Next Article in Journal
Why Behavioral Steering Falls Short: Agency, Practice, and Virtue Ethics in Sustainable Consumption
Previous Article in Journal
Correction: Alinejad et al. A Charge/Discharge Plan for Electric Vehicles in an Intelligent Parking Lot Considering Destructive Random Decisions, and V2G and V2V Energy Transfer Modes. Sustainability 2022, 14, 12816
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 18
Issue 4
10.3390/su18041826
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Experimental Tests of an Affective Pathway to Pro-Environmental Intentions

by
Johannes Klackl
1,*,
Isabella Uhl-Hädicke
1,*,
Tilmann Fowles
2 and
Eva Jonas
1
1
Department of Psychology, University of Salzburg, 5020 Salzburg, Austria
2
Applied Social Science, Dortmund University of Applied Sciences and Arts, 44227 Dortmund, Germany
*
Authors to whom correspondence should be addressed.
Sustainability 2026, 18(4), 1826; https://doi.org/10.3390/su18041826
Submission received: 21 November 2025 / Revised: 30 January 2026 / Accepted: 4 February 2026 / Published: 11 February 2026
(This article belongs to the Section Psychology of Sustainability and Sustainable Development)
Browse Figures
Versions Notes

Abstract

Empirical evidence suggests that affective responses to climate change are predictive of pro-environmental attitudes, intentions, and behavior and may thus contribute to sustainability. However, that evidence is mostly correlational. We present experiments designed to test the mediational hypothesis that exposure to climate change information elicits negative affective responses, which in turn increase pro-environmental intentions. In Study series 1 (8 studies, total n = 819), we used a measurement-of-mediation design. We provided participants with information about the devastating effects of climate change or other non-threatening information. Neither providing that information nor the negative affective responses to that information were associated with participants’ subsequent pro-environmental intentions. In Study 2 (n = 135), we employed a manipulation-of-mediator design, which was a more stringent test of mediation. We asked people to either intensify or suppress their feelings while they were exposed to information about climate change or information unrelated to climate change. Deepening one’s feelings increased negative affect and reduced positive affect but did not significantly raise pro-environmental intentions. Together, these results suggest that climate change information reliably produces affective responses, but these do not translate into pro-environmental intentions in the short term.

1. Introduction

People around the world are increasingly concerned about environmental degradation caused by global ecological problems, such as climate change. A number of similar psychological constructs, such as climate anxiety [1,2], eco-anxiety [3,4,5], global warming worry [6], and climate distress, have been proposed to capture the cognitive, affective, and behavioral responses to the worsening state of the biosphere. A multinational study of 10,000 children and young people found that the majority of them experienced feelings of sadness, anxiety, anger, powerlessness, helplessness, and guilt [7]. These results are supported by other studies examining more general negative feelings or worries about climate change [8,9,10]. While not necessarily causal, these feelings can be accompanied by somatic symptoms [11], poor sleep quality, and poor mental health [12], although the prevalence of these impairments is generally low [2,13].
A central question in this context is whether strong emotional engagement with climate change can catalyze both individual and collective action against climate change and thus contribute to sustainability. While some studies found exposure to media on climate change to be positively related to pro-environmental intention [14] and behavior [15,16], it was also found to be negatively related to some kinds of pro-environmental behavior, such as protest participation intention [17].
A closer examination of the evidence reveals that most studies on the link between affective responses to climate change and pro-environmental behavior rely on correlational designs, which are unsuitable for testing causal relationships. It thus remains unclear whether affective responses to environmental problems drive pro-environmental behavior, whether the reverse is true, or whether both affect and behavior are guided by other variables, such as environmental values or identity [18]. Experimental research is necessary to establish the causal links between people’s affective and behavioral responses to climate change.
Experimental studies suggest that anxious responses to climate change information can elicit negative affect and increase pro-environmental behavior intentions [19,20,21], whereas others have not found such effects [22]. To summarize, in the domain of climate change, it is unclear whether affective responses to climate change play a key role in promoting pro-environmentalism.
In the studies reported here, we aimed to test the mediational hypothesis that exposure to climate-related information would lead to negative affective responses that would then carry forward into pro-environmentalism. In Study series 1, we employed a standard experimental design to probe mediation, known as the measurement-of-mediation design. We provided participants with either information about the devastating effects of climate change or various control conditions and investigated whether the resulting affective responses mediated subsequent pro-environmental intentions [23]. In Study 2, we employed a manipulation-of-mediator design. In addition to manipulating exposure to climate change information, we manipulated affect itself by asking people to reflect on or suppress their feelings. The idea is that experimentally strengthening negative affect, which is the alleged mediating process, should amplify pro-environmental responses.

2. Study Series 1

In Study series 1, our goal was to explore whether exposure to negative information about climate change promotes pro-environmental intentions through increased negative affect and/or reduced positive affect, using eight datasets, two of which had already been published [24,25]. The remaining datasets have accumulated in our file drawer over the past decade. In all studies that comprise Study series 1, we manipulated climate change salience by exposing participants to information about the detrimental effects of climate change or to information on climate change-unrelated topics. Subsequently, we measured people’s levels of positive and negative affect, as well as their pro-environmental intentions. The fact that we repeatedly tested the same hypothesis across several similar yet slightly different studies provided a good way to assess the replicability of the effect and its robustness to methodological variation. The large sample size enabled a powerful statistical test of the mediational hypothesis.

2.1. Materials and Methods

2.1.1. Participants

We conducted eight studies with 871 participants (see Table 1 for study-specific information). To increase statistical power, we also added and re-analyzed data from two already-published papers, namely Study 1c [24] and Study 1g [25]. In previous publications, we did not test the mediational hypothesis that climate change salience would promote pro-environmental action through both positive and negative affect. Thus, the data from these studies were re-analyzed to test a novel hypothesis. The sample sizes of the studies were not predetermined. Participants of all studies except Study 1c were students at the University of Salzburg. Study 1c was an intercultural study conducted in Austria and Argentina, drawing on more diverse samples. The ethics review board of the University of Salzburg approved each study.

2.1.2. Design

Across all studies in Study series 1, we manipulated climate change salience by confronting participants with a list of scientific findings about the devastating effects of climate change (e.g., “Some research results suggest a sea level rise of at least half a meter up to two meters by the year 2100. Within 300 years, a rise of 2.5 m to 5.1 m is possible.”), or with a list of scientific findings with no mention of climate change (e.g., “The surface of the Earth is approximately 70.7% water and 29.3% land.”). We will refer to these groups as the climate change—salient (CCS) group and the climate change—not salient (CCNS) group. The number and wording of the findings differed across studies (see Table S1 in the Supplementary Materials for a complete list of the items used in each study). To increase participants’ engagement with the material, we asked them to indicate whether they were already aware of each finding by placing a check mark next to it. In both conditions, we asked participants to write down the three pieces of information they found most important. This was done to prevent people from merely skimming the texts, to increase elaboration, and ensure that participants processed the information more deeply, thereby strengthening the manipulation.
Some studies posed exceptions to these general rules. In Study 1d, participants were subjected to one of three conditions: a condition that highlighted the local (spatially close) effects of climate change, a condition that highlighted its global effects, or a control condition in which participants read about the latest fashion colors. Negative affect was higher in the global condition (M = 0.21; SD = 0.15) than in the control condition (M = 0.18; SD = 0.18). In contrast, the local group had the lowest negative affect (M = 0.17; SD = 0.13). Positive affect was highest in the control condition (M = 0.45; SD = 0.18), somewhat lower in the global group (M = 0.43; SD = 0.16), and lowest in the local group (M = 0.38; SD = 0.13). Despite the differences between the local and global conditions, we combined and contrasted them with the control condition. Another particularity of Study 1d was that the scientific facts were presented as a text rather than a list. Participants were only required to list the three most essential findings, rather than indicating which ones they already knew.
Study 1e also deviated from the norm, since the climate change salience manipulation either highlighted the temporally close (15 years from now) or the temporally distant (75 years from now) effects of climate change. In the control condition, participants read about global geographical facts. Negative affect was higher in the distant condition (M = 0.29; SD = 0.15), followed by the close condition (M = 0.26; SD = 0.14) and the control condition (M = 0.21; SD = 0.15). Positive affect was highest in the control condition (M = 0.39; SD = 0.16), similarly high in the distant condition (M = 0.39; SD = 0.14), and lowest in the close condition (M = 0.37; SD = 0.14). We combined the temporally close and distant conditions and contrasted them with the control condition.
In Study 1d, there was also a patriotism priming manipulation. At the beginning of the study, one group of participants received information about which things Austrians were most proud of. The list included beautiful landscapes, rich traditions, diverse cuisine, and a high quality of life. The other group received information about the most popular hot drinks in Austria. The list included coffee, hot chocolate, and tea. This manipulation did not significantly affect any of the measures of interest (negative affect: t(147) = 1.749; p = 0.082; positive affect: t(147) = 0.684; p = 0.495; pro-environmental intentions: t(140) = −0.553; p = 0.581).

2.1.3. Procedure

At an early stage of each study, we manipulated the salience of climate change. Immediately following this manipulation, we measured people’s affective responses to the materials they had just been exposed to. Then, depending on the study, we included additional measures before participants filled in the pro-environmental intention scale.

2.1.4. Measures

Manipulation check. Following the climate change manipulation, we asked participants how worried they felt after having been confronted with the stimulus material on a 6-point scale ranging from ‘not at all’ to ‘a lot.’ In Study 1h, we used a 5-point scale.
Positive and negative affect. In each study, we administered the Positive and Negative Affective Schedule scales [26]. The instrument consists of a 10-item positive affect scale (e.g., active, excited, strong) and a 10-item negative affect scale (e.g., distressed, upset, scared).
Pro-environmental intentions. In each study, we assessed participants’ pro-environmental intentions using different variants of the SEU-3 [27], e.g., “I am (still) willing to bring my vacation habits more in line with environmental protection (e.g., avoiding long-distance flights, gentle tourism)”. The scale has seven content areas—energy saving, societal commitment, waste separation and recycling, sports and leisure, shopping, environmentally friendly transportation, and water conservation—all of which represent ways to reduce one’s contribution to climate change. In addition to intentions, the SEU-3 also includes items to measure pro-environmental attitudes (e.g., “I think it is unreasonable to ask people to check their electricity and gas consumption regularly,” reverse-coded) and behavior (e.g., “In my free time I use the car, for example for trips, short vacations, visits or trips to leisure activities,” reverse-coded). We only used the items designed to measure pro-environmental intentions to build the scale. Any reverse-coded items were carefully recoded. Study 1c deviated from the norm by using a custom 12-item pro-environmental intention scale [24]. Table S2 in the Supplementary Materials provides an overview of which pro-environmental intention items were used in each study.
We excluded three items from the “environmentally friendly transportation” content area because of low item-scale correlations and/or inconsistent loadings (“Even if public transportation were better and cheaper than driving, I would prefer the car” [reverse-coded], “If I know that I will have to wait a long time in front of a red light, construction site or railroad crossing, I will (continue to) turn off the engine”, “In the future, when I am traveling by motor vehicle, I will (continue to) turn off the engine during long stops at traffic lights and during traffic jams”). We believe that because the participants in Study series 1 were mostly young and many were students, many did not own a car. Furthermore, we excluded the item “I would by no means be willing to pay more money for my drinking water, even if the additional proceeds are used to finance measures against increasing drinking water pollution” due to inconsistent loadings. The poor performance of this item could be due to the fact that, in Austria, most drinking water is sourced from groundwater and springs and is of good quality despite minimal treatment.
Furthermore, we conducted two measurement invariance analyses to determine whether the items in both variants of the SEU-3 captured pro-environmental intentions similarly across studies (see Supplementary Materials). The items of the short SEU-3 version loaded positively on the pro-environmental intentions construct, suggesting configural invariance. However, metric or scalar invariance was not achieved. The measurement invariance analysis of the long version of the SEU-3 indicated metric invariance but not scalar invariance.
Additional measures. In Study 1e, we administered the Behavior Identification Form [28] and a self-efficacy scale [29]. In Study 1d, we administered the Behavior Identification Form [28], a scale that asked participants to evaluate a list of 44 social groups [30], a self-efficacy scale [29], and a prototype scale measuring the personal importance of environmental action. In Study 1b, we administered a 16-item measure of hedonic, egoistic, altruistic, and biospheric values [31] and a three-item environmental self-identity scale [18]. At some point in each study, we asked participants to indicate their nationality, occupation, field of study (if applicable), marital status, place of residence, highest educational attainment, and the number of children they had and planned to have. These measures are not considered for our hypotheses and are therefore not reported in this paper.

2.1.5. Data Analysis

We used R version 4.5.2 [32] to analyze the data. We used a variety of Likert scales to measure affect and pro-environmental intentions across the many studies and measures reported in this paper; for example, some ranged from 1 to 5, while others ranged from 1 to 7. To increase the comparability of scores across studies, we rescaled them so that the smallest possible score corresponded to 0 and the highest possible score corresponded to 1, using the rescale function of the car package [33]. In addition, we coded the climate change salience manipulation with 1 (climate change—salient) and 0 (climate change—not salient).
First, we examined the bivariate relationships among all measures as well as the effect of the climate change salience manipulation on each measure (see Table 2). To account for the multi-study structure of the dataset, we used multilevel correlations [34].
To test the mediational hypothesis that experimentally manipulated climate change salience increases pro-environmental intentions via positive and negative affect, we specified a multilevel mediation model using lavaan [35], with observations nested within studies. The model decomposed effects into within-study (Level 1) and between-study (Level 2) components by specifying separate model blocks for each level and using study as the clustering variable. All variables were treated as continuous. This multilevel specification corresponds to a random-intercept model for study. We modeled the climate change salience manipulation (1 = climate change—salient; 0 = climate change—not salient) as the initial predictor, positive and negative affect as parallel mediators, and pro-environmental intentions as the outcome. Due to non-normality in the distributions of negative affect and pro-environmental intentions and in the residuals, we fitted the model using robust maximum-likelihood estimation to obtain robust standard errors and scaled test statistics under violations of normality assumptions.
We initially planned to use a more complex model that allowed not only random intercepts for each study but also the slopes of the regression paths to vary across studies, which resulted in a singular fit. The correlations among multiple random effects were +/− 1. We interpreted this as evidence that such a model was too complex and opted for a simpler one. We simplified the model in a stepwise fashion, as recommended by Barr et al. [36]. By allowing only the intercepts of the fixed effects (but not their slopes) to vary across studies while not allowing their slopes to do so, we achieved a non-singular fit. We did not estimate the standard errors of the indirect effect using a resampling-based method such as bootstrapping due to a lack of consensus on how to implement such procedures in multilevel models.
To examine whether any mediation effects may have been present in some of the studies in Study series 1, we conducted a series of study-specific fixed-effect analyses (see Table S6 in Supplementary Materials).

2.2. Results

We first investigated the differences between the climate change—salient (CCS) and climate change—not salient (CCNS) groups regarding affect and pro-environmental intentions. The CCS group reported significantly higher negative affect but did not differ significantly regarding positive affect and pro-environmental intentions (see Table 2). Multilevel correlations between the affect and pro-environmental intention measures indicated that positive and negative affect were weakly, but positively, related to each other (see Table 2). The correlation between PA and NA did not differ markedly between the CCS (r(429) = 0.20, p < 0.001) and CCNS conditions (r(387) = 0.15, p = 0.004). Moreover, positive, but not negative, affect was weakly and positively related to pro-environmental intentions.
Next, we interpreted the results of our multilevel mediation meta-analysis across studies to test whether climate change salience increases pro-environmental attitudes by increasing negative affect and/or reducing positive affect. Variance decomposition indicated substantial variability within and between studies. The ICC suggested that roughly 10% of the total variance in pro-environmental intentions was attributable to between-study differences (see Table 3).
At the within-study level, CCS was positively associated with negative affect. However, negative affect was not associated with PEI. The resulting within-study indirect effect was not significant. The meta-analysis across Study series 1 revealed no significant indirect, direct, or total effects (see Figure 1). For a complete description of the model results, see the Supplementary Materials.

2.3. Discussion

In Study series 1, we relied on data from a series of similar experiments to test whether exposure to climate change information and the negative affective responses resulting from that exposure would promote pro-environmental intentions. The large sample size enabled a powerful statistical test of the mediational hypothesis. Across these experimental studies, we found that informing people about the impact of climate change increased negative affect and reduced positive affect but did not change pro-environmental intentions. Moreover, people’s affective responses did not mediate the effect of climate change salience on pro-environmental intentions. These findings are based on similar, yet slightly different, studies.
We observed a small positive correlation between positive affect and negative affect. This finding is unusual, as previous research has typically reported small negative correlations between these constructs [26,37]. The magnitude of the correlation did not differ substantially between the climate change—salient and climate change—non-salient conditions, suggesting that the positive correlation was not attributable to the experimental manipulation. Rather, the results may indicate affective ambivalence among some participants. This possibility should be examined in future research.
The materials used in the climate change—salient and climate change—non-salient groups varied across studies. In some studies, participants were presented with information about global climate change versus general facts about planet Earth. In others, they received information about local climate change consequences versus local geographical facts. In one study, a text about the latest fashion colors served as the control material. These differences introduced additional between-study variance in the constructs and effects of interest. Moreover, some materials in the non-salient conditions may have elicited mild positive affect. Consequently, it remains unclear whether the observed reduction in positive affect following climate change salience reflects a genuine response to the climate related information or whether it is partly an artifact of one or more control conditions. Nonetheless, when each study was analyzed separately, none showed evidence of a significant mediating effect of positive or negative affect (see Supplemental Table S6 in Supplementary Materials). This suggests that the overall null mediational effect is unlikely to be attributable to these methodological variations.

3. Study 2

In Study 2, we performed an additional test of the hypothesis that exposure to climate change information would promote pro-environmental intentions via affective responses. This time, we manipulated not only the salience of climate change but also people’s negative affective responses to it. We achieved this by instructing some participants to engage deeply with the climate change-related information, while instructing others to process it superficially, without becoming emotionally involved [38]. We expected that extensive contemplation of climate change and its significance for one’s life would elicit stronger affective responses [39]. If negative affective responses lead to pro-environmental intentions, one would expect participants who were exposed to climate change information and instructed to get emotionally involved to report the highest pro-environmental intentions.
Furthermore, unlike Study series 1, which used text-based manipulations to make the issue of climate change salient, Study 2 relied on short film clips taken from a documentary on climate change. In general, motion pictures elicit stronger affective responses than text messages [40]. In combination with the affective intervention, this study design aimed to better detect a possible effect of climate salience on PEI via negative affective states.
Additionally, we not only measured participants’ pro-environmental intentions but also asked them how successful they were at translating those intentions into actual behavior by surveying them again two weeks after the experiment. This enabled us to determine whether the climate change salience and reflection manipulations affected actual pro-environmental behavior, and whether affect would mediate these effects.

3.1. Materials and Methods

3.1.1. Participants

As this intervention was not used in previous studies, there were no reference effect sizes. Therefore, we sampled as many participants as possible in the given time frame. N = 135 participants took part in our study. Given the 2 × 2 study design, a power analysis using G*Power 3.1.9.7 revealed that this resulted in 82% power to find a medium-sized effect [41]. The sample consisted mainly of university students. Psychology students received course credits for participating. We excluded one participant from the climate change—salient group because they seemed unaware that the video they had seen was about climate change. The mean age of the final sample (n = 134) was 24.54 years (SD = 6.43). A total of 89 participants identified as female, and 45 identified as male. The ethics review board of the University of Salzburg approved the study.

3.1.2. Design

Study 2 employed a 2 (reflection vs. suppression) × 2 (climate change salience) design. Consequently, participants were randomly assigned to one of the four possible combinations of these conditions: climate change—salient/reflection (n = 36), climate change—salient/suppression (n = 23), climate change—not salient/reflection (n = 34), climate change—not salient/suppression (n = 41).
Climate change salience. To manipulate climate change salience, we used two 3-min video sequences from the documentary HOME. The film illustrates the human impact on large-scale planetary ecological systems through innovative aerial shots. Participants in the climate change—salient (CCS) group watched a video clip highlighting the negative impacts of climate change on the biosphere. Participants in the climate change—not salient (CCNS) group watched a similar video clip that did not mention climate change. The movie is freely accessible because it is part of Creative Commons. The videos are available at https://www.dailymotion.com/video/x72xadm (accessed on 1 July 2025) and https://www.dailymotion.com/video/x72xado (accessed on 1 July 2025). We used a multiple-choice question to see whether people thought the video was about and provided six possible answers (“about the dangers of climate change for humanity”; “about the beauty of nature”; “about the cradle of life”; “about the natural cycle”; “about the interaction of different living beings on earth”; “nature in general”). We excluded one participant in the climate change—not salient group who did not select the correct answer, “about the dangers of climate change for humanity”.
Reflection vs. suppression. Following Roth and colleagues [38], we provided participants with different instructions for managing their emotions while watching a subsequent video. In the reflection condition, participants were informed that we were primarily interested in the feelings and thoughts evoked by the specific video content. We instructed them to reflect on how the things portrayed in the film would affect their lives. In the suppression condition, we informed participants that we were primarily concerned with the technical aspects of the film clip and less concerned with its content. We asked them to watch the video in an unemotional, detached way and to pay special attention to technical aspects (color, sound, sharpness, etc.).

3.1.3. Procedure

We disguised the study as a “Nature video evaluation study”. After arriving in the lab, participants took place in front of a computer that presented the experimental material and measures. Participants completed the study at their own pace. First, they were required to provide information about their age, gender, nationality, education, current profession, and political orientation. Next, they completed the Biospheric Value Measure [31] and the Schwartz Value Survey [42]. Afterwards, we manipulated reflection vs. suppression by providing participants with one of two instructions for watching the video (reflect or suppress). Then, participants moved on to the video, which either highlighted the threatening consequences of climate change or did not. To watch the video, participants had to switch to full-screen mode and put on headphones.
After watching the video, we administered measures of how much people worried in response to the video and to what extent they successfully followed the reflection (or suppression) instruction. We also gave participants a multiple-choice comprehension question to ensure that they understood the video was about climate change. Finally, participants received the pro-environmental intentions scale [27]. Two weeks after the lab session, we invited participants via email to an online survey that included a follow-up measure of pro-environmental behavior.

3.1.4. Measures

Negative affect. To measure the degree to which participants reported negative affect in response to watching the videos, we asked participants to self-report their feelings using the following statements: “because of the video I feel…” “worried”, “anxious”, “uncertain”, “frightened”, and “threatened” (α = 0.94). We added additional positive and neutral adjectives (amused, cheerful, interested, lively, surprised, serene) to balance the scale’s overall emotional valence. We used a 7-point Likert scale, ranging from “completely disagree” to “completely agree.”
Pro-environmental intentions. We included 13 items from the SEU-3 [27] (e.g., I will (continue to) leave the car in the future if I can use a bus, train, or bicycle instead; α = 0.78). We excluded one item that negatively correlated with the scale mean. We used a 7-point Likert scale, ranging from “completely disagree” to “completely agree.”
Pro-environmental behavior. For the follow-up measurement, we used a modified version of the SEU-3 [27] that included all relevant pro-environmental behaviors that could be performed in two weeks (e.g., “In the past two weeks left the car behind if it was possible to take the bus, train or the bicycle”; α = 0.65). We used a 7-point Likert scale ranging from “completely disagree” to “completely agree”. A total of 54 of 134 participants completed the follow-up measure.
Additional measures. We also administered the Biospheric Value Orientation Scale [31] (“Respecting the Earth”, “Protecting the environment”, “Unity with nature”, and “Preventing pollution”), and the Schwartz Value Survey [42]. We administered the Purity subscale of the ethnocentrism scale [43]. This facet of ethnocentrism is mainly concerned about the rejection of outgroup members to benefit the ingroup, e.g., Our culture would be much better off if we could keep people from different cultures out (α = 0.83). We used a 7-point Likert scale ranging from “completely disagree” to “completely agree”. These measures are not relevant to our hypotheses and are therefore not reported in this paper.

3.1.5. Data Analysis

In Study 2, we used the lavaan package [35] to specify a moderation model. To investigate attrition in the follow-up measure of pro-environmental behavior, we compared completers and non-completers based on negative affect (t(132) = −0.78; p = 0.440; 95% CI [−0.06; 0.03]), pro-environmental intentions (t(132) = −0.78; p = 0.440; 95% CI [−0.06; 0.03]), pro-environmental intentions (t(132) = −0.53; p = 0.598; 95% CI [−0.07; 0.13]) and experimental manipulation (χ2(3, N = 135) = 2.16; p = 0.540). The results revealed no evidence of an attrition bias. Furthermore, we utilized maximum-likelihood estimation to retain as much information as possible in the model. Model diagnostics revealed no clear signs of heteroskedasticity or non-normality of residuals.
We modeled the climate change salience (CCS) manipulation (0 = climate change—not salient; 1 = climate change—salient) as the predictor, pro-environmental intentions as the outcome, and the reflection manipulation (0 = suppression; 1 = reflection) as a moderator. Pro-environmental intentions and pro-environmental behavior were modeled as outcomes. As in Study series 1, we rescaled the continuous variables (affect, pro-environmental intentions, and pro-environmental behavior) so that the lowest possible value was 0, and the highest possible value was 1.

3.2. Results

First, we investigated the zero-order correlations between the variables we measured. Negative affect was positively associated with pro-environmental intentions. Pro-environmental intentions and behavior were positively correlated (see Table 4).
As in Study series 1, instructing people to reflect on (rather than suppress) their feelings increased negative affect and reduced positive affect (see Table 4 and Figure 2), indicating that the reflection manipulation was effective. In addition, as expected based on Study series 1, participants reported higher negative affect and lower positive affect after the climate change video compared with after the nature video, as reflected in a main effect of climate change salience (CCS). Confirming the findings of Study series 1, climate change salience had no significant effect on pro-environmental intentions or pro-environmental behavior. Notably, the two manipulations exerted no interactive influence on any of the measures. That is, the climate change video led people to experience more negative affect but not to report higher pro-environmental intentions, thus providing no support for the hypothesis that negative affect mediates the effect of salient climate change on pro-environmental intentions or behavior. To summarize, watching a video about climate change and reflecting on one’s feelings led to more negative affect but not to more pro-environmental intentions or behavior.

3.3. Discussion

Study 2 had two goals: The first was to replicate the central finding of Study 1, albeit with a video-based manipulation rather than a text-based one. A climate change video increased negative affect. The second goal was to determine whether inducing additional negative affect would enhance the climate change video’s ability to boost pro-environmental intentions. This was not the case. Although instructing people to reflect on their feelings increased their negative affect and decreased their positive affect, it did not increase their pro-environmental intentions. The third goal of Study 2 was to see whether climate change salience would also translate into actual behavior. This was not the case.

4. General Discussion

We set out to test whether negative affective responses to climate change would drive pro-environmental action. The hypothesis is based on the finding that people who behave pro-environmentally report more negative emotional responses to climate change [2,6,12]. If these negative emotional responses play a causal role in how climate change awareness leads to climate action, using reminders of climate change to provoke such negative affective responses should (at least temporarily) result in higher pro-environmental intentions. In Study series 1, we tested this hypothesis in various situations by manipulating climate change salience and measuring positive and negative affect after people received the information, as well as their pro-environmental intentions. The extent of their negative (or positive) affective responses did not mediate the effect of climate change salience on pro-environmental intentions, thus providing no support for this mediation hypothesis. In Study 2, we conducted another test of mediation by asking people to reflect on (rather than suppress) their negative affective responses, in addition to manipulating climate change salience. Reminders of climate change did not affect pro-environmental intentions, regardless of whether people’s affective experience was experimentally intensified through reflection. Together, these results indicate that although reminders of climate change can easily elicit negative affective responses, these responses are not associated with the pro-environmental intentions people report afterwards.
Based on the present findings, it would be premature to conclude that negative affect (or other types of affect) plays no role in the emergence of pro-environmentalism in the face of climate change. This is because there are several alternative explanations for this pattern of results. First, the climate change salience manipulations we used in our studies may be too weak to affect people’s intentions or behavior. Evidence suggests that the relationship between climate anxiety and PEB appears to follow an inverted U-shaped pattern: moderate levels of climate anxiety are most conducive to PEB, whereas very low or very high levels of anxiety may reduce engagement [44,45]. Similarly, the potential of climate change information to change pro-environmental intentions may depend on the amount of negative affect elicited. Second, much pro-environmental behavior is habitual and therefore resistant to change [46]. Even though pro-environmental intentions may be less stable, as they do not necessarily involve awareness of a habit [47], a single exposure to climate change may not be sufficient to cause significant changes. More extensive, repeated exposure to climate change information could yield measurable effects on people’s intentions and behavior. Third, longitudinal studies have provided evidence that negative affect can also follow from pro-environmental behavior, raising doubts about the unidirectionality of the relationship [48,49]. Finally, the absence of evidence for a link between pro-environmental intentions and affect may be due to the nature of affect itself. An affect denotes a positive or negative feeling towards an event or object but is considered less intense and less differentiated when compared to an emotion [50]. Emotions such as anxiety, sadness, anger, guilt, or shame have been described in the context of climate change [51,52]. They are thought to be more strongly associated with cognitive processing, judgment, decision-making, and behavior.
Theory and prior research suggest that humans often fail to respond to climate change with adequate action because they cope with climate change similarly to how they cope with their own mortality: they engage in strategies aimed at managing the discomfort associated with the problem through avoidance, rationalization, or denial of climate change and its consequences [53,54]. Scholars have referred to these strategies as proximal defense. Since climate change can serve as a stark reminder of death and vulnerability, these defenses may partly explain why reminders of climate change do not reliably elicit protective, pro-environmental behavior. While such maneuvers may be adaptive for coping with existential givens like mortality, they are likely counterproductive for tackling climate change. Study 2 empirically tested this hypothesis but did not confirm it; asking people to suppress (rather than reflect on) thoughts of climate change did not affect pro-environmental intentions or behaviors, regardless of whether climate change was experimentally made salient. However, more reliable tests are needed to determine the role of these defenses in climate action.
Regarding the question of whether fear messaging is harmful or beneficial for climate engagement, the present paper’s findings do not suggest that climate change communication should abandon discussing the threatening nature of climate change or avoid negative affect, as proposed previously [55,56]. We found no evidence that providing information about climate change ‘backfired’ and lowered pro-environmental intentions. Even deliberately enhancing negative affect in Study 2 did not affect pro-environmental intentions or behavior. Thus, for constructive engagement with climate change, negative affect is neither helpful nor harmful, although it may potentially cause unwanted mental health side effects [2,13,57].

Limitations

The studies reported here are subject to several limitations. First, the information presented in the various control conditions of Study series 1 and Study 2 was not only less threatening but also unrelated to climate change. Hence, our manipulations conflated climate change with threat. On the one hand, these two things are separable in principle, yet difficult to separate in practice, as it is almost impossible to discuss climate science without also highlighting its threatening consequences. Another limitation is that we primarily assessed pro-environmental intentions rather than actual behavior; even in Study 2, where we measured pro-environmental behavior, we relied on participants’ self-reports of behavior rather than objective measures.
Our samples consisted mainly of university students, limiting generalizability to the general population. All studies were conducted in laboratory or online settings using hypothetical scenarios. This paper includes no field studies, which further limits its applicability to real-world settings. A further limitation is that the studies presented here are cross-sectional and thus cannot provide evidence on the longer-term effects of repeated exposure to climate change information and the subsequent negative emotional episodes associated with it. Longitudinal studies are needed to investigate these processes. Longitudinal studies would allow researchers to track changes in climate anxiety and environmental engagement over time, providing insights into developmental trajectories and potential feedback loops between pro-environmentalism and emotion.
Furthermore, other variables not included in the current analysis may moderate the effects. Previous studies have indicated that efficacy beliefs are important in motivating people with high problem awareness to take action [58,59]. As indicated by the extended parallel process model, strong fear appeals in combination with low efficacy can lead to fear-control responses, such as denial, rather than to danger-control behavior [60]. This is likely because climate change represents a collective action problem characterized by delayed benefits, the diffusion of responsibility, and intergenerational costs, which weaken personal agency to mitigate climate risks [61,62,63,64]. These fear-control responses have been reported in the context of climate threats [24,25,65].
Furthermore, preexisting perspectives related to climate change, such as environmental self-identity or biospheric values, may determine whether making climate change salient increases pro-environmental behavior [66]. This underlines the ambivalent role of nature in the context of mortality salience. Individuals who strongly identify with nature may be more inclined to protect it when confronted with thoughts of their own death. In contrast, others may instead respond by focusing more on anthropocentric concerns [67].

5. Conclusions

The present research aimed to test the hypothesis that negative affective responses to climate change causally promote pro-environmental intentions and behavior. Across two complementary experimental approaches—a multi-study series using text-based manipulations and a laboratory experiment employing video-based stimuli—we consistently found that reminders of climate change reliably increased negative affect but did not translate into higher pro-environmental intentions or behavior. Moreover, even when negative affect was experimentally intensified through reflection, it did not enhance pro-environmental responses.
These findings challenge the assumption that eliciting negative affect is an effective short-term strategy for fostering sustainability and climate action. While affective reactions to climate change are easily triggered, they appear insufficient to change behavioral intentions in the short term. This does not imply that affect plays no role in climate engagement or that it cannot contribute to pro-environmental intentions and behavior in the long term. Still, our findings help understand the role of affect in communication strategies aimed at fostering sustainability.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/su18041826/s1. Table S1: An overview of the texts used in each study of Study series 1; Table S2: All pro-environmental intention items used in each study of Study series 1; Table S3: Factor loadings for the items of the long version of the pro-environmental intentions scale used in Studies 1d–e; Table S4: Factor loadings for the items of the short version of the pro-environmental intentions scale used in Studies 1a–b and 1f–h; Table S5: Multi-group CFA fit indices for the configural, metric and scalar invariance models; Table S6: Study-specific fixed-effects mediation models of Study series 1.

Author Contributions

Conceptualization, J.K., I.U.-H., T.F. and E.J.; methodology, I.U.-H., T.F. and J.K.; formal analysis, J.K.; investigation, I.U.-H. and T.F.; data curation, J.K.; writing—original draft preparation, J.K.; writing—review and editing, J.K., T.F. and I.U.-H.; visualization, J.K.; supervision, E.J. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University of Salzburg (protocol code EK-GZ:21/2016; date of approval: 8 August 2016).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data and the analysis code are openly available at the Open Science Framework at https://osf.io/a8pzw (accessed on 1 July 2025).

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Clayton, S. Climate Anxiety: Psychological Responses to Climate Change. J. Anxiety Disord. 2020, 74, 102263. [Google Scholar] [CrossRef]
  2. Whitmarsh, L.; Player, L.; Jiongco, A.; James, M.; Williams, M.; Marks, E.; Kennedy-Williams, P. Climate Anxiety: What Predicts It and How Is It Related to Climate Action? J. Environ. Psychol. 2022, 83, 101866. [Google Scholar] [CrossRef]
  3. Coffey, Y.; Bhullar, N.; Durkin, J.; Islam, M.S.; Usher, K. Understanding Eco-Anxiety: A Systematic Scoping Review of Current Literature and Identified Knowledge Gaps. J. Clim. Change Health 2021, 3, 100047. [Google Scholar] [CrossRef]
  4. Usher, K.; Durkin, J.; Bhullar, N. Eco-anxiety: How Thinking about Climate Change-related Environmental Decline Is Affecting Our Mental Health. Int. J. Ment. Health Nurs. 2019, 28, 1233–1234. [Google Scholar] [CrossRef] [PubMed]
  5. Clayton, S.; Karazsia, B.T. Development and Validation of a Measure of Climate Change Anxiety. J. Environ. Psychol. 2020, 69, 101434. [Google Scholar] [CrossRef]
  6. Verplanken, B.; Marks, E.; Dobromir, A.I. On the Nature of Eco-Anxiety: How Constructive or Unconstructive Is Habitual Worry about Global Warming? J. Environ. Psychol. 2020, 72, 101528. [Google Scholar] [CrossRef]
  7. Hickman, C.; Marks, E.; Pihkala, P.; Clayton, S.; Lewandowski, R.E.; Mayall, E.E.; Wray, B.; Mellor, C.; Van Susteren, L. Climate Anxiety in Children and Young People and Their Beliefs about Government Responses to Climate Change: A Global Survey. Lancet Planet. Health 2021, 5, e863–e873. [Google Scholar] [CrossRef]
  8. Niedzwiedz, C.; Katikireddi, S.V. Determinants of Eco-Anxiety: Cross-National Study of 52,219 Participants from 25 European Countries. Eur. J. Public Health 2023, 33, ckad160.069. [Google Scholar] [CrossRef]
  9. Gebhardt, N.; van Bronswijk, K.; Bunz, M.; Müller, T.; Niessen, P.; Nikendei, C. Scoping review of climate change and mental health in Germany—Direct and indirect impacts, vulnerable groups, resilience factors. J. Health Monit. 2023, 8, 122–149. [Google Scholar] [CrossRef]
  10. Collery, A.; Niedzwiedz, C.L. Climate Change Worry and the Association with Future Depression and Anxiety: Cross-National Analysis of 11 European Countries. BMJ Ment. Health 2025, 28, e301318. [Google Scholar] [CrossRef]
  11. Pitron, V.; Lemogne, C.; Clayton, S.; Léger, D.; Van Den Bergh, O.; Witthöft, M. Climate Change Anxiety and Its Association with Somatic Symptom Distress and Idiopathic Environmental Intolerances: A Cross-Sectional Study. J. Psychosom. Res. 2024, 187, 111937. [Google Scholar] [CrossRef]
  12. Ogunbode, C.A.; Pallesen, S.; Böhm, G.; Doran, R.; Bhullar, N.; Aquino, S.; Marot, T.; Schermer, J.A.; Wlodarczyk, A.; Lu, S.; et al. Negative Emotions about Climate Change Are Related to Insomnia Symptoms and Mental Health: Cross-Sectional Evidence from 25 Countries. Curr. Psychol. 2023, 42, 845–854. [Google Scholar] [CrossRef]
  13. Wullenkord, M.C.; Tröger, J.; Hamann, K.R.S.; Loy, L.S.; Reese, G. Anxiety and Climate Change: A Validation of the Climate Anxiety Scale in a German-Speaking Quota Sample and an Investigation of Psychological Correlates. Clim. Change 2021, 168, 20. [Google Scholar] [CrossRef]
  14. Paek, H.-J.; Hove, T. Mechanisms of Climate Change Media Effects: Roles of Risk Perception, Negative Emotion, and Efficacy Beliefs. Health Commun. 2024, 39, 3426–3435. [Google Scholar] [CrossRef] [PubMed]
  15. Huang, H. Media Use, Environmental Beliefs, Self-Efficacy, and pro-Environmental Behavior. J. Bus. Res. 2016, 69, 2206–2212. [Google Scholar] [CrossRef]
  16. Vrselja, I.; Pandžić, M.; Rihtarić, M.L.; Ojala, M. Media Exposure to Climate Change Information and Pro-Environmental Behavior: The Role of Climate Change Risk Judgment. BMC Psychol. 2024, 12, 262. [Google Scholar] [CrossRef]
  17. Thaker, J. Cross-Country Analysis of the Association between Media Coverage and Exposure to Climate News with Awareness, Risk Perceptions, and Protest Participation Intention in 110 Countries. Environ. Commun. 2024, 18, 233–249. [Google Scholar] [CrossRef]
  18. Van Der Werff, E.; Steg, L.; Keizer, K. The Value of Environmental Self-Identity: The Relationship between Biospheric Values, Environmental Self-Identity and Environmental Preferences, Intentions and Behaviour. J. Environ. Psychol. 2013, 34, 55–63. [Google Scholar] [CrossRef]
  19. Van Zomeren, M.; Spears, R.; Leach, C.W. Experimental Evidence for a Dual Pathway Model Analysis of Coping with the Climate Crisis. J. Environ. Psychol. 2010, 30, 339–346. [Google Scholar] [CrossRef]
  20. Skurka, C.; Romero-Canyas, R.; Joo, H.H.; Acup, D.; Niederdeppe, J. Emotional Appeals, Climate Change, and Young Adults: A Direct Replication of Skurka et al. (2018). Hum. Commun. Res. 2021, 48, 147–156. [Google Scholar] [CrossRef]
  21. Skurka, C.; Niederdeppe, J.; Romero-Canyas, R.; Acup, D. Pathways of Influence in Emotional Appeals: Benefits and Tradeoffs of Using Fear or Humor to Promote Climate Change-Related Intentions and Risk Perceptions. J. Commun. 2018, 68, 169–193. [Google Scholar] [CrossRef]
  22. Ettinger, J.; Walton, P.; Painter, J.; DiBlasi, T. Climate of Hope or Doom and Gloom? Testing the Climate Change Hope vs. Fear Communications Debate through Online Videos. Clim. Change 2021, 164, 19. [Google Scholar] [CrossRef]
  23. Baron, R.M.; Kenny, D.A. The Moderator–Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations. J. Personal. Soc. Psychol. 1986, 51, 1173–1182. [Google Scholar] [CrossRef]
  24. Uhl, I.; Klackl, J.; Hansen, N.; Jonas, E. Undesirable Effects of Threatening Climate Change Information: A Cross-Cultural Study. Group Process. Intergroup Relat. 2018, 21, 513–529. [Google Scholar] [CrossRef]
  25. Uhl, I.; Jonas, E.; Klackl, J. When Climate Change Information Causes Undesirable Side Effects: The Influence of Environmental Self-Identity and Biospheric Values on Threat Responses/Cuando La Información Sobre El Cambio Climático Tiene Efectos Indeseados: La Influencia de La Identidad Ambiental y de Los Valores Biosféricos En La Respuesta Ante Una Amenaza. PsyEcol. Biling. J. Environ. Psychol. 2016, 7, 307–334. [Google Scholar] [CrossRef]
  26. Watson, D.; Clark, L.A.; Tellegen, A. Development and Validation of Brief Measures of Positive and Negative Affect: The PANAS Scales. J. Personal. Soc. Psychol. 1988, 54, 1063–1070. [Google Scholar] [CrossRef]
  27. Schahn, J.; Damian, M.; Schurig, U.; Füchsle, C. Konstruktion und Evaluation der dritten Version des Skalensystems zur Erfassung des Umweltbewußtseins (SEU-3). Diagnostica 2000, 46, 84–92. [Google Scholar] [CrossRef]
  28. Vallacher, R.R.; Wegner, D.M. Levels of Personal Agency: Individual Variation in Action Identification. J. Personal. Soc. Psychol. 1989, 57, 660–671. [Google Scholar] [CrossRef]
  29. Schwarzer, R.; Jerusalem, M. General Self-Efficacy Scale. In Measures in Health Psychology: A User’s Portfolio. Causal and Control Beliefs; Windsor: Albany, NY, USA, 2012. [Google Scholar]
  30. Fritsche, I.; Cohrs, J.C.; Kessler, T.; Bauer, J. Global Warming Is Breeding Social Conflict: The Subtle Impact of Climate Change Threat on Authoritarian Tendencies. J. Environ. Psychol. 2012, 32, 1–10. [Google Scholar] [CrossRef]
  31. Steg, L.; Perlaviciute, G.; Van Der Werff, E.; Lurvink, J. The Significance of Hedonic Values for Environmentally Relevant Attitudes, Preferences, and Actions. Environ. Behav. 2014, 46, 163–192. [Google Scholar] [CrossRef]
  32. R Core Team. R: A Language and Environment for Statistical Computing; R Core Team: Vienna, Austria, 2025. [Google Scholar]
  33. Fox, J.; Weisberg, S. An R Companion to Applied Regression, 3rd ed.; Sage: Thousand Oaks, CA, USA, 2019. [Google Scholar]
  34. Makowski, D.; Wiernik, B.M.; Patil, I.; Lüdecke, D.; Ben-Shachar, M.S. Correlation: Methods for Correlation Analysis [R Package] 2022. Available online: https://CRAN.R-project.org/package=correlation (accessed on 1 July 2025).
  35. Rosseel, Y. Lavaan: An R Package for Structural Equation Modeling. J. Stat. Softw. 2012, 48, 1–36. [Google Scholar] [CrossRef]
  36. Barr, D.J.; Levy, R.; Scheepers, C.; Tily, H.J. Random Effects Structure for Confirmatory Hypothesis Testing: Keep It Maximal. J. Mem. Lang. 2013, 68, 255–278. [Google Scholar] [CrossRef]
  37. Breyer, B.; Bluemke, M. Deutsche Version der Positive and Negative Affect Schedule PANAS (GESIS Panel); Zusammenstellung sozialwissenschaftlicher Items und Skalen (ZIS); GESIS—Leibniz-Institut für Sozialwissenschaften: Mannheim, Germany, 2016. [Google Scholar] [CrossRef]
  38. Roth, G.; Benita, M.; Amrani, C.; Shachar, B.-H.; Asoulin, H.; Moed, A.; Bibi, U.; Kanat-Maymon, Y. Integration of Negative Emotional Experience versus Suppression: Addressing the Question of Adaptive Functioning. Emotion 2014, 14, 908–919. [Google Scholar] [CrossRef] [PubMed]
  39. Ayduk, Ö.; Kross, E. Analyzing Negative Experiences Without Ruminating: The Role of Self-Distancing in Enabling Adaptive Self-Reflection. Soc. Personal. Psychol. 2010, 4, 841–854. [Google Scholar] [CrossRef]
  40. Yadav, A.; Phillips, M.M.; Lundeberg, M.A.; Koehler, M.J.; Hilden, K.; Dirkin, K.H. If a Picture Is Worth a Thousand Words Is Video Worth a Million? Differences in Affective and Cognitive Processing of Video and Text Cases. J. Comput. High. Educ. 2011, 23, 15–37. [Google Scholar] [CrossRef]
  41. Faul, F.; Erdfelder, E.; Buchner, A.; Lang, A.-G. Statistical Power Analyses Using G*Power 3.1: Tests for Correlation and Regression Analyses. Behav. Res. Methods 2009, 41, 1149–1160. [Google Scholar] [CrossRef]
  42. Schwartz, S.H. Universals in the Content and Structure of Values: Theoretical Advances and Empirical Tests in 20 Countries. In Advances in Experimental Social Psychology; Elsevier: Amsterdam, The Netherlands, 1992; Volume 25, pp. 1–65. [Google Scholar]
  43. Bizumic, B.; Duckitt, J.; Popadic, D.; Dru, V.; Krauss, S. A Cross-cultural Investigation into a Reconceptualization of Ethnocentrism. Eur. J. Soc. Psychol. 2009, 39, 871–899. [Google Scholar] [CrossRef]
  44. Chapman, D.A.; Peters, E. Examining the (Non-Linear) Relationships between Climate Change Anxiety, Information Seeking, and pro-Environmental Behavioral Intentions. J. Environ. Psychol. 2024, 99, 102440. [Google Scholar] [CrossRef]
  45. Coates, Z.; Kelly, M.; Brown, S. The Relationship between Climate Anxiety and Pro-Environment Behaviours. Sustainability 2024, 16, 5211. [Google Scholar] [CrossRef]
  46. Linder, N.; Giusti, M.; Samuelsson, K.; Barthel, S. Pro-Environmental Habits: An Underexplored Research Agenda in Sustainability Science. Ambio 2022, 51, 546–556. [Google Scholar] [CrossRef]
  47. Mahardika, H.; Thomas, D.; Ewing, M.T.; Japutra, A. Comparing the Temporal Stability of Behavioural Expectation and Behavioural Intention in the Prediction of Consumers Pro-Environmental Behaviour. J. Retail. Consum. Serv. 2020, 54, 101943. [Google Scholar] [CrossRef]
  48. Pavani, J.-B.; Nicolas, L.; Bonetto, E. Eco-Anxiety Motivates pro-Environmental Behaviors: A Two-Wave Longitudinal Study. Motiv. Emot. 2023, 47, 1062–1074. [Google Scholar] [CrossRef]
  49. Veijonaho, S.; Hietajärvi, L.; Ojala, M.; Salmela-Aro, K. From Distress to Action?—A Three-Wave Longitudinal Study of Climate Change Distress, pro-Environmental Behavior, and Coping Strategies among Finnish Adolescents. J. Environ. Psychol. 2025, 105, 102676. [Google Scholar] [CrossRef]
  50. Brosch, T. Affect and Emotions as Drivers of Climate Change Perception and Action: A Review. Curr. Opin. Behav. Sci. 2021, 42, 15–21. [Google Scholar] [CrossRef]
  51. Pihkala, P. Toward a Taxonomy of Climate Emotions. Front. Clim. 2022, 3, 738154. [Google Scholar] [CrossRef]
  52. Marczak, M.; Wierzba, M.; Zaremba, D.; Kulesza, M.; Szczypiński, J.; Kossowski, B.; Budziszewska, M.; Michałowski, J.M.; Klöckner, C.A.; Marchewka, A. Beyond Climate Anxiety: Development and Validation of the Inventory of Climate Emotions (ICE): A Measure of Multiple Emotions Experienced in Relation to Climate Change. Glob. Environ. Change 2023, 83, 102764. [Google Scholar] [CrossRef]
  53. Wolfe, S.E.; Tubi, A. Terror Management Theory and Mortality Awareness: A Missing Link in Climate Response Studies? WIREs Clim. Change 2019, 10, e566. [Google Scholar] [CrossRef]
  54. Smith, L.K.M.; Ross, H.C.; Shouldice, S.A.; Wolfe, S.E. Mortality Management and Climate Action: A Review and Reference for Using Terror Management Theory Methods in Interdisciplinary Environmental Research. WIREs Clim. Change 2022, 13, e776. [Google Scholar] [CrossRef]
  55. O’Neill, S.; Nicholson-Cole, S. “Fear Won’t Do It”: Promoting Positive Engagement with Climate Change Through Visual and Iconic Representations. Sci. Commun. 2009, 30, 355–379. [Google Scholar] [CrossRef]
  56. Moser, S.C.; Dilling, L. Making Climate HOT. Environ. Sci. Policy Sustain. Dev. 2004, 46, 32–46. [Google Scholar] [CrossRef]
  57. Hastings, G.; Stead, M.; Webb, J. Fear Appeals in Social Marketing: Strategic and Ethical Reasons for Concern. Psychol. Mark. 2004, 21, 961–986. [Google Scholar] [CrossRef]
  58. Bockarjova, M.; Steg, L. Can Protection Motivation Theory Predict Pro-Environmental Behavior? Explaining the Adoption of Electric Vehicles in The Netherlands. Glob. Environ. Change 2014, 28, 276–288. [Google Scholar] [CrossRef]
  59. Van Der Werff, E.; Steg, L. One Model to Predict Them All: Predicting Energy Behaviours with the Norm Activation Model. Energy Res. Soc. Sci. 2015, 6, 8–14. [Google Scholar] [CrossRef]
  60. Witte, K. Putting the Fear Back into Fear Appeals: The Extended Parallel Process Model. Commun. Monogr. 1992, 59, 329–349. [Google Scholar] [CrossRef]
  61. Hurlstone, M.J.; Price, A.; Wang, S.; Leviston, Z.; Walker, I. Activating the Legacy Motive Mitigates Intergenerational Discounting in the Climate Game. Glob. Environ. Change 2020, 60, 102008. [Google Scholar] [CrossRef]
  62. Reser, J.P.; Bradley, G.L. Fear Appeals in Climate Change Communication. In Oxford Research Encyclopedia of Climate Science; Oxford University Press: Oxford, UK, 2017. [Google Scholar]
  63. Tannenbaum, M.B.; Hepler, J.; Zimmerman, R.S.; Saul, L.; Jacobs, S.; Wilson, K.; Albarracín, D. Appealing to Fear: A Meta-Analysis of Fear Appeal Effectiveness and Theories. Psychol. Bull. 2015, 141, 1178–1204. [Google Scholar] [CrossRef]
  64. Gardiner, S.M. A Perfect Moral Storm: Climate Change, Intergenerational Ethics and the Problem of Moral Corruption. Environ. Values 2006, 15, 397–413. [Google Scholar] [CrossRef]
  65. Fritsche, I.; Jonas, E.; Kessler, T. Collective Reactions to Threat: Implications for Intergroup Conflict and for Solving Societal Crises. Soc. Issues Policy Rev. 2011, 5, 101–136. [Google Scholar] [CrossRef]
  66. Bolderdijk, J.W.; Gorsira, M.; Keizer, K.; Steg, L. Values Determine the (In)Effectiveness of Informational Interventions in Promoting Pro-Environmental Behavior. PLoS ONE 2013, 8, e83911. [Google Scholar] [CrossRef]
  67. Fritsche, I.; Hoppe, A. We Supernaturals. In Handbook of Terror Management Theory; Elsevier: Amsterdam, The Netherlands, 2019; pp. 157–178. [Google Scholar]
Sustainability 18 01826 g001
Figure 1. A visual summary of the multilevel mediation analysis results in Study series 1. Climate change salience increased negative affect and reduced positive affect. However, climate change salience did not influence pro-environmental intentions, neither directly nor indirectly, via affect. The numbers reflect unstandardized regression coefficients. *** = p < 0.001; * p < 0.05.
Figure 1. A visual summary of the multilevel mediation analysis results in Study series 1. Climate change salience increased negative affect and reduced positive affect. However, climate change salience did not influence pro-environmental intentions, neither directly nor indirectly, via affect. The numbers reflect unstandardized regression coefficients. *** = p < 0.001; * p < 0.05.
Sustainability 18 01826 g001
Sustainability 18 01826 g002
Figure 2. A visual illustration of the results of Study 2. Both climate change salience and reflecting on (vs. suppressing) emotions increased negative affect and lowered positive affect. There were no significant interactions between the two manipulations.
Figure 2. A visual illustration of the results of Study 2. Both climate change salience and reflecting on (vs. suppressing) emotions increased negative affect and lowered positive affect. There were no significant interactions between the two manipulations.
Sustainability 18 01826 g002
Table 1. Basic characteristics of the studies in Study series 1 (Studies 1a–1h), in which participants either received information about climate change (climate change—salient; CCS) or no information about climate change (climate change—not salient; CNS). The table presents sample sizes, gender distributions, age, experimental and control materials, internal consistencies of the measures, and manipulation checks. Manipulation check averages and difference tests for both groups suggest that in each study, participants in the climate change salience (CCS) group felt more threatened than those in the climate change—not salient (CCNS) group.
Table 1. Basic characteristics of the studies in Study series 1 (Studies 1a–1h), in which participants either received information about climate change (climate change—salient; CCS) or no information about climate change (climate change—not salient; CNS). The table presents sample sizes, gender distributions, age, experimental and control materials, internal consistencies of the measures, and manipulation checks. Manipulation check averages and difference tests for both groups suggest that in each study, participants in the climate change salience (CCS) group felt more threatened than those in the climate change—not salient (CCNS) group.
Sample SizeAgeMaterials aInternal Consistency (α)Manipulation Check (0–1)
CC SalientCC Not SalientCCS > CCNS
StudyNNfemaleMSDCC SalientCC Not Salient Positive AffectNegative AffectPEIMSDMSD
1a624130.8911.64LocalLocal geo0.880.90.750.700.290.230.3***
1b14010729.611.78LocalLocal geo0.840.880.80.770.230.090.17***
1c995531.7212GlobalGlobal geo0.860.920.90.770.230.290.29***
1d703624.846.46Local + GlobalFashion colors0.830.850.850.290.280.160.23***
1e13010321.794.63GlobalGlobal geo0.760.760.890.760.170.130.19***
1f1196624.33.77LocalLocal geo0.80.850.790.860.150.120.19**
1g954523.737.72LocalLocal geo0.860.870.770.680.240.120.19***
1h1047624.110.35LocalLocal geo0.820.90.920.700.230.230.21***
Total81952926.169.62
Note. *** p < 0.001; ** p < 0.01. a Materials: In the climate change—salient group, participants were confronted with local and/or global consequences of climate change, whereas in the climate change—not salient group, participants were confronted with local (local geo) or global (global geo) geographical facts or the latest fashion colors.
Table 2. Means, standard deviations, and correlations among the measures used in Study series 1. Means and standard deviations are reported separately for the group that received information about climate change (climate change—salient; CCS) and those that did not (climate change—not salient; CCNS).
Table 2. Means, standard deviations, and correlations among the measures used in Study series 1. Means and standard deviations are reported separately for the group that received information about climate change (climate change—salient; CCS) and those that did not (climate change—not salient; CCNS).
Climate Change—Not Salient (CCNS; N = 431)Climate Change—Salient (CCS; N = 388)CCS > CCNSr(818)
VariableM (SD)M (SD)t(813.10)12
1. Negative affect0.13 (0.14)0.28 (0.20)12.80 ***
2. Positive affect0.43 (0.18)0.41 (0.16)−1.930.132 ***
3. Pro-environmental intention0.62 (0.17)0.63 (0.17)−0.07−0.0020.087 *
Note. M and SD represent mean and standard deviation, respectively. * indicates p < 0.05. *** indicates p < 0.001.
Table 3. Results of the multilevel mediation meta-analysis across studies. At the within-study level, climate change salience reliably increased negative affect and reduced positive affect. However, climate change salience did not lead to significant changes in pro-environmental intentions, neither indirectly (via affect) nor directly.
Table 3. Results of the multilevel mediation meta-analysis across studies. At the within-study level, climate change salience reliably increased negative affect and reduced positive affect. However, climate change salience did not lead to significant changes in pro-environmental intentions, neither indirectly (via affect) nor directly.
EffectEstimateSE95% CIp
LLUL
Fixed effects (within-study)
  CCS → NA0.1410.0330.0750.206<0.001
  CCS → PA−0.0240.010−0.043−0.0060.011
  NA → PEI−0.0280.036−0.0980.0420.430
  PA → PEI0.0920.060−0.0250.2100.124
  CCS → PEI (direct effect)0.0070.013−0.0190.0330.590
  CCS → NA → PEI−0.0040.005−0.0140.0060.450
  CCS → PA → PEI−0.0020.001−0.0050.0010.118
  Total effects0.0010.010−0.0190.0210.930
Random effects
  Within-study variance0.0290.0040.0220.036<0.001
  Between-study variance0.0030.002−0.0010.0080.133
ICC0.102----
Note. Number of studies = 8; total N = 819; CCS = climate change salience; NA = negative affect; PA = positive affect; PEI = pro-environmental intention; ICC = intraclass coefficient; CI = confidence interval; LL = lower limit; UL = upper limit.
Table 4. Affect, pro-environmental intentions, and pro-environmental behavior (PEB) in Study 2; left: means and standard deviations; middle: effects of the reflection and climate change salience manipulations and their interaction; right: bivariate correlations.
Table 4. Affect, pro-environmental intentions, and pro-environmental behavior (PEB) in Study 2; left: means and standard deviations; middle: effects of the reflection and climate change salience manipulations and their interaction; right: bivariate correlations.
b (SE)r(135)
VariableMSDReflectionCCSReflection x CCS123
1. Negative affect0.380.290.1 (0.05) *0.42 (0.04) ***−0.03 (0.07)
2. Positive affect0.380.19−0.09 (0.04) *−0.25 (0.03) ***0.03 (0.05)−0.61 **
3. Pro-environmental intentions0.760.13−0.02 (0.03)0.05 (0.03)−0.02 (0.05)0.21 *−0.10
4. Pro-environmental behavior0.700.17−0.03 (0.06)−0.01 (0.05)0.03 (0.08)0.03−0.030.75 **
Note. M and SD are used to represent mean and standard deviation, respectively. * indicates p < 0.05. ** indicates p < 0.01. *** indicates p < 0.001.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Klackl, J.; Uhl-Hädicke, I.; Fowles, T.; Jonas, E. Experimental Tests of an Affective Pathway to Pro-Environmental Intentions. Sustainability 2026, 18, 1826. https://doi.org/10.3390/su18041826

AMA Style

Klackl J, Uhl-Hädicke I, Fowles T, Jonas E. Experimental Tests of an Affective Pathway to Pro-Environmental Intentions. Sustainability. 2026; 18(4):1826. https://doi.org/10.3390/su18041826

Chicago/Turabian Style

Klackl, Johannes, Isabella Uhl-Hädicke, Tilmann Fowles, and Eva Jonas. 2026. "Experimental Tests of an Affective Pathway to Pro-Environmental Intentions" Sustainability 18, no. 4: 1826. https://doi.org/10.3390/su18041826

APA Style

Klackl, J., Uhl-Hädicke, I., Fowles, T., & Jonas, E. (2026). Experimental Tests of an Affective Pathway to Pro-Environmental Intentions. Sustainability, 18(4), 1826. https://doi.org/10.3390/su18041826

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop