Emotional Processes in
Climate Communication:
Distinct emotions, in-group actions,
and pro-environmental behaviour
Ragnheiður Torfadóttir Student number: 11363585 Research Master’s Thesis
Supervisor: Dr. Anke Wonneberger 28. 06. 2019
Abstract
If we are to avoid catastrophic climate change, individual action is needed. This paper seeks to augment the scientific understanding on how people can be motivated to mitigation action, both in public and private spheres. Emotions are a crucial factor in this, because of their motivating qualities. Distinct emotions have been found to have differing action
potential, and also to lead to differing behavioural intentions. Incorporating elements from the cognitive theory of stress, this paper analysed the effects of in-group portrayal on
pro-environmental behavioural intentions, mediated by anger, guilt, and pride.
In two between-subjects experiments, the respondents’ nation – the in-group – was portrayed in a news article as having either a large or small carbon footprint compared to other countries. This evoked emotional reactions which mediated the effects of the messages on their pro-environmental behavioural intentions. Anger was consistently found to be the most motivating emotion, influencing both public and private-sphere behaviours, as well as willingness to pay. More research is needed to validate these results, using better scales to measure private-sphere behaviours, and measuring emotional reactions more precisely.
This paper adds to the literature of appraisal theories by analysing concrete behavioural intentions – for the first time to the researchers’ best knowledge – as outcomes of anger, guilt, and pride.
Emotional Processes in Climate Communication:
Distinct emotions, in-group actions, and pro-environmental behaviour Climate change is the biggest challenge that humanity faces in the 21st century.
Although awareness has risen sharply in the last few years, the global community is still very far from reaching mitigation targets needed to limit global warming to “well below 2°C,” the target set by the Paris Climate Accord in 2015 (United Nations, 2015, p. 3). Global warming has already reached 1.1°C compared to pre-industrial levels and is on course to hit 1.5°C, even if emission of greenhouse gasses into the atmosphere would cease immediately (Mauritsen & Pincus, 2017). If we are to limit carbon emissions enough to keep global warming under 2°C this century, it’s clear that an effort is needed from individuals in all levels of society, in every country in the world.
Understanding what drives individuals to act on climate change is of the utmost importance. The cumulative impact of individual behaviours can be considerable (Clarke, Heinonen, & Ottelin, 2017), and political action is indispensable for holding governments and conglomerates accountable for their emissions and inaction in preventing them (Wallace-Wells, 2019). Analysing the effects of news media messages is appropriate because they are the most common channel for people to learn about the problem (Leiserowitz, Smith, & Marlon, 2010).
In the last decade, literature on environmental communication has increasingly turned attention to distinct emotional reactions to messages about climate change (e.g. Feldman & Hart, 2016; Ferguson & Branscombe, 2010; O’Neill & Nicholson-Cole, 2009; Smith & Leiserowitz, 2014) and environmental issues in general (e.g. Bissing-Olson, Fielding, & Iyer, 2016; Meijers, Remmelswaal, & Wonneberger, 2018; Wonneberger, 2018). Emotions are not simply irrational or ancillary reactions, but responses that have developed over millions of years of natural selection. Distinct emotions have been found to correspond to different
coping strategies, and can be viewed as a system of coping (Roseman, 2013; Roseman, Wiest, & Swartz, 1994). Therefore, distinct emotional reactions can be expected to result in different behaviour.
Harth, Leach & Kessler (2013) found different emotional reactions to messages about in-group responsibility for climate change to have differing attitudinal outcomes. Following their findings, this study analyses emotional reactions to messages about the participants’ in-group – news about the carbon footprint of their nation – and how the distinct emotions influence behavioural intentions. This study expands on the research by Harth et al. (2013) by not only analysing attitudinal outcomes or intent-oriented behaviours, but looking at how emotional mechanisms influence impactful pro-environmental behavioural intentions (PEB).
Elements from appraisal theories – the cognitive theory of stress in particular (Homburg & Stolberg, 2006; Lazarus, 1991) – are also incorporated to better understand the antecedents of discrete emotions. Primary appraisal was measured as the perceived threat of climate change to the individual and two dimensions of secondary appraisal were measured, self-efficacy and collective self-efficacy. This paper adds to the literature of appraisal theories by analysing concrete behavioural intentions – for the first time to the researchers’ best knowledge – as outcomes of anger, guilt, and pride.
Membership in groups is a very important determinant of self-image and identity (Morrison, Decety, & Molenberghs, 2012), which is why media messages about the
performance of the in-group are expected to influence the group’s members. The evaluation of groups is primarily relational, defined in terms of comparisons to other social groups (Tajfel, 1982). A message comparing the behaviour of the in-group to others is expected to induce emotional responses, positive or negative, depending on the favourability of the comparison. The in-group is operationalised as the nation and the stimulus showed the
respondents’ home countries’ carbon footprint as either larger or smaller than that of other countries.
A negative message about the carbon footprint of the participant’s home country is expected to evoke guilt or anger, but a positive message to evoke pride. These emotions are expected to affect PEB in different ways. The aim is to analyse the effects of communication about climate change on PEB, how this is mediated by distinct emotions toward the in-group and moderated by cognitive factors, with the aim of furthering the scientific understanding on how PEB can be encouraged. Two studies were conducted to answer the following research questions, one using an international sample and another using samples of Iceland and the Netherlands.
How do messages about country performance affect pro-environmental behavioural intentions? Is this effect mediated by anger, guilt, and/or pride? Can these emotional
reactions be further explained by perceived threat (primary appraisal) or efficacy (secondary appraisal)?
Theoretical Framework Social identity theory
Social identity theory (Tajfel & Turner, 1986) suggests that self-image has both an individual and a group component. The group component, social identity, is derived from the groups people perceive themselves to belong to. Tajfel and Turner (1986) define social groups as groups that individuals use to define themselves and others define them by. One such social group, that most people belong to, is their nation. Morrison et al. (2012) supported this with a neuroscientific study on group membership by showing that words associated with the in-group, including nationality, activated regions in the brain previously linked to the self and personal identity.
The concept of social identity is especially interesting, due to its potential to evoke emotional reactions. A discrepancy between an individual’s values and the behaviour of a group can lead to group-level feelings of guilt (Doosje, Branscombe, Spears, & Manstead, 1998). Beliefs about the performance of the nation are expected to affect people’s beliefs about their own behaviour and cause emotional reactions. The emotional mechanism will be further explored in the following chapter.
Appraisal theory and distinct emotions
Before discussing effects of emotions, it is important to pay attention to their onset and characteristics. Emotional reactions do not only differ between the different stimulants that induce emotions, but also differ greatly between the people who perceive the stimulus and their situations. Appraisal theorists believe that this is partly due to differing cognitive evaluations, or appraisals, of the message or situation (for overview see Roseman & Smith, 2001). Appraisal theories assume that “thought and emotion are largely inseparable”
(Ellsworth, 2013, p. 125), and how a situation is perceived – the appraisals of it – determine whether and which emotion is evoked. Anger, guilt, and pride have been chosen as the focus of this study because of their interesting and differing characteristics.
Anger arises when a slight or attack is perceived, towards the self, others, or personal
values (Lazarus, 1991) and is associated with appraisal of certainty that something bad has happened (Lemer & Keltner, 2001). When people are angry, they are likely to explain their emotions in active ways, and it is associated with a high action potential (Roseman et al., 1994). According to Valentino, Gregorowicz, and Groenendyk: “Anger is the body’s response to a frustrating situation in which we possess a sense of control and believe our future actions will lead to success in dealing with the problem at hand.” (2009, p. 311).
Guilt is a self-conscious emotion that occurs when an individual perceives themselves as
is considered to be an emotion of relatively low action potential (Roseman, 2013), but can generate a feeling of responsibility to act (Basil, Ridgway, & Basil, 2006). It has been found to influence intentions to repair damage that has been done (Harth et al., 2013; Roseman et al., 1994), and to increase charitable donation intentions (Basil et al., 2006; Wonneberger, 2018). Doosje et al. (1998) found that guilt can also arise as a result of violations by the in-group that the individuals are not directly responsible for.
Pride is another self-conscious emotion and it arises in association with behaviour that is
in accordance with personal norms or when goals are reached (Lewis, 2008). Although negative emotions have generally received more attention, positive emotions can also be strong influencers of behaviour (Ellsworth, 2013). Pride as a result of previous behaviours can act as a motivator for future positive behaviours, allowing the individual to sustain the
positive emotion (Bissing-Olson et al., 2016). Primary appraisal
Lazarus (1991) categorizes appraisals as primary or secondary. He describes primary appraisal as a person’s stake in the outcomes of an event or message. He names it primary “because without a stake there is no potential for an emotion” (Lazarus, 1991, p. 827). Two dimensions of primary appraisal are considered: valance and relevance.
The valance of the message, or whether the message is consistent with the receivers’ goals, determines if the emotion felt will be negative or positive (Lazarus, 1991). In this study, a message about the countries’ small carbon footprint is expected to be congruent with the respondents’ goals and result in positive emotional reactions, while a negative message about the in-groups’ large carbon footprint is expected to be incongruent with the respondents goals and result in negative emotions.
H1. Those who are exposed to a negative message about national emissions will express
more (a) anger and (b) guilt and those who are exposed to a positive message about national emissions will express more (c) pride.
The relevance of the message determines the potential for emotion, as something that we deem unimportant is not likely to evoke emotions. Thus, climate change is not likely to evoke strong emotional reactions unless it is perceived as relevant to the individual. When an
individual sees climate change as a threat to themselves or their family, the issue can be said to be relevant to them. Perceived threat is expected to influence the extent to which the stimulus results in emotional reactions.
H2. Those who express higher threat beliefs about climate change will experience
stronger emotional reactions to messages about national emissions.
Secondary appraisal
Secondary appraisal concerns the options available to cope with a situation (Lazarus, 1991). If an individual thinks that climate change is an important issue that needs to be
addressed – but simultaneously feels that they are unable to address it themselves or even as a member of a group – then this can be expected to lead to negative emotions with low action potential, such as sadness, apathy, or even fatalism (Frijda, 2016; Mayer & Smith, 2018). However, if they feel capable of impacting the situation, this will likely increase the chances of emotions that have more action potential arising. Two dimensions of secondary appraisal will be considered, self-efficacy and collective efficacy. Self-efficacy is the perceived
competence of the individual to engage in actions that are potentially impactful, and collective
efficacy refers to the belief that a group can collectively solve their problems (Bandura, 1982).
Anger is thought to arise in threatening situations, when some agency to cope with the situation is felt (Lemer & Keltner, 2001). Valentino et al. (2009) analysed emotional
arise in individuals who were confident about their ability to participate in the political process. Thus, if a person believes that climate change needs to be combatted, and her group is not doing what is needed to address it, she might become angry at the group, but only if she feels that her efforts could potentially have some impact (high self- and/or collective
efficacy).
H3. Secondary appraisal will moderate the effect of messages on anger, so that those
with higher (a) self-efficacy or (b) collective efficacy feel more anger in the negative condition than those with lower self-efficacy.
However, if a person feels that she could do something about it (high self-efficacy) but simultaneously knows that she is not doing enough (negative portrayal of in-group actions), this will likely evoke guilt. Ferguson and Branscombe (2010) found collective guilt to arise when a message about the human causes of climate change portrayed the effects of climate change in less severe terms, and thus repair seemed possible. They suggest that one possible explanation for this is that the severe risk message may have negatively influenced the participants efficacy beliefs, causing them to resort to fatalism or justifying their actions. Thus, a discrepancy between a person’s values and the behaviour of her group can result in feeling guilt, and this is more likely if she believes she can have an impact (high self- or collective efficacy).
H4. Secondary appraisal will moderate the effect of messages on guilt, so that those with
higher (a) self-efficacy or (b) collective efficacy feel more guilt in the negative condition than those with lower self-efficacy.
On the other hand, if she feels that her actions as a group member are impactful (high collective efficacy) and the group is doing what is needed (positive portrayal if in-group), this might lead to feeling pride.
H5. Collective efficacy will moderate the effect of messages on pride, so that those with
higher collective efficacy will feel more pride in the positive condition than those with lower collective efficacy.
Effects of discrete emotions on PEB
This study analyses discrete emotions – focusing on distinct emotions like anger and guilt rather than dimensions such as valance or arousal – because the discrete emotions paradigm contributes “elements that provide the nuance necessary to explain human action more fully” (Nabi, 2010, p. 154). Smith and Leiserowitz (2014) found that discrete emotions alone could explain 50% of the variance in public global warming policy support, while a holistic affect variable could only explain 28%. Distinct emotional reactions to political information have been found to influence message processing, learning, and behavioural outcomes in different ways (Marcus & Mackuen, 1993; Valentino, Hutchings, Banks, & Davis, 2008). The outcome of a message can therefore be expected to differ depending on which emotions it evokes in the receiver.
This is consistent with findings from Harth et al. (2013), who found three distinct emotions to influence three different kinds of climate change attitudes. They conducted an experiment where the responsibility of the in-group – Germany – for environmental outcomes was manipulated. In one condition the in-group’s environmental damage was highlighted and in the other the in-group’s environmental protection efforts were highlighted. Both guilt and anger were higher in the damage condition and pride higher in the protection condition. They found guilt to predict intentions to repair damage, anger to predict favouring the punishment
of wrongdoers, and pride to predict willingness to support environmental protection efforts that favour the in-group. If different emotions, even arising from the same stimulus, have
Although Harth et al. (2013) named their dependent variables behavioural intentions, they were measured with very abstract items that measured general attitude rather than behavioural intentions. This study follows up on their findings and analyses whether such effects can be found on more tangible behavioural intentions. By doing that, the study follows a recommendation by Stern (2000) and Swim et al. (2011) to analyse behaviours that have large potential effects on emissions. Stern (2000) distinguishes between environmentally significant behaviours in the public and private spheres. PEB in the public sphere is, for example, protesting and petitioning, and PEB in the private-sphere is individual actions such as driving less or buying locally grown produce. This study measures intentions to engage in the individual private-sphere actions that have the biggest impact on emissions (Wynes & Nicholas, 2017), as well as public sphere behaviours that have the potential to bring about societal or political change. Although Stern’s model includes the same predictors for all behaviours, he expects different behaviours to have different antecedents, as he also found in previous studies (Dietz, Stern, & Guagnano, 1998; Stern, Dietz, Abel, Guagnano, & Kalof, 1999).
This study also expands on Harth et al. (2013) by comparing more countries and by analysing the role of appraisals. This allows further exploration of the emotional mechanism through which media messages affect different behavioural intentions. In the following sections, the potential behavioural outcomes of anger, guilt, and pride will be outlined. Since no study has analysed the effects of emotional responses to messages on private and public sphere PEB, the following hypotheses are explorative in nature.
Anger. Anger experienced as a reaction to the message might be directed towards the authorities that can be blamed for acting selfishly or irresponsibly, or towards other members of the in-group, who are perceived to be responsible for inaction. Such anger is expected to lead to action readiness, especially towards political action, such as participating in
demonstrations, petitioning, or participating actively in environmental organizations. Negative in-group portrayal is expected to increase public sphere PEB if it is mediated by anger, as anger is a better motivator of political action than other emotions (Leach, Iyer, & Pedersen, 2006; Valentino, Brader, Groenendyk, Gregorowicz, & Hutchings, 2011).
H6. Negative in-group portrayal will evoke anger, which in turn predicts public sphere
PEB.
Guilt.Reading about their countries’ large carbon footprint, people may be reminded of their own transgressions and feel guilty, or feel collective guilt over their in-group’s
behaviour. Leach and his colleagues have found guilt to be a weak predictor of political action, relative to anger, and more associated with abstract goals than willingness to engage in specific action (Iyer & Leach, 2008; Leach et al., 2006). In a climate change context, they found guilt to increase willingness to engage in vaguely defined actions to repair the damage that has been done (Harth et al., 2013). Doosje et al. (1998) found collective guilt to lead to both positive attitudes toward non-specific compensation behaviours and intentions to engage in what they call specific behaviour, which was measured as willingness to pay to charity.
Other studies that have analysed the effects of guilt on environmental outcomes have found guilt to result in relatively low-cost behavioural intentions, such as choosing energy-efficient lightbulbs (Ferguson & Branscombe, 2010), in willingness to pay green taxes (Ferguson & Branscombe, 2010), and willingness to donate to charity (Basil et al., 2006; Wonneberger, 2018).
H7. Negative portrayal of in-group behaviour induces guilt in participants, which in turn
increases (a) private-sphere PEB intentions and (b) willingness to pay.
Pride. Pride can be expected to have a positive effect on behavioural outcomes, because people are motivated to sustain the emotion with continued positive actions. Bissing-Olson, Fielding, and Iyer (2016) used an experience sampling design to analyse the effects of pride
and guilt on everyday private-sphere PEB. They found that pride about previous behaviour was positively related to PEB in the next time period (2.5 hours). Onwezen, Antonides, and Bartels (2013; 2014) analysed the effects of anticipated pride and guilt in two papers, and found both to have positive effects on private-sphere PEB.
H8. Positive in-group portrayal evokes pride, which in turn is predictive of
private-sphere PEB intentions.
The messages in this study are expected to evoke the emotions anger, guilt, and/or pride, and this process is expected to be influenced by primary and secondary appraisals. The emotions are further expected to result in different kinds of PEB, as summarised in Figure 1.
Figure 1. Hypothesised model of emotional mechanisms in response to messages about the in-group. Note. All arrows present positive effects. Willingness to pay was only measured in the second study.
Method
Two three-group between-subjects experimental studies were conducted to test the hypotheses. In the first study (S1) an international sample was used and an English-language questionnaire. Appraisal theories generally assume that appraisal-emotion processes are universal and data largely supports this view (Roseman & Smith, 2001). This does not
exclude the possibility that different cultures may have different appraisals of events, but suggests that the same appraisals will result in the same emotions, regardless of culture.
To test the effects of the nation as in-group more precisely with more relevant messages, the second study (S2) drew on samples from two countries. The countries chosen were two Northern-European countries with large carbon footprints, Iceland and the Netherlands. The questionnaire was translated into the respective languages.
Sample
In both studies, convenience sampling of the researcher’s network was used, along with the website Survey Circle (NS1 = 212 ; NS2 = 272). Both samples were highly educated, with
median education levels of a bachelor’s degree. Participants in S1 were from 47 different countries, but the majority was European (n = 162). Participants in S2 were from Iceland (n = 188) and the Netherlands (n = 84). In S2, mean age was considerably higher in the Icelandic group (M = 42.53, SD = 15.64) than in the Dutch group (M = 28.59, SD = 8.14). This was also reflected in the income level, as the median income of the Icelandic group was two brackets above that of the Dutch sample. More detailed demographic data is displayed in Table 5.1
Procedure
Participants in S1 were asked to answer an online questionnaire in English, which started with an introduction and informed consent page, leading to a pre-questionnaire that measured demographics, perceived threat of climate change, self-efficacy and collective efficacy in the climate change context. They were then randomly allocated to one of the experimental groups, positive portrayal of in-group, negative portrayal of in-group, or control condition. After reading the stimulus they were presented with a post-questionnaire that measured emotional reactions to the stimulus, PEB intentions in the private and public sphere, climate
change concern, self-reported knowledge of the subject, and a manipulation check. Finally, they were presented with a debriefing statement, where people were informed about the incorrect information they received and were provided with a link detailing the actual carbon footprint of their countries.
In S2, the procedure was the same, although national identity was also measured in the post-test and the questionnaire was presented in Icelandic and in Dutch.
Stimulus
In S1 the in-group was operationalised as the respondents’ country of origin, by inserting information from the pre-questionnaire into the stimulus text. The text otherwise differed only depending on the experimental condition, and not depending on actual country-specific values. The messages portrayed the in-group in a negative, neutral, or positive light. All messages included a passage on the risks of climate change and explained how “a recent study” analysed the carbon footprint of multiple countries. Tajfel and Turner (1986) expect the evaluations of the in-group to be predominantly based on comparison to other social groups. Such social comparison was introduced by comparing the country to “neighbouring countries.” The control stimulus (n = 73) then stated that the participant’s country was close to the average carbon footprint. The negative (n = 69) stimulus presented the country as having one of the biggest carbon footprints, and the positive stimulus (n = 70) presented the country as having one of the smallest carbon footprints. To underline the responsibility of the nation, a paragraph followed explaining this by attributing blame or credit to government, industry, and individual actions. Full stimulus materials can be found in Appendix A. A pilot test was conducted to test the stimulus material, for more information see Appendix B.
In S2, the stimulus materials were similar to S1. The negative condition stated that the country was doing poorly, and that the country’s carbon footprint was 55% higher than the EU average in the Dutch text and the Nordic countries’ average in the Icelandic text. The
positive condition stated that the country was doing well compared to its neighbours, that the footprint was 5% lower than the EU/Nordic average. The control condition stated that the carbon footprint was similar to neighbouring countries.
Measures
Primary appraisal. Cognitive level of perceived threat was measured with a 5-item scale from Hartmann et al. (2014), with questions such as: “How likely is it that climate change will affect you in your lifetime?” All response scales had 7 points and question-specific labels.
Secondary appraisal. Scales from Van Zomeren, Spears, and Leach (2010) were used to measure self- and collective efficacy with regards to climate change. Self-efficacy was
measured with 5 items, such as: “There are simple things I can do that reduce the negative consequences of the climate crisis.” A 7-point agreement scale was offered. Collective efficacy was measured with three items, including: “To what extent do you think that people can jointly prevent the negative consequences of the climate crisis?” This was measured on a 7-point scale with labels from 0: “Not at all” to 6: “A great deal.”
Emotions. After being exposed to the stimulus, participants were asked on unipolar scales to what extent the text had evoked the following emotions. In S1, these emotions were measured: anger, sadness, disgust, fear, anxiety, guilt, hope, pride, enthusiasm, and happiness. Additional emotions were included so as not to prompt participants too clearly to what was being analysed.
In S2, emotions were measured more precisely by identifying both the subject and the object of the emotion felt. Items were therefore phrased: “I, as an Icelander/Dutchman, feel __ about our behaviour,” identifying the object as not only the individual but the individual as part of the national group and the subject as the nation. Group-based guilt, anger, and pride were measured with items from Harth et al. (2013). The words to measure anger were: angry,
annoyed, irritated, enraged and disgusted. To measure guilt, the words guilty, regretful, remorseful, sorry, complicit, and ashamed were used. Pride was measured with the words proud, successful, superior, happy, satisfied, and confident.
PEB intentions. This survey measured the individual changes that have the largest impact on lowering emissions: having one fewer child, living car-free, avoiding air-travel, and eating a plant-based diet (Wynes & Nicholas, 2017). Intentions to engage in these behaviours were measured with the very active phrasing “I will” to further try to capture actual
behavioural intentions, rather than attitudes towards the behaviour. These are relatively
difficult changes to make, so three less effortful behaviours, that also have a potential to lower an individual’s emissions were added: choosing local products, paying to offset emissions of flights, and recycling (only in S2). Public sphere PEB was measured using a 5 item scale from Hart & Feldman (2016), to which one item was added to measure voting intentions. In S2, two items were added to measure the participants willingness to pay (WTP), willingness to pay for eco-friendly products and support for a green tax.2
Manipulation check and credibility. One item was used to measure if the article had the intended effect: “I believe my country is doing poorly in climate change mitigation.” And another item was used to measure if people perceived the article as intended: “The news article showed my country in a positive light.” Three items were used to measure the credibility of the news article, measuring whether the respondents thought it was credible, newsworthy, and whether they believed it.3
2 For all PEB items the answering scale provided was 0 “I will not do this”, 1 “I am very unlikely to do this”, 2
“I am unlikely to do this”, 3 “I might do this”, 4 “I am likely to do this,” 5 “I am very likely to do this,” 6 “I am certain I will do this.”
Statistical Analysis
One-way ANOVAs were performed to test whether the manipulation worked, whether they evoked the expected emotional reactions, and if the credibility differed between conditions.
Measurement model
To test the measurement model, confirmatory factor analyses were conducted using full information maximum likelihood in IBM AMOS version 25. Multivariate kurtosis was high in both measurement models (c.r.S1 = 16.26; c.r.S2 = 17.84) and therefore bootstrapping was
used to evaluate model fit. The data had no ill-scaled variances and missing cases were addressed using regression imputation with AMOS.
Private-sphere PEB intentions were split into two factors for both groups. Smaller changes such as choosing local products, changing diet to be more environmentally friendly, offsetting emissions of flights, and avoiding air-travel made one factor made one factor, but larger lifestyle changes such as living car-free and having fewer children another was included as an additional factor – named lifestyle changes.
In S1, seven items were negatively skewed and were therefore squared before further analysis. Bollen-Stine bootstrapping (1000 times) revealed that the CFA model with private-sphere behaviours split into two factors, and the voting item loading on the smaller changes factor, had a good fit, Bollen-Stine p = .396.4 The factor loadings from the CFA model are shown in Table 6.
In S2, collective efficacy items were highly correlated (ranging from .86 to .91), and therefore the computed mean of the three items was included in the measurement model as a parcel. Behavioural intentions did not form reliable factors in a CFA model, and were therefore included as computed parcels. They were measured as a list of very different
behaviours, and could be seen as an index for behavioural intentions, scales that measures one latent construct.
Bollen-Stine bootstrapping (1000 times) revealed that the measurement model was valid for both groups, p = .111.5 The measurement model had construct-level metric invariance and equivalent construct variances and covariances. Merging the factors anger and guilt resulted in the exact same model fit and the two factors were very highly correlated (Pearson’s r = .98, p < .001). The full factors could not be included in a path model because of
multicollinearity, and therefore the single items measuring anger and guilt were used as proxies for the measurement of these emotions. This new measurement model had a good fit (Bollen-Stine p = .129).6
All other factors and parcels had acceptable discriminant validity. Factor loadings ranged from .55 to .91, four factor loadings were under the suggested .70. The low factor loadings were deemed acceptable because the overall model fit was good, and this model was used to compute latent variables for the following analysis.
Hypotheses testing
Because interaction effects were hypothesised, structural regression models were not feasible. Instead, structural equation path models, using the latent variables computed from the CFA model, were used to test the model displayed in Figure 1.
In S1, the model had multivariate kurtosis (c.r. = 26.86), and therefore bootstrapping was used to assess the model fit. After removing all insignificant paths, the final model had a very good model fit: χ2 (16) = 17.55, p = .351, CFI < .99, RMSEA = .02 95% CI [.00;.07], Bollen-Stine bootstrapping p = .542. For coefficients of the structural model, see Table 8.
5 Model fit: χ2 (316) = 1159.67, p = .000, CFI = .95, RMSEA = .04 95% CI [.04;.05], Bollen-Stine p = .111 6 Model fit: χ2 (387) = 555.085, p = .000, CFI = .92, RMSEA = .04 95% CI [.03;.05], Bollen-Stine p = .129
In S2, a good model fit was found, with no equality constraints between the Icelandic and Dutch groups.7 After removing insignificant effects on emotions, adding equality constraints where possible, the final model had a good fit: χ2 (86) = 93.35, p = .278, CFI < .99, RMSEA = .02 90% CI [.00;.04]. Constraining the effects of self-efficacy on emotions or behaviours to be equal resulted in a significant loss of model fit. The coefficients of the structural model are displayed in Tables 9 and 10.
In addition to the path models, Hayes PROCESS tool in SPSS was used as a complimentary analysis to confirm mediation effects through the distinct emotions on behaviours.
Results S1
Randomisation worked for gender, age, and education.8 The three conditions all differed significantly on perception of how positively their country was portrayed.9 Participants’ perception of how poorly their country is doing in mitigation differed significantly between the positive condition (M = 3.64, SD = 1.70) and the other two (Mneg = 4.46, SD = 1.23, p =
.004; Mcontrol = 4.37, SD = 1.40, p = .017)10, suggesting that the positive and negative
manipulations had the intended effects, but the control condition also resulted in negative evaluations of the country’s performance. The credibility of the news article was rated above the midpoint,11 but was significantly lower among the positive condition (M = 2.91, SD = 1.46), compared to the negative (M = 3.68, SD = 1.29, p = .002) and control (M = 3.86, SD = 1.24, p < .001) conditions.12
7 Model fit: χ2 (25) = 35.35, p = .894, CFI < .99, RMSEA = .00 95% CI [.00;.03], Bollen-Stine p = .915 8
Gender: χ2 (4) = 2.14, p = .710; Age: F (2,211) = 1.12, p = .328; Education: F (2,211) = 1.52, p = .222
9 Levene’s test for equality of variance showed that equal variances could not be assumed (Levene’s F [2,209] =
4.90, p = .008) and therefore a robust test of equality of means was conducted: Welch’s F (2,139.00) = 171.69, p < .000. All Games-Howell post-hoc differences were significant at p < .001
10
Levene’s F (2,209) = 4.08, p = .018); Welch’s F (2,137.41) = 5.69, p = .004
11 One sample t-test, using the test value 3: t(211)= 5.11, p < .001 12 F (2,211) = 10.18, p < .001
Appraisals and emotions
The research first sought to increase understanding on the antecedents of emotional reactions. A message about the in-group’s behaviour was expected to result in negative or positive
emotional reactions, depending on the valance of the message. The negative condition resulted in more anger and guilt than the positive and control
conditions and less pride, confirming H1.13 See means in Table 1.
Threat appraisal was expected to increase the
potential for emotion, thus increasing the strength of emotional reactions (H2). The path model (see Table 8) showed that threat appraisal had a significant positive effect on both anger (b = .17, p = .008) and guilt (b = .15, p = .016), but a negative effect on pride (b = -13, p = .028). Therefore, H2 was retained only for the negative emotions, but rejected for pride.
Both (a) self- and (b) collective efficacy were expected to interact with the negative condition, increasing its effects on both anger (H3) and guilt (H4). Although self-efficacy had a positive effect on anger (b = .12, p < .001) and guilt (b = .08, p = .003), no interactions were found and therefore H3a and H4a were rejected. Contrary to what was expected, collective efficacy had negative effects on anger (b = -.85, p < .001) and guilt (b = -.52, p = .016), so H3b and H4b were also rejected.
13 Anger: Levene‘s F = 7.93 (2,209), p < .001; Welch’s F (2, 136.58) = 27,99, p < .001 – Games-Howell
post-hoc test was used when equality of variances could not be assumed; Guilt: Levene‘s F = 0.19 (2,209), p = .804; F (2, 209) = , p < .001;
Pride: Levene‘s F = 5.82 (2,209), p = .003; Welch’s F (2, 136.46) = 47,35, p < .001
Table 1. Emotional reactions in S1. n M SD Anger Negative Control Positive 69 73 70 5.30 3.88* 2.61* 1.77 2.35 2.50 Guilt Negative Control Positive 69 73 70 4.61 4.01 2.27* 2.20 2.26 2.15 Pride Negative Control Positive 69 73 70 1.17 1.99* 4.61* 1.67 1.87 2.45 Note. *Bonferroni and Games-Howell post-hoc test revealed that mean is significantly different from the mean displayed above, p < .05
The positive condition had a significant effect on pride, but only for those with higher values of collective efficacy (b = 1.07, p < .001). H5, predicting that collective efficacy would positively moderate the effect of the positive condition on pride was therefore retained. Behavioural outcomes of emotions
The second aim of the study was to analyse the effects of the different emotions and how emotions act as mediators of messages on behaviour. The messages had direct effects on both public and private-sphere behaviours, but not on lifestyle changes. Relative indirect effects from PRCESS analysis are displayed in Table 2.
The negative message evoked anger (b = 1.58, p < .001) and the positive message resulted in less anger, compared to the control (b = -1.19, p < .001). Anger, in turn, had a positive effect on public sphere behaviour (b = 0.51, p < .001). Anger was found to mediate the effect of messages on public sphere behaviours, confirming H6.
Table 2.
Mediated effects of messages on behaviours through emotions, in S1.
Behaviour Message Emotion Effect (SE) 95% CI
Public sphere Negative message Anger Guilt Pride 0.75 (.31)* 0.04 (.13) 0.51 (.22)* 0.22; 1.45 -0.23; 0.31 0.15; 1.00 Positive message Anger Guilt Pride -0.56 (.28)* -0.10 (.27) -1.43 (.43)* -1.20; -0.11 -0.62; 0.48 -2.38; -0.66 Lifestyle changes Negative message Anger Guilt Pride -0.00 (.05) -0.01 (.02) 0.04 (.03) -0.10; 0.09 -0.06; 0.02 -0.00; 0.10 Positive message Anger Guilt Pride 0.00 (.04) 0.02 (.04) -0.12 (.07) -0.08; 0.07 -0.05; 0.12 -0.28; 0.00 Private-sphere Negative message Anger Guilt Pride 0.04 (.02)* 0.01 (.01) 0.04 (.02)* 0.004; 0.08 -0.01; 0.03 0.01; 0.07 Positive message Anger Guilt Pride -0.03 (.02)* -0.02 (.02) -0.12 (.03)* -0.07; -0.002 -0.05; 0.02 -0.19; -0.05 Note. These estimates are from PROCESS model 4. Threat appraisal, self-efficacy and collective efficacy were used as control variables in the models.
As expected, the negative message caused increased guilt (b = .71, p = .045) and the positive condition resulted in decreased guilt (b = -1.66, p < .001). Guilt, in turn, predicted private-sphere behaviours (b = .01, p = .048). This was not found to be a significant mediation of message effects. Guilt was found to have no effect on lifestyle changes. H7, predicting that guilt mediates the effects of messages on private-sphere behaviours, was not confirmed.
The positive condition increased pride for those with higher collective efficacy, as predicted in H5. Pride was found to have negative effects on all behavioural intentions, contrary to what was expected (public b = -.60, p < .001; private b = -.05, p < .001; public b = -.07, p < .001). H8 was therefore rejected. The PROCESS analysis showed that pride
mediated the negative effects of messages on public and private-sphere behaviour, but not on lifestyle changes.
Results S2
Randomisation worked for gender, age, education, and household income for both countries.14 The three conditions all differed significantly on perception of how positively their country was portrayed.15 Participants’ perception of how poorly their country is doing in mitigation did not differ significantly between conditions,16 suggesting that the manipulations did not have the intended effect of influencing the respondents’ evaluation of their in-group. The credibility of the news article was rated above the midpoint,17 but was significantly lower in the positive condition, compared to the negative and control conditions (p < .001).18
14 Iceland: Gender: χ2 (4) = 5.78, p = .216; Age: F (2,185) = 1.88, p = .155; Education: Levene’s F (2,185) =
7.81, p = .001, Welch’s F (2,2.91) = 2.91, p = .058; Income: F (2,180) = 1.58, p = .208
Netherlands: Gender: χ2 (4) = 3.77, p = .707; Age: F (2,81) = 0.58, p = .564; Education: Levene’s F (2,81) = 3.26, p = .043, Welch’s F (2,48.27) = 0.88, p = .423; Income: F (2,80) = 0.09, p = .915
15 F (2,271) = 115.85, p < .001, Bonferroni post-hoc differences were all significant at p < .001 16
Mpos = 3.81, SD = 1.23; Mneg = 4.19, SD = 1.60; Mcontrol = 3.80, SD = 1.45; F (2,271) = 2.07, p = .128
17 One sample t-test: t(271)= 8.04, p < .001, M = 3.67, test value = 3 18 M
Appraisals and emotions
The negative condition and control condition resulted in more anger and guilt and less pride than the positive,
confirming H1.19 Threat appraisal interacted with messages, so that it had a positive effect on anger for those who saw the negative message (b = 0,91, p = .028). Threat appraisal had a significant positive effect on guilt (b = 1.17, p < .001) and a negative effect on pride (b = -0.69, p < .001).
Consistent with S1, H2 was accepted for the negative emotions. See means in Table 3.
H3 suggested that (a) self- and (b) collective efficacy would increase the effects of the negative condition on anger. The effects of self-efficacy on emotions could not be set to be equal across groups. In the Icelandic group, self-efficacy had a positive effect on anger (b = 0.74, p <.001), but in the Dutch group it was found to have a negative effect (b = -0.42, p = .027). Self-efficacy was not found to interact with the messages, so H3a was rejected for both countries. Collective efficacy was not found to affect anger on its own or in interaction with conditions, H3b was therefore rejected.
Self-efficacy (a) and collective efficacy (b) were expected to strengthen the effect of the negative condition on guilt (H4). In the Icelandic group self-efficacy had a direct effect on guilt (b = 0.60, p = .002), but in the Dutch group this was not found (b = -.27, p = .121). Collective efficacy did not affect guilt in the model. No moderating effects of self- or collective efficacy were found on guilt, and H4 was therefore rejected.
19 Anger: F (2, 269) = 13.23, p < .001 Guilt: F (2, 269) = 12.44, p < .001 Pride: F (2, 269) = 31.50, p < .001 Table 3. Emotional reactions in S2. n M SD Anger Negative Control Positive 96 91 85 4.76 4.29 3.12* 2.40 2.14 2.04 Guilt Negative Control Positive 96 91 85 4.94 4.36 3.32* 2.24 2.29 2.15 Pride Negative Control Positive 96 91 85 -.79 -.26 1.55* 1.65 1.35 1.55 Note. *Bonferroni post-hoc test revealed that mean is
significantly different from the mean displayed above, p < .05
Collective efficacy was not found to affect pride as expected and therefore H5 was rejected. However, in the Icelandic group self-efficacy was found to increase pride for those in the positive condition (b = 0.60, p = .046).
Table 4.
Relative indirect effects of messages on behaviours through emotions, in S2.
Behaviour Message Emotion Effect (SE) 95% CI
Public sphere Negative message Anger Guilt Pride .07 (.04) .06 (.04) .05 (.04) -.00; .16 -.01; .16 -.02; .14 Positive message Anger Guilt Pride -.11 (.06)* -.07 (.05) -.10 (.07) -.24; -.01 -.18; .01 -.24; .03 Lifestyle changes Negative message Anger Guilt Pride .01 (.03) .03 (.04) -.02 (.04) -.05; .08 -.04; .12 -.11; .05 Positive message Anger Guilt Pride -.02 (.05) -.03 (.04) .05 (.07) -.12; .08 -.13; .05 -.09; .19 Private-sphere Negative message Anger Guilt Pride 07 (.05) .06 (.05) .06 (.04) -.00; .18 -.00; .18 -.00; .16 Positive message Anger Guilt Pride -.11 (.06)* -.08 (.05) -.12 (.07) -.25; -.01 -.19; .01 -.26; .01 Willingness to pay Negative message Anger Guilt Pride .07 (.05) .04 (.04) .02 (.04) -.00; .19 -.03; .13 -.05; .10 Positive message Anger Guilt Pride -.12 (.07)* -.05 (.07) -.04 (.07) -.27; -.02 -.15; .04 -.18; .08 Note. These estimates are from PROCESS model 4. Threat appraisal, self-efficacy and collective efficacy were used as control variables in the models.
*Mediation is significant at the .05 level Behavioural outcomes of emotions
Messages were found to have direct effects on behaviours (see coefficients in Tables 9 and 10). The second aim of the study was to see how distinct emotions could mediate the effects of messages on behaviours. Relative indirect effects from PROCESS mediation analysis are displayed in Table 4. Anger was found to mediate effects of seeing the positive condition on public sphere behaviour, private-sphere behaviour, and willingness to pay (WTP).
Guilt and pride were not found to mediate effects of messages on behaviours, and H7 and H8 are therefore rejected.
Discussion
Emotions and rational thought have often been seen as opposites, but this view is rejected by appraisal theories. Emotions can be viewed as a mechanism that helps humans cope with diverse situations and it has been hypothesized that certain emotions – rather than certain situations – are likely to cause certain ways of coping (Roseman, 2013; Roseman et al., 1994). This paper analysed whether different emotions in reaction to a message about climate change would result in different behavioural intentions, representing different coping strategies.
One of the two main aims of the study was to increase understanding on how emotions form as reactions to media messages. The paper analysed emotions toward the in-group, following up on findings from Harth et al. (2013), who found differing emotions toward the in-group to have different attitudinal outcomes. Membership in groups is an important determinant of self-image and identity, and messages about the in-groups’ performance were therefore expected to induce emotional reactions. Pre-existing appraisals that participants have in relation to climate change were measured, as these were expected to influence the formation of emotional reactions.
Anger and guilt were evoked by the negative message and pride by the positive message, as expected, in both studies. Primary appraisal, perception of climate change as threatening to individuals and their families, also predicted negative emotional reactions. In S2, primary appraisal moderated the effect of the messages, increasing the effects of the negative
condition on anger. Primary appraisal was not found to influence pride, but this may be due to the limited conceptualisation of primary appraisal to the individual as only perceived threat. Future studies should expand on this by measuring also the perceived threat to the
individuals’ norms and values – as people may perceive climate change as an important issue without appraisals of direct impacts on themselves.
Self-efficacy was found to increase anger and guilt for all participants, and it also increased pride for those who saw the positive message in the Icelandic group in S2. This is consistent with predictions of appraisal theories, which expect that when people feel that their actions can make a difference, they are likely to feel more strongly about an issue.
Collective efficacy was found to have a similar moderating effect on pride in S1, but surprisingly showed negative effects on anger and guilt. In S2, however, collective efficacy did not have any significant effects on emotions or directly on behaviours. This is surprising, as previous studies have found collective efficacy to be a stronger predictor for problem-focused coping and PEB than self-efficacy (e.g. Chen, 2015; Homburg & Stolberg, 2006). Jugert et al. (2016) found, “based on social identity theorizing that groups shape individual psychology through their capacity to define the individual self” (p. 16), that collective
efficacy works to increase PEB by increasing self-efficacy. If self-efficacy mediates the effect of collective efficacy on emotions and PEB, including self-efficacy in the model as another predictor may have lessened the effect found of collective efficacy. More analysis is needed to test whether this is the case.
The second main aim of the study was to see whether the emotions evoked by the
stimulus would mediate the effects of messages on pro-environmental behaviours in different ways. This is an important area of study, as knowing what motivates people to act can help communicators get across the urgency of climate change in more influential ways.
Guilt has been found to motivate private-sphere behaviours (Doosje et al., 1998; Ferguson & Branscombe, 2010) and result in willingness to donate to charity (Basil et al., 2006; Wonneberger, 2018), but was not found to predict any behavioural intentions in this study. Guilt is an uncomfortable feeling, in many ways similar to cognitive dissonance, and is
likely to result in motivation to reduce it (Stice, 1992). If reducing guilt is a primary
motivation, this is likely to result in minimal action that is sufficient for reduction. This study measured high-impact behaviours that are in many cases also difficult to carry out, and if guilt is likely to result in minimal behaviours that serve to reduce the feeling, then this may not have been captured by these measures.
Contrary to what was hypothesized, pride was found to have negative effects on all behavioural intentions in S1, and no effects in S2. Although this contradicts previous studies that have found people’s pride regarding previous actions and anticipated pride over future actions to positively influence behaviour (Bissing-Olson et al., 2016; Onwezen et al., 2013), this is not inconsistent with findings from Harth et al. (2013). Theyfound pride to increase support for “in-group favouring environmental protection,” which was measured with items such as “Germany alone cannot save the climate of the world; therefore we should not take more costs” (Harth et al., 2013, p. 23) – essentially measuring support for not acting to mitigate climate change.
There is some evidence that pride is not one uniform emotion, but a term that captures two very different phenomena – authentic pride and hubristic pride. This distinction could explain the very different outcomes found in studies on the effects of pride. Both types are thought to stem from appraisals that a positive event is caused by the self but are further distinguished by whether the pride is attributed to an unstable or stable internal cause. If the cause is perceived to be unstable and controllable, the pride is authentic, but if it is stable and global – i.e. “because I am great” – then the pride is hubristic (Tracy, Shariff, & Cheng, 2010). Thus, authentic pride regarding previous actions might result in PEB, while hubristic pride has no such effect or even a negative effect on PEB. Future research should pay attention to this distinction when measuring pride and analyse whether the different kinds of pride have differing behavioural outcomes.
Of the three emotions analysed, anger had by far the strongest effects on behavioural intentions. Anger consistently mediated the effects of messages, increasing public sphere behavioural intentions, as expected. Anger was also found to motivate private-sphere
behaviours and willingness to pay. Lerner and Keltner (2001) showed that when in an angry state, people make more optimistic risk judgements and choices, compared to people
experiencing fear. In the case of climate change, optimistic judgements of outcomes can be beneficial, as they decrease the chance of resorting to fatalism and increase the chance of problem-focused coping. This might increase the positive effect of anger on PEB.
Limitations
Following recommendations from Swim et al. (2011), this paper sought to analyse impact-oriented PEB. No existing scale was found that measures private-sphere intentions to mitigate climate change, and therefore a new scale was constructed for this. Due to the limited scope of this project, the scale could not be tested. Future research should amend this
shortcoming by developing a reliable scale of private-sphere PEB.
Another point of improvement would be to determine people’s pre-existing emotions towards climate change. This study asked participants about the emotions the stimulus evoked, but the message is far from being perceived in a vacuum. Climate change is a common topic in the media, and it is safe to assume that everyone that participated in the study had some pre-existing appraisals towards it, and already felt some degree of emotional response before being presented with the stimulus.
Finally, the limited resources of this project unfortunately affected the quality of the sampling. For convenience’s sake, the sampling was mostly obtained and administered via the researcher’s network, which limited its reach and size. Because of the possibility of self-selection, the sample might also have been unusually environmentally concerned. Another drawback of this technique is that it does not allow for response rate to be measured, and
therefore the extent of this bias cannot be estimated. The sample sizes are also smaller than is recommended for a model of this size, which leads to a low statistical power (Kline, 2011). Conclusion
The findings of these two studies show that a message about the performance of the respondents’ nation – the in-group – can motivate pro-environmental behavioural intentions. This effect was found to be mediated through anger, but not through guilt, and pride served only to reduce behavioural intentions. More research is needed to better understand the relationship between emotions and behaviour, but this paper has shown that emotions can be a crucial factor in the motivation of environmentally significant behaviours. This research has added to the literature of appraisal theories and of environmental communication by showing that anger serves as a motivator for action, not only in the political sphere but also in the private-sphere.
References
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist,
37(2), 122–147.
Basil, D. Z., Ridgway, N. M., & Basil, M. D. (2006). Guilt appeals: The mediating effect of responsibility. Psychology and Marketing, 23(12), 1035–1054.
https://doi.org/10.1002/mar.20145
Bissing-Olson, M. J., Fielding, K. S., & Iyer, A. (2016). Experiences of pride, not guilt, predict pro-environmental behavior when pro-environmental descriptive norms are more positive. Journal of Environmental Psychology, 45, 145–153.
https://doi.org/10.1016/j.jenvp.2016.01.001
Chen, M. F. (2015). Self-efficacy or collective efficacy within the cognitive theory of stress model: Which more effectively explains people’s self-reported proenvironmental behavior? Journal of Environmental Psychology, 42, 66–75.
https://doi.org/10.1016/j.jenvp.2015.02.002
Clarke, J., Heinonen, J., & Ottelin, J. (2017). Emissions in a decarbonised economy? Global lessons from a carbon footprint analysis of Iceland. Journal of Cleaner Production, 166, 1175–1186. https://doi.org/10.1016/j.jclepro.2017.08.108
Dietz, T., Stern, P. C., & Guagnano, G. A. (1998). Social structural and social psychological bases of environmental concern. Environment and Behavior, 30(4), 450–471.
https://doi.org/10.1177/001391659803000402
Doosje, B., Branscombe, N. R., Spears, R., & Manstead, A. S. R. (1998). Guilty by
association: When one’s group has a negative history. Journal of Personality and Social
Psychology, 75(4), 872–886. https://doi.org/10.1037/0022-3514.75.4.872
Ellsworth, P. C. (2013). Appraisal theory: Old and new questions. Emotion Review, 5(2), 125–131. https://doi.org/10.1177/1754073912463617
Feldman, L., & Hart, P. S. (2016). Using Political Efficacy Messages to Increase Climate Activism. Science Communication, 38(1), 99–127.
https://doi.org/10.1177/1075547015617941
Ferguson, M. A., & Branscombe, N. R. (2010). Collective guilt mediates the effect of beliefs about global warming on willingness to engage in mitigation behavior. Journal of
Environmental Psychology, 30(2), 135–142. https://doi.org/10.1016/j.jenvp.2009.11.010
Frijda, N. H. (2016). The evolutionary emergence of what we call “emotions.” Cognition and
Emotion, 30(4), 609–620. https://doi.org/10.1080/02699931.2016.1145106
Harth, N. S., Leach, C. W., & Kessler, T. (2013). Guilt, anger, and pride about in-group environmental behaviour: Different emotions predict distinct intentions. Journal of
Environmental Psychology, 34, 18–26. https://doi.org/10.1016/j.jenvp.2012.12.005
Homburg, A., & Stolberg, A. (2006). Explaining pro-environmental behavior with a cognitive theory of stress. Journal of Environmental Psychology, 26(1), 1–14.
https://doi.org/10.1016/j.jenvp.2006.03.003
Iyer, A., & Leach, C. W. (2008). Emotion in inter-group relations. European Review of Social
Psychology, 19(1), 86–125. https://doi.org/10.1080/10463280802079738
Jugert, P., Greenaway, K. H., Barth, M., B??chner, R., Eisentraut, S., & Fritsche, I. (2016). Collective efficacy increases pro-environmental intentions through increasing self-efficacy. Journal of Environmental Psychology, 48, 12–23.
https://doi.org/10.1016/j.jenvp.2016.08.003
Kline, R. (2011). Principles and Practice of Structural Equation modeling. (T. D. Little, Ed.) (3rd ed.). The Guilford Press.
Lazarus, R. S. (1991). Progress on a cognitive-motivational-relational theory of emotion.
American Psychologist, 46(8), 819.
explain willingness for political action. PSPB, 32(9), 1232–1245. https://doi.org/10.1348/014466606X99360
Leiserowitz, A., Smith, N., & Marlon, J. (2010). Americans’ Knowledge of Climate Change.
Yale Project on Climate Change Communication.
Lemer, J. S., & Keltner, D. (2001). Fear, Anger, and Risk. Journal of Personality and Social
Psychology, 81(1), 146–159.
Lewis, M. (2008). Self-Conscious Emotions: Embarrassment, Pride, Shame, and Guilt. In
Handbook of Emotions (pp. 742–756). https://doi.org/10.2307/2076468
Marcus, G. E., & Mackuen, M. B. (1993). Anxiety, Enthusiasm, and the Vote: The Emotional Underpinnings of Learning and Involvement During Presidential Campaigns. The
American Political Science Review, 87(3), 672–685.
Mauritsen, T., & Pincus, R. (2017). Committed warming inferred from observations. Nature
Climate Change, 7(9), 652–655. https://doi.org/10.1038/nclimate3357
Mayer, A., & Smith, E. K. (2018). Unstoppable climate change? The influence of fatalisstic beliefs about climte change on behavioural change and willingness to pay
cross-nationally. Climate Policy, 19(4), 511–523. https://doi.org/10.1080/14693062.2018.1532872
Meijers, M. H. C., Remmelswaal, P., & Wonneberger, A. (2018). Using Visual Impact Metaphors to Stimulate Environmentally Friendly Behavior: The Roles of Response Efficacy and Evaluative Persuasion Knowledge. Environmental Communication. https://doi.org/10.1080/17524032.2018.1544160
Morrison, S., Decety, J., & Molenberghs, P. (2012). The neuroscience of group membership.
Neuropsychologia, 50(8), 2114–2120.
https://doi.org/10.1016/j.neuropsychologia.2012.05.014
Research The Case for Emphasizing Discrete Emotions in Communication Research.
Communication Monographs, 77(2), 153–159.
https://doi.org/10.1080/03637751003790444
O’Neill, S., & Nicholson-Cole, S. (2009). “Fear Won’t Do It” Visual and Iconic Representations. Science Communication, 30(3), 355–379.
https://doi.org/10.1177/1075547008329201
Onwezen, M. C., Antonides, G., & Bartels, J. (2013). The Norm Activation Model: An exploration of the functions of anticipated pride and guilt in pro-environmental behaviour. Journal of Economic Psychology, 39, 141–153.
https://doi.org/10.1016/j.joep.2013.07.005
Onwezen, M. C., Bartels, J., & Antonides, G. (2014). The self-regulatory function of
anticipated pride and guilt in a sustainable and healthy consumption context. European
Journal of Social Psychology, 44(1), 53–68. Retrieved from
http://onlinelibrary.wiley.com/doi/10.1002/ejsp.1991/abstract
Roseman, I. J. (2013). Appraisal in the emotion system: Coherence in strategies for coping.
Emotion Review, 5(2), 141–149. https://doi.org/10.1177/1754073912469591
Roseman, I. J., & Smith, C. A. (2001). Appraisal theory: Overview, assumptions, varieties, controversies. In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal Processes
in Emotion: Theory, methods, research (pp. 3–19). New York: Oxford University Press.
Roseman, I. J., Wiest, C., & Swartz, T. S. (1994). Phenomenology, Behaviors and Goals Differentiate Discrete Emotions. Journal of Personality and Social Psychology, 67(2), 206–221.
Smith, N., & Leiserowitz, A. (2014). The role of emotion in global warming policy support and opposition. Risk Analysis, 34(5), 937–948. https://doi.org/10.1111/risa.12140 Stern, P. C. (2000). New Environmental Theories: Toward a Coherent Theory of
Environmentally Significant Behavior. Journal of Social Issues, 56(3), 407–424. https://doi.org/10.1111/0022-4537.00175
Stern, P. C., Dietz, T., Abel, T. D., Guagnano, G. A., & Kalof, L. (1999). Value belief norm theory of support for social movements: The case of environmentalism. Human Ecology
Review, 6(2), 81–97. https://doi.org/10.2307/2083693
Stice, E. (1992). The Similarities Between Cognitive Dissonance and Guilt : Confession as a Relief of Dissonance. Current Psychology: Research & Reviews, 11(1), 69–77.
Swim, J. K., Stern, P. C., Doherty, T. J., Clayton, S., Reser, J. P., Weber, E. U., … Howard, G. S. (2011). Psychology’s Contributions to Understanding and Addressing Global Climate Change. American Psychologist, 66(4), 241–250.
https://doi.org/10.1037/a0023220
Tajfel, H. (1982). Social Psychology of Intergroup Relations. Annual Review of Psychology,
33(1), 1–39. https://doi.org/10.1146/annurev.ps.33.020182.000245
Tajfel, H., & Turner, J. C. (1986). The Social Identity Theory of Intergroup Behavior. In
Political Psychology (pp. 276–293). https://doi.org/10.4324/9780203505984-16
Tracy, J. L., Shariff, A. F., & Cheng, J. T. (2010). A Naturalist’s View of Pride. Emotion
Review, 2(2), 163–177. https://doi.org/10.1177/1754073909354627
United Nations. (2015). Paris Agreement. Paris, France. https://doi.org/10.1017/s0020782900004253
Valentino, N. A., Brader, T., Groenendyk, E. W., Gregorowicz, K., & Hutchings, V. L. (2011). Election night’s alright for fighting: The role of emotions in political participation. Journal of Politics, 73(1), 156–170.
https://doi.org/10.1017/S0022381610000939
Valentino, N. A., Gregorowicz, K., & Groenendyk, E. W. (2009). Efficacy, emotions and the habit of participation. Political Behavior, 31(3), 307–330.
https://doi.org/10.1007/s11109-008-9076-7
Valentino, N. A., Hutchings, V. L., Banks, A. J., & Davis, A. K. (2008). Is a Worried Citizen a Good Citizen? Emotions, Political Information Seeking, and Learning via the Internet.
Political Psychology, 29(2), 247–273. Retrieved from
http://www.jstor.org/stable/20447114
van Zomeren, M., Spears, R., & Leach, C. W. (2010). Experimental evidence for a dual pathway model analysis of coping with the climate crisis. Journal of Environmental
Psychology, 30(4), 339–346. https://doi.org/10.1016/j.jenvp.2010.02.006
Wallace-Wells, D. (2019). The Uninhabitable Earth: Life After Warming. New York: Tim Duggan Books.
Wonneberger, A. (2018). Environmentalism—A Question of Guilt? Testing a Model of Guilt Arousal and Effects for Environmental Campaigns. Journal of Nonprofit and Public
Sector Marketing, 30(2), 168–186. https://doi.org/10.1080/10495142.2017.1326873
Wynes, S., & Nicholas, K. A. (2017). The climate mitigation gap: education and government recommendations miss the most effective individual actions. Environmental Research
Appendix A Stimulus Materials Stimulus Study 1
[Country] was be replaced by the country that the participant reported coming from, in the pre-questionnaire. The differences between the condition texts are underlined, but this emphasis was not present in the experiment. Before reading the news story, the participants was this text: “Please read the following news article. You will be able to continue to the next page after 30 seconds.”
Negative in-group portrayal:
[Country]’s Carbon Footprint is Big and Getting Bigger
A recent study of the ecological impact of 50 countries worldwide shows that the carbon footprint of [country] is among the largest in the world.
Carbon footprint is the amount of greenhouse gasses that is released in each country per capita and it’s a good measure for how much a country contributes to accelerating global warming, which has already reached 1.1°C, compared to pre-industrial levels. The countries with the biggest carbon footprint are the biggest polluters and most to blame for climate change.
[Country] is one of the top 10 emitters in the study, with 30.1 tons of CO2 emissions a year per capita. This is worse than in the last measure, in 2017, when 28.3 tons were emitted.
The carbon footprint of [Country] is much larger than that of the lowest emitter, Ethiopia (0.1 tons of CO2 equivalents per person), but somewhat smaller than that of the top emitter, Qatar (45.4 tons of CO2 equivalents per person).
This bad score is caused by the large amount of fossil fuels burned to power people’s cars and various industries. Also to blame is the large amount of goods imported to the country from across the globe.
Positive in-group portrayal:
[Country]’s Carbon Footprint is Small and Getting Smaller
A recent study of the ecological impact of 50 countries worldwide shows that the carbon footprint of [country] is among the smallest in the world.
Carbon footprint is the amount of greenhouse gasses that is released in each country per capita and it’s a good measure for how much a country contributes to slowing down global warming, which has already reached 1.1°C, compared to pre-industrial levels. The countries were ranked in order of smallest to biggest users of greenhouse gasses.
[Country] is in the top 15 of all 50 countries in the study, with mere 12.1 tons of CO2 emissions. This is better than in the last measure, in 2017, when 12.7 tons were emitted.
The carbon footprint of [Country] is far less than that of the top emitter, Qatar (45.4 tons of CO2 equivalents per person), but somewhat higher that of the lowest measured, which was Guatemala (2.9 tons of CO2 equivalents per person).
This good position is partly due to the large amount of renewable energy produced and used to service domestic buildings and industries. Carbon-reducing initiatives by individuals, corporations, and the government have also been successful at reducing [country]’s carbon footprint.
Neutral in-group portrayal:
Recent Study Analyses [Country]’s Carbon Footprint
A recent study of the ecological impact of 50 countries worldwide shows that the carbon footprint of [country] is similar to neighbouring countries.
Carbon footprint is a good measure for how much a country contributes to global warming, which has already reached 1.1°C, compared to pre-industrial levels. It is the amount of greenhouse gasses that is released in each country, per person. The study included a list of countries by the size of their carbon footprint.
