentions. Media exposure also positively predicted increased climate anxiety. Implications for climate change mitigation are discussed.

“…We are on a highway to climate hell with our foot on the accelerator” … “Humanity has a choice: cooperate or perish.”

This was how United Nations Secretary-General António Guterres described the impact of climate change at the 27th Conference of the Parties to the United Nations Framework Convention on Climate Change (COP27) in the fall of 2022 (United Nations, 2022). Although the direct impacts of climate change, from rising sea levels to increased extreme weather events and wildfires, have received substantial attention (Harvey, 2016; United Nations, n.d.), less attention has been given to the indirect effects of climate change on mental health (e.g., anxiety, stress, etc.). Specifically, younger generations exhibit indirect effects termed “climate anxiety” (Morling, 2020). Climate change anxiety is found to be more common among young people in the U.S., especially those aged 18–29 years with approximately 7 in 10 American young adults reporting feeling anxious about climate change (Hickman et al., 2021). Climate anxiety can stem from direct experiences of extreme weather or disasters or indirect exposure to climate media information (Ogunbode et al., 2019), with studies showing that media exposure (frequency and attention) to climate change is a prominent factor that contributes significantly to young people’s feelings about climate change (Brulle, Carmichael, & Jenkins, 2012; Loll, Schmatz, von Lonski, Cremer, & Richter, 2023).

However, some scholars have argued that climate change anxiety might not be solely maladaptive (Crandon, Scott, Charlson, & Thomas, 2022). Despite this knowledge, climate change anxiety is still portrayed in the media and some academic literature as a pathological phenomenon with a negative and detrimental influence and needs to be eradicated, usually by emphasizing avoidance-based coping interventions like tuning off information sources (Ojala, 2012; Wu, Snell, & Samji, 2020). This highlights the need to further understand the complexities associated with the disparate levels of climate anxiety among young people and how climate anxiety influences pro-environmental action.

While extreme levels of anxiety could impair people from responding to environmental challenges and potentially lead to feelings of despair, research also suggests the pertinent role of efficacy beliefs in enhancing pro-environmental behaviors (Burnham & Ma, 2017; Skurka, Troy, Cui, & Gil de Zúñiga, 2023). Perceived efficacy beliefs are defined by Bandura (1998) as “beliefs in one’s capabilities to organize and execute the courses of action required to produce given levels of attainments” (p. 53). Perceived efficacy beliefs, which play a pivotal role in driving behavioral outcomes (Feldman & Hart, 2016), encompass different dimensions, with self-efficacy and collective efficacy being prominent among them. Self-efficacy refers to an individual’s confidence in his or her ability to make a meaningful impact or take effective action, while collective efficacy involves an individual’s belief in the effectiveness of group actions (Bandura, 2000). Although there has been considerable exploration of the relationship between self-efficacy beliefs and pro-environmental behaviors, the collective dimensions of efficacy beliefs have received comparatively limited attention (Mzimela & Moyo, 2023). Scholars have underscored the collective nature of climate change as a pervasive issue affecting all individuals, potentially limiting one’s perceived capacity to address it independently (Koletsou & Mancy, 2011).

Therefore, this study investigated the potential adaptive and maladaptive nature of climate change anxiety in terms of its capacity to motivate or discourage proactive climate change behaviors. It contends that while media exposure may influence climate anxiety among young people of college age, there could be a nonlinear trend in the effects of minimal and extreme levels of climate anxiety on pro-environmental behavior, such that minimal levels of climate anxiety can motivate individuals to engage in positive actions that mitigate climate change, whereas extreme levels of climate anxiety can begin to impair their psychological well-being and reduce their perceptions of agency and efficacy. The study also explored the relationship between efficacy beliefs and pro-environmental behavioral intentions. There are important implications for climate change mitigation, adaptation, and general climate communication. The findings enhance understanding of the nuances related to climate change anxiety, specifically concerning exposure to climate change information through the media. Additionally, it provides insights into how climate change communication relates to the psychological processes of young individuals.

Literature Review

Media Exposure and Climate Change Anxiety

Media exposure can be defined as “the extent to which audience members have encountered specific messages or classes of messages/media content” (Slater, 2004, p. 168). Although this conceptualization of media exposure has been the subject of scholarly debate (see de Vreese & Neijens, 2016), Chaffee and Schleuder (1986) provided a fundamental distinction between passive media exposure, which is incidental, and active media seeking, which involves deliberate cognitive effort. This postulation takes into consideration the level of attention and intention that individuals have when they consume media information about climate change.

While some studies have operationalized passive media exposure as incidental encounters with media or frequency of media use (Loy, Hamann, & Reese, 2020), active media exposure is often operationalized as the degree of attention devoted to media content (Slater & Rasinski, 2005). The current study followed Slater’s (2004) recommendations in operationalizing both passive and active climate media exposure separately as the frequency of media use and attention to climate change news, respectively. In exploring both constructs of media exposure, the study offers a nuanced exploration of the complex dynamics underlying the passive and active nature of media exposure to climate change.

Traditional and social media exposure to climate change events are prominent sources of climate change-related psychological distress (Brulle, Carmichael, & Jenkins, 2012; Loll, Schmatz, von Lonski, Cremer, & Richter, 2023). This notion is grounded in Crandon, Scott, Charlson, and Thomas (2022) socio-ecological perspective on climate anxiety, which argues that individuals, especially children and young adults, are influenced by an interplay of direct and indirect environmental factors organized in a hierarchical structure ranging from micro (individuals, family) and meso (communities) to exo (media, governmental structures) and macro systems (cultural forces). They noted that the media is part of the exosystems, influencing climate anxiety, as unrestrained access “to sensationalist media, as well as news and images about global disasters, environmental destruction, and death,” has the potential to induce “a sense of despair and helplessness” (Crandon, Scott, Charlson, & Thomas, 2022, p. 127).

The media also plays an agenda-setting role and is “largely responsible for the transmission of scientific knowledge and its incorporation into daily experiences” (Maran & Begotti, 2021, p. 2; McCombs, 2004). The representations of climate change issues by the media usually translate into the perceptions of the audience, the salience with which they attend to such issues, and invariably, the impacts such media representations have on their psychological well-being (Andrews & Caren, 2010). Brulle, Carmichael, and Jenkins (2012) investigated the factors responsible for individuals’ anxiety about climate change. Their results revealed that media coverage was one of four main factors that predict anxiety about future climate conditions among young people. Furthermore, Maran and Begotti (2021) found that the more younger Italian adults pay attention to media messages about climate change, the more they experience climate anxiety. Taken together, it seems plausible to hypothesize:

H1: Media exposure (frequency of exposure and attention) to climate change information will be positively related to climate anxiety.

Is Climate Change Anxiety Adaptive, Maladaptive, or Both?

Several studies have noted that climate anxiety is a prominent mental health factor that influences the well-being of young people (Crandon, Scott, Charlson, & Thomas, 2022; Hickman et al., 2021; Loll, Schmatz, von Lonski, Cremer, & Richter, 2023). Clayton and Karazsia (2020) identified two main dimensions of climate anxiety: cognitive-emotional and functional impairments. While the former relates to cognitive and affective orientations of anxiety, the latter refers to the impairment of daily functionalities with deleterious consequences such as difficulty concentrating or sleeping. While there have been multiple conceptualizations of climate-related psychological conditions with varied implications (see Clayton & Karazsia, 2020), Climate anxiety can serve as a standard conceptualization with validated measurement to capture the psychological distresses associated with climate change (Wullenkord, Tröger, Hamann, Loy, & Reese, 2021).

The debates around the adaptive-maladaptive nature of climate change anxiety are relevant for understanding climate anxiety’s impact on climate action. Dodds (2021) argues on the side of the negative impacts of climate change anxiety on behavior, suggesting that all levels of climate anxiety could be detrimental to prosocial behaviors. They explain that triggering neurological systems associated with anxiety about climate change may also result in “denying the reality of climate change (it does not exist, it is a conspiracy), or denying our losses, dependency or responsibility (nature might die but we will be fine; it is caused by other humans or natural causes, the Chinese or the sun, not me)” (Dodds et al., p. 224). On the other hand, Clayton and Karazsia (2020) postulated that climate anxiety can be both adaptive and maladaptive. They argued that anxiety and worry play crucial and associative roles in the survival of both humans and animals because it involves experiencing negative emotions that come with physical symptoms and a forward-looking sense of worry. Because anxiety is future-oriented, it can motivate individuals to make necessary preparations for upcoming tasks, such as performing or taking precautions against potential threats (Barlow, 2004; Clayton & Karazsia, 2020). In this sense, it could be beneficial for individuals to experience minimal levels of anxiety about climate crises to engage in pro-environmental actions. Crandon, Scott, Charlson, and Thomas (2022) further suggested the nonlinearity of the association between climate anxiety and pro-environmental behavior, noting that while “adaptive climate anxiety may facilitate problem-solving and pro-environmental behavior, maladaptive climate anxiety becomes overwhelming. This overwhelming level of anxiety may involve fear, helplessness, hopelessness, powerlessness and an avoidance of the threat” (p.123).

However, Crandon, Scott, Charlson, and Thomas (2022) acknowledged that whether climate change anxiety is adaptive or maladaptive for pro-environmental behavioral outcomes is unclear and constitutes an empirical question. Similarly, Brulle and colleagues highlighted the lack of research on climate change anxiety among young people, even though people younger than 30 years were likely to have higher levels of anxiety than people of other age groups (Brulle, Carmichael, & Jenkins, 2012). Moreover, none of these studies explored the nuanced differences between individuals with lower and higher levels of climate anxiety. Nonetheless, it is not unreasonable to expect a curvilinear relationship between climate change anxiety and pro-environmental behavior, such that low levels of climate change anxiety could predict positive pro-environmental behaviors, as such behavior represents simple concern or worry about climate change, while higher levels may become cognitively constraining and could be associated with a decrease in intentions to engage in pro-environmental actions. Given the mixed findings on the relationship between climate change anxiety and pro-environmental actions, the following research question was posed.

RQ1: Is there a curvilinear relationship between climate change anxiety and pro-environmental action intentions?

Efficacy Beliefs and Pro-Environmental Behavioral Intentions

Although extreme levels of anxiety could impair people from responding to environmental challenges and potentially lead to a perceived lack of control, efficacy beliefs conceptualize an individual’s ability to take control of different aspects of their lives (Bandura, 1998, 2000). Bandura defined perceived efficacy as “beliefs in one’s capabilities to organize and execute the courses of action required to produce given levels of attainments” (Bandura, 1998, p. 53). This implies the subjective interpretation of one’s ability to execute actions that lead to an anticipated outcome. Perceived efficacy is a strong predictor of human behavior across multiple disciplines, from education (Goddard, 2001) to sports (Chow & Feltz, 2014) and health (Klassen & Klassen, 2018).

Efficacy beliefs have been extensively explored in relation to climate change adaptation and perceptions (Bieniek-Tobasco, Rimal, McCormick, & Harrington, 2020; van Valkengoed & Steg, 2019). Hornsey, Chapman, and Oelrichs (2021) describe climate change efficacy as one’s belief in one’s ability to effectively respond to climate change issues. Efficacy beliefs are thought to exist on different dimensions across different disciplines (Bieniek-Tobasco, Rimal, McCormick, & Harrington, 2020; Chow & Feltz, 2014; Goddard, 2001). However, the most common dimensions are the self- and collective efficacy dimensions. Self-efficacy emphasizes the individual and their subjective ability to respond to an issue (Bandura, 1998). Collective efficacy refers to one’s belief in the ability of their group (Bandura, 2000).

While research on climate efficacy beliefs has often emphasized the individual-oriented dimension of efficacy beliefs (Klassen & Klassen, 2018; Loy, Hamann, & Reese, 2020), climate change issues are typically viewed as social dilemmas that necessitate collective action (Koletsou & Mancy, 2011). Chen (2015) also noted that “people are social beings and rely on each other to find solutions to problems for improving their quality of life” (p. 72). Therefore, the concept of collective efficacy has been suggested as an essential dimension of efficacy in this context, as it addresses the collaborative aspect of addressing climate change. In the context of collective-oriented objectives such as climate change mitigation and adaptation, one’s perception of their group’s ability could then serve to boost their ability to engage in climate action (Jugert et al., 2016). While acknowledging the significance of individual self-efficacy, individuals could also draw from the communal sharing of knowledge and resources, coupled with a collective commitment to enhancing their social systems, especially in environmentally related circumstances where they might perceive their personal capacity as insufficient.

Extant research indicates that collective efficacy beliefs are strongly related to climate-adaptive behaviors (Bieniek-Tobasco, Rimal, McCormick, & Harrington, 2020; Skurka, Troy, Cui, & Gil de Zúñiga, 2023). Homburg and Stolberg (2006) found among adults that pro-environmental behavior hinges on individuals’ perceptions of collective efficacy. Thaker, Howe, Leiserowitz, and Maibach (2019) also discovered that among Indians, increased collective efficacy was associated with greater endorsement of climate policies and increased climate activism behaviors. Because these studies were conducted on different populations with varying demographics, we intend to add to the literature by focusing on college students in the U.S. To that end, we hypothesize the following:

H2a: College students’ (a) self-efficacy beliefs and (b) collective efficacy beliefs positively predict pro-environmental behavioral intentions.

Methods

This study employed a survey method whereby participants were selected using a nonrandom convenience sampling method. A priori power analysis was computed using G*Power (Faul, Erdfelder, Lang, & Buchner, 2007) for multiple regression with a small-to-medium effect size (f² = 0.04–0.09) based on previous studies (Huang, 2016; Whitmarsh et al., 2022), with a power of 0.80, an alpha level of 0.05, for four predictors. It was determined that N = 304 participants would be necessary to detect the anticipated effects in the analysis. A total of 440 students from a Western university participated in the survey. The Institutional IRB approved the study [IRB # STUDY00000781]. The participants were also offered extra credit, as permitted by the supervising instructor for their courses. Students who opted out of participation still received extra credit. The sampling method for this study was deemed suitable due to the peculiarity of the target population, people of college age. The sampling method aligns with the goals of the current study, which seeks to explore the media usage of college students. Additionally, the study is exploratory in its attempt to investigate the unique relationships between variables rather than generalizing to the general population (Wimmer & Dominick, 2013). Thus, convenience sampling was an appropriate method for the current study.

Measures

Media Exposure to Climate Change News

Following Maran and Begotti (2021) and Ogunbode et al. (2019), media exposure was measured using two separate measures: a single item measuring attention to climate change news with a question stem, “Typically, how attentive or inattentive are you to information about climate change?” (1= Very Inattentive, 5 = Very Attentive), and a 13-item scale measuring the frequency of media use (e.g., TV, radio, YouTube, Facebook, etc.; see supplementary materials for full survey items). Responses to the items ranged from less often than once a week (1) to over 10 times per day (8). Following previous research (Huang, 2016; Loy, Hamann, & Reese, 2020), the frequency of media use was averaged into a composite variable, where a higher score indicates a greater frequency of media use (α =.75).

Efficacy Beliefs

Efficacy beliefs were measured using a 10-item scale adapted from Huang (2016) and Jugert et al. (2016) with two dimensions (self-efficacy and collective efficacy). Sample items include “My actions can contribute to reducing the impacts of climate change.” Responses range from not sure (1) to extremely confident (5). Self-efficacy was averaged into a composite variable where a higher score indicates greater self-efficacy (=.83). The collective efficacy items were also averaged into a composite variable where a higher score indicates greater collective efficacy (α =.87).

Climate Change Anxiety

Climate change anxiety was measured using the 13-item Climate Change Anxiety Scale (CCAS) scale (Clayton & Karazsia, 2020) with responses ranging from Never (1) to Almost Always (5). Climate change anxiety items were averaged into a composite variable where a higher score indicates higher climate change anxiety. (α =.84)

Pro-Environmental Behavioral Intentions

Using Coleman, Thorson, Jimenez, and Vinton (2022), pro-environmental behavioral intentions were measured using an 8-item scale with statements such as “Make lifestyle changes for reasons related to protecting the climate.” The scale ranges from “Extremely Unlikely” (1) to “Extremely Likely” (5). These items were combined into a composite variable where a higher score indicates greater intentions to engage in pro-environmental behaviors (α =.86). Although this approach is limited by the fact that intentions do not always translate to actual behavioral outcomes, the use of pro-environmental intentions as a proxy for climate action is supported by related research (Baden, McIntyre, & Homberg, 2019; Feldman & Hart, 2016).

Demographic Information

Demographic information such as age, sex, educational level, race/ethnicity, and political affiliation was also collected as covariates. (See supplementary material for a full description of items).

Analytical Strategy

All the analyses and data cleaning were performed with R programming software (R Core Team, 2023). Predictor variables were centered to reduce multicollinearity and allow for interpretability (Iacobucci, Schneider, Popovich, & Bakamitsos, 2016). Missing data were imputed (see supplementary material). Robust regression models were fitted and estimated using the M-estimator with iterative reweighted least squares (I.R.W.L. S) optimization technique. Compared to regular OLSs, they are more resistant to outliers and deviations from normality, and these methods can produce less biased estimates of parameters when the model errors are heteroscedastic (Roy, 1995; Sainani, 2012). The tests were conducted using the Robustbase package in the R environment. The use of robust regression techniques such as the iterative reweighted least squares method is well established in the literature and has been shown to produce reliable and less biased results (Box & Watson, 1962; Ganguly, 2014; Johnston & Faulkner, 2021). They also do not produce typical goodness-of-fit tests, such as F statistics (Ou, Hall, Rodney, & Stremler, 2022; Western, 1995). Robust standard errors were reported for the models.

Results

After the data were completely cleaned (see supplementary material), the final sample was obtained (n = 398). As shown in Table 1 (See Appendix A), the sample (Age_median = 20 years) was largely composed of sophomores (45%), women (67%), and predominantly White (70%). Nearly half (48%) identified as somewhat liberal. The means, standard deviations, and correlations for the main variables are shown in Table 2 (see Appendix B).

The Relationship Between Media Exposure and Climate Anxiety

The first research hypothesis predicted that media exposure (frequency of media use/attention to climate change news) is related to climate change anxiety. As shown in Table 3 (Appendix C), media exposure variables explained approximately 33% of the variance in climate anxiety (R² = 0.333, SE =.458), with the frequency of media use (β =.05, p < .001) and attention given to climate change news (β =.34, p < .001) significantly predicting climate anxiety. When another robust regression was fit, controlling for political affiliation, gender, education, ethnicity, and age, the model improved, explaining approximately 39% of the variance in climate anxiety (R² = 0.387, SE =.458). The frequency of media use (β =.05, p < .001) and attention given to climate change news (β =.24, p < .001) significantly predicted climate anxiety. Political affiliation, level of education, and gender were significant covariates in the model.

Curvilinear Relationship Between Climate Change Anxiety and Pro-Environmental Behavioral Intentions

A polynomial term to the second degree (quadratic term) was also created for climate anxiety (Cohen, Nahum-Shani, & Doveh, 2010; Jackofsky, Ferris, & Breckenridge, 1986) to examine its curvilinear relationship with pro-environmental behavioral intentions. Climate change anxiety and its quadratic term explained approximately 30% of the variance in pro-environmental behavioral intentions (R² = 0.299, SE = .681). When the model included political affiliation, gender, and age as covariates, the model explained 40% of the variance in pro-environmental behavioral intentions (R² = 0.40, SE = .633). As Table 4 (Appendix D) shows, the climate anxiety quadratic term was significant and negative. These findings suggest a nonlinear relationship between climate anxiety and pro-environmental behavioral intentions such that increases in climate anxiety predict more pro-environmental behavior but only up to a point after which the relationship becomes negative. Figure 1 (see Figure document) shows the change in the slope of the relationship between climate change anxiety and pro-environmental behavioral intentions.

Figure 1. Curvilinear relationship between climate anxiety and pro-environmental behavioral intentions.

Efficacy Beliefs and Pro-Environmental Behavioral Intentions

The second hypothesis predicted that college students’ a) self-efficacy and b) collective efficacy beliefs would positively predict adaptive climate change behavioral intentions. A robust regression model was fitted without covariates. Self-efficacy and collective efficacy beliefs explained approximately 20% of the variance in pro-environmental behavioral intentions. As Table 5 (Appendix E) shows, when the model was fitted while controlling for political affiliation, education, gender, ethnicity, and age, it accounted for approximately 38% of the variance in pro-environmental behavioral intentions (R² =.382, SE = .636). Self-efficacy (β = 0.20, p < .05) and collective efficacy (β = 0.13, p < .05) were found to be significant predictors of pro-environmental behavioral intentions. Political affiliation was the only significant covariate in the model.

Discussion

Climate change anxiety — particularly among young people in colleges — has become a topic of scholarly attention in the U.S. While there has been an implicit assumption that climate change anxiety is a pathological phenomenon that requires eradication through suggested coping interventions, the study finds support for both adaptive and maladaptive levels of climate change anxiety. At lower and moderate levels, climate change anxiety might predict positive action intentions; at higher levels, however, individuals might exhibit climate change anxiety at maladaptive and therefore problematic levels. Crandon and colleagues agreed that while maladaptive levels of climate change anxiety could result in cognitive and functional impairment, climate change anxiety should not be pathologized because some levels might be beneficial for climate change action (Crandon, Scott, Charlson, & Thomas, 2022). This is based on the argument that climate change anxiety could be classified as a rational response to a global environmental threat and that the role of the media is to mediate representations of realities in a way that resonates with public consciousness (Crandon, Scott, Charlson, & Thomas, 2022). Hence, as young people become more concerned about the mediated risks of climate change, there are likely levels of anxiety that might drive them to take preventive and adaptive actions. However, an overwhelming level of anxiety might then cognitively and functionally impair them from taking action (Clayton & Karazsia, 2020).

While the findings indicate that extreme climate anxiety may hinder intentions to engage in pro-environmental action, one confounding variable could be generalized anxiety. As Brulle, Carmichael, and Jenkins (2012) found, individuals with heightened climate anxieties tended to have more generalized anxiety as well. Hence, it could be the case that a preexisting anxiety disorder may exacerbate climate worries over time. Due to the limitation of this study as a cross-sectional panel, future studies could inform on how climate anxiety and generalized anxiety may fuel one another bidirectionally in relationship to climate action intentions.

The findings also revealed that those who paid more attention to and were frequently exposed to climate change information were likely to have increased anxiety. These results resonate with a body of empirical research that consistently underscores the pivotal role of media exposure in shaping climate change anxiety levels (Brulle, Carmichael, & Jenkins, 2012; Maran & Begotti, 2021). These findings seem logical given that climate change information in the media tends to be negatively biased with a sensationalist emphasis on threats and impacts (Harrison, Macmillan, & Rudd, 2020; Hart & Feldman, 2014). This further supports the socio-ecological perspective on climate anxiety postulated by Crandon and colleagues, who argued that media exposure is one of several factors that influence climate anxiety (Crandon, Scott, Charlson, & Thomas, 2022). It also lends credence to the agenda-setting perspective that the media can influence the public’s attention and perception of how salient certain issues are by selecting and emphasizing what information to present. This, in turn, can affect the mental state of the audience (Andrews & Caren, 2010).

Furthermore, relevant dimensions of efficacy, such as the collective perspective, have received less attention despite climate change issues typically being viewed as social dilemmas that necessitate collective action (Koletsou & Mancy, 2011). The current study showed that among college students in the U.S., self-efficacy beliefs were more predictive of pro-environmental intentions than collective efficacy beliefs. Exploratory analysis also found no interaction effects between self and collective efficacy beliefs. One explanation could be cultural cognition that emphasizes individualism over group identity and social responsibility. Klassen (2004) suggested that individualistic cultures such as the United States have an independent view of self, with emphasis on “consciousness, independence, individual initiative, and right to privacy” (p. 208). Hence, they are likely to emphasize their self-efficacy rather than their collective efficacy. However, collective efficacy might be common in collectivistic cultures such as those of countries in the Global South and Asia, which are more likely to emphasize “We” consciousness with a propensity toward communality, interdependence, and group identity. It is worth mentioning that pro-environmental behaviors can be categorized into two distinct spheres: public and private (Miller, Rice, Gustafson, & Goldberg, 2022). This division can help elucidate certain distinctions between self-efficacy and collective efficacy, as public sphere behaviors often necessitate collective action and coordination. However, the present study relied on previous empirical works to merge both public and private sphere behaviors to provide a comprehensive overview of pro-environmental behavioral intentions (Huang, 2016; Miller, Rice, Gustafson, & Goldberg, 2022). This dynamic becomes an empirical question that could be further explored to understand how the different dimensions of efficacy beliefs differ in their relationships with different dimensions of pro-environmental behaviors.

An unsurprising finding was the predictive utility of political affiliation; liberal leaning respondents tend to show more climate anxiety and were likely to engage in pro-environmental action. This finding also consistently aligns with the literature that political ideology influences climate change cognitive and behavioral outcomes (Huang, 2016; Wullenkord, Tröger, Hamann, Loy, & Reese, 2021).

Limitations

This study also has limitations. In addition to the cross-sectional nature of the study, a nonrandom sampling technique was employed, which might limit the generalizability of the findings. However, the homogeneity of the population (college students) might allow important insights to be drawn from a study that could apply to the U.S. college student population, as studies have found convenience sampling suitable for such populations (Etikan, Sulaiman, & Rukayya, 2016). Nonetheless, future studies should attempt to generalize the findings by drawing a representative sample of college students using an appropriate sampling frame to investigate these relationships. For transparency, it is important to note that these issues should be taken into strong consideration when interpreting the findings of this study.

Practical Implications

This study has implications for academic and public discourse about climate change anxiety, media coverage of climate change, and climate communication in general. First, it contributes to understanding climate change anxiety by being among the few research studies to explicitly examine the curvilinear relationship between climate change anxiety and pro-environmental behavioral intentions. Instead of placing all levels of climate change anxiety in the same bucket as negative and problematic for action, it is more productive to understand the distinction between the beneficial levels of the phenomenon and the extent to which it stops becoming beneficial. As Crandon and colleagues succinctly put it, “of particular importance is to distinguish rational, adaptive responses to climate change (for example, climate distress/climate empathy) from clinically maladaptive climate anxiety (for example, clinical climate anxiety/adjustment disorder)” (Crandon, Scott, Charlson, & Thomas, 2022, p. 129).

The study also implies the potential roles of self and collective efficacy beliefs in mitigating extreme levels of climate anxiety, which was shown to impair pro-environmental actions. A robust sense of efficacy not only boosts self-esteem but also guards against vulnerability, potential mental health disorders, and stress (Innocenti et al., 2023). While extreme levels of climate change anxiety may impair action, focusing on ways to enhance perceived self and collective efficacy may drive individuals to seek ways to mitigate the negative impact on their mental health as well as engage in pro-environmental actions.

Informational sources such as the news media must also pay attention to how coverage of climate change impacts public well-being. The media should be contributing to the multidimensional steps taken in addressing both the direct and indirect impacts of climate change, especially among young people. Hence, there is a need to reevaluate the approach of climate change news coverage by scrutinizing the negativity, and sensationalism with which climate change news is covered. In this way, news can contribute to addressing the varying impacts of climate change on the planet’s health and the people involved.

Acknowledgments

The author would like to thank Dr. Senyo Ofori Parku, Shane Burrell, and Dr. Sara Weston for their support and feedback on this manuscript.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Data availability Statement

The data that support the findings of this study are available in the OSF repository (https://osf.io/pndaw/?view_only=19e04c842d8645a5be77619845b6aaa2)

Supplemental Data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10810730.2024.2354370.

Appendix

A

Table A1. Descriptive table of demographics

Characteristic	N = 398
Education
Sophomore	179 (45%)
Junior	157 (40%)
Senior	47 (12%)
Freshman	13 (3.3%)
Graduate/Professional Degree	1 (0.25%)
Preferred to Not Answer	1 (0.25%)
Gender
Woman	265 (67%)
Man	117 (29%)
Prefer to self-describe	5 (1.3%)
Transgender	5 (1.3%)
Non-binary/third gender	4 (1.0%)
Prefer not to say	2 (0.5%)
Race/Ethnicity
White	277 (70%)
More than one	40 (10%)
Asian/Pacific Islander	30 (7.5%)
Hispanic/Latin(o/a/x)	28 (7.0%)
Black/African-American	15 (3.8%)
Other	6 (1.5%)
Native American/American Indian	2 (0.5%)
Political Affiliation
Somewhat Liberal	189 (48%)
Very Liberal	110 (28%)
Centrist	65 (16%)
Somewhat Conservative	24 (6.1%)
Very Conservative	7 (1.8%)
Preferred To Not Answer	3

Appendix B

Table A2. Means, standard deviations, and zero-order correlations with confidence intervals

Variable	M	SD	1	2	3	4	5
1. Attention to Media	2.73	0.94
2. Frequency of Media Use	4.45	1.73	.15**
			[.05, .24]
3. Climate Anxiety	2.04	0.60	.52**	.23**
			[.44, .59]	[.13, .32]
4. Self-Efficacy	2.95	1.05	.30**	.19**	.26**
			[.21, .39]	[.10, .29]	[.16, .35]
5. Collective Efficacy	3.22	1.00	.25**	.10*	.20**	.72**
			[.15, .34]	[.00, .20]	[.10, .29]	[.67, .76]
6. Pro-environmental Behavioral Intentions	3.43	0.82	.53**	.15**	.50**	.40**	.38**
			[.46, .60]	[.05, .25]	[.42, .57]	[.31, .48]	[.29, .46]

Note. M and SD are used to represent mean and standard deviation, respectively. α; Cronbach’s Alpha. Values in square brackets indicate the 95% confidence interval for each correlation. * indicates p < .05. ** indicates p < .01. Frequency of media use was measured on an eight-point scale. All other variables are on five-point scales.

Appendix C

Table A3. Robust linear regressions with climate anxiety as the dependent variable

	Model 1 w/o covariates	Model 2 with covariates
Predictors	Standardized beta	Standardized beta
(Intercept)	−0.03	−0.92
Frequency of Media Use	0.05 **	0.05**
Attention to News	0.34**	0.29**
†Political Affiliation		−0.09**
†Education Levels		−0.11**
Age		0.05
Ethnicity (White = 1)		0.01
Gender (Female = 1)		0.20**
R^2	0.333	0.387
R^2 adjusted	0.330	0.376
Robust SE	0.458	0.458

Note. * indicates p < .05. ** indicates p < .01.

†Political Affiliation and Education Levels were measured as interval variables

Appendix D

Table A4. Robust r.egressions with pro-environmental behavioral intentions as the criterion

	Model 1 w/o covariates	Model 2 with covariates
Predictor	Standardized beta	Standardized Beta
(Intercept)	0.16 **	0.32
Climate Anxiety	0.78**	0.61**
+Climate Anxiety (Squared)	−0.36**	−0.22*
†Political Affiliation		−0.31**
Education Levels		−0.02
Age		0.04
Gender (Female = 1)		0.06
Ethnicity (White = 1)		0.09
R^2	0.299	0.402
R^2 adjusted	0.295	0.391
Robust SE	0.681	0.633

Standardized beta is reported, * indicates p < .05. * indicates p < .01. +Political Affiliation was measured as an interval (1=Liberal, 5 = Conservative) + Climate Anxiety squared represents the polynomial (quadratic) term.

Appendix E

Table A5. Robust regressions with pro-environmental behavioral intentions as dependent variable

	Model 1w/o covariates	Model 2with covariates
Predictors	Standardized beta	Standardized beta
(Intercept)	2.24**	3.25**
Self-Efficacy	0.20**	0.20**
Collective efficacy	0.19**	0.13**
Ethnicity (White = 1)		0.13
Age		−0.01
Gender Binary (Female = 1)		−0.00
Education Levels		−0.00
†Political Affiliation		−0.36**
R^2	0.204	0.369
R^2 adjusted	0.199	0.358
Robust SE	0.713	0.644

Note. Standardized betas are reported * indicates p < .05. ** indicates p < .01. SE= Standard Error

†Political Affiliation and Education Levels were measured as interval variables