Predicting Pro-environmental Intention and Behavior Based on Justice Sensitivity, Moral Disengagement, and Moral Emotions – Results of Two Quota-Sampling Surveys

Nicolai, Susanne; Franikowski, Philipp; Stoll-Kleemann, Susanne

doi:10.3389/fpsyg.2022.914366

ORIGINAL RESEARCH article

Front. Psychol., 24 June 2022

Sec. Environmental Psychology

Volume 13 - 2022 | https://doi.org/10.3389/fpsyg.2022.914366

This article is part of the Research Topic Current Trends in Environmental Psychology, volume I View all 35 articles

Predicting Pro-environmental Intention and Behavior Based on Justice Sensitivity, Moral Disengagement, and Moral Emotions – Results of Two Quota-Sampling Surveys

$\r\nSusanne Nicolai*$ Susanne Nicolai^1*

Philipp Franikowski²

Susanne Stoll-Kleemann¹

¹Department of Sustainability Science and Applied Geography, Institute of Geography and Geology, University of Greifswald, Greifswald, Germany
²Department of Personality Psychology and Psychological Assessment, Institute of Psychology, University of Greifswald, Greifswald, Germany

The effects of climate change lead to increasing social injustice and hence justice is intrinsically linked to a socio-ecological transformation. In this study, we investigate whether justice sensitivity motivates pro-environmental intention (PEI) and behavior (PEB) and, if so, to what extent emotions and moral disengagement determine this process. For this purpose, we conducted two quota-sampling surveys (Study 1: N = 174, Study 2: N = 880). Multiple regression analyses in both studies suggest that a higher perception of injustice from a perpetrator’s, beneficiary’s, and observer’s perspective is associated with an increased PEI. However, moral disengagement best predicted PEB and PEI. Guilt and authentic pride were found to be emotional predictors of PEI. Additionally, mediation analyses demonstrated that guilt mediates the connection between both perpetrator and beneficiary sensitivity and PEI. These results suggest that when the predominant originators of climate change (i.e., individuals from industrialized countries) perceive global climate injustice from the perspective of a beneficiary or a perpetrator, they experience guilt and have a higher PEI. Based on this mechanism, it seems promising to render global injustice more salient to those responsible for activities that lead to climate change to motivate them to adapt their behavior. The role of moral disengagement and victim sensitivity as barriers to pro-environmental behavior is discussed in this context.

Introduction

The latest report by the Intergovernmental Panel on Climate Change [IPCC] (2022) suggests that it is not only natural conditions that determine who is most affected by the climate crisis, but rather the living conditions of the individuals. Limited economic resources, suffering from inequality, enduring political instabilities and wars, or limited access to basic resources such as clean water are only a few examples of disadvantages faced by individuals due to climate change. Ultimately, human rights, such as the right to life, health, and subsistence are endangered by the consequences of climate-damaging behavior, i.e., behavior that emits a lot of greenhouse gases and, therefore, forces climate change. Due to unequally distributed resources (e.g., money, political power), climate change disproportionally affects the Global South, younger generations, women, and people with low income more intensely than the Global North, older generations, men, and people with high income (Intergovernmental Panel on Climate Change [IPCC], 2022). In all of these cases, the same disproportion can be found, namely that the individuals who are least responsible for climate change are more affected by its consequences than those most responsible for it. For example, from 1990 to 2015, the global wealthiest 10% were responsible for 46% of emission growth whereas the global poorest 50% were responsible for only 6% of the total emission growth (Kartha et al., 2020). As a consequence, discussions concerning climate change comprise ethical and moral issues regarding justice (Caney, 2010; Boom et al., 2016; De Smet et al., 2016; Alves and Mariano, 2018). Therefore, the term climate justice incorporates various justice principles in the domain of climate change. These include distributive justice, which considers the allocation of costs and benefits among people, procedural justice, which points out who can participate in decision-making processes, and recognition, which involves respectful and appropriate “consideration of diverse cultures and perspectives” (Intergovernmental Panel on Climate Change [IPCC], 2022, p. 9).

Justice was found to be a fundamental human motive (Montada, 2007; Baumert et al., 2013a). Therefore, we hypothesized that justice is not only the target but also a cause of pro-environmental intention, and, in turn, a motivator of pro-environmental behavior. Hence, we investigated whether justice sensitivity motivates individuals to behave (more) pro-environmentally. In this article, we apply findings from justice sensitivity research to the context of the current climate crisis and investigate the predictive power of justice sensitivity toward the intention of a pro-environmental behavioral change. Moreover, we investigate the associations of pro-environmental behavior change with moral emotions and moral disengagement.

In the context of climate justice, people in industrialized countries can be regarded as the beneficiaries of the detrimental living conditions of people from less developed countries, since the standard of living in industrialized countries is directly connected to the exploitation of natural and human resources in less developed countries (Dorninger et al., 2021; Hickel et al., 2022). However, many people may not be completely aware of this fact as the actions that result in driving climate change are probably perceived as an “unintentional, if unfortunate, side effect” of goal-directed behavior (Markowitz and Shariff, 2012). Markowitz and Shariff (2012) call this phenomenon the blamelessness of unintentional action. Unintentionally caused consequences are judged less harshly than equally severe but intentionally caused consequences (Guglielmo et al., 2009). However, the existence or lack of an intention does not change the consequences of an action. Therefore, we hypothesize that people need to identify themselves as responsible to renounce their privileges and act in solidarity with the disadvantaged. Moreover, we assume that this perception is a motivator for pro-environmental behavior change and, therefore, is pivotal to achieving climate justice. Consequently, we focus on justice sensitivity in this article. In the following, we contextualize the role of justice sensitivity on pro-environmental behavior and intention by investigating the joint prediction of pro-environmental behavior and intention with justice sensitivity, emotions, and moral disengagement.

Justice was found to be a fundamental human motive (Montada, 2007; Baumert et al., 2013a). That is, individuals generally seek to establish justice while avoiding injustice (Lerner, 1977), and hence individuals want themselves and others to be treated fairly and are willing to behave according to justice principles (Baumert et al., 2013b). To establish justice, individuals are willing to give up their own advantages (Engel, 2011), act sustainably (Kals and Becker, 2006), or reject a beneficial but unfair deal (Sutter et al., 2020). A meta-analysis based on 182 studies on collective behavior confirmed that the more strongly individuals perceive a situation to be unjust, the more likely they are to participate in collective action (such as protest behavior) to improve the situation (van Zomeren et al., 2008). Baumert et al. (2013b) summarize:

Assuming that justice is a fundamental motive for individuals means that the perception of a potential injustice triggers emotional reactions (e.g., anger, moral outrage, compassion, guilt) and urges the individual to act in order to restore justice or to avoid the injustice. Hence, the concept of a human justice motive implies the assumption of a psychological link between the perception of (potential) injustice and affective and behavioral reactions (p. 161).

Although justice is a fundamental motive, individuals differ in their judgments concerning whether a situation (e.g., the climate crisis) is just or unjust. These differing assessments were found to be due to different principles of justice that vary across different individuals, contexts, situations, and relationships (Deutsch, 1985; Jasso et al., 2016; Sabbagh and Schmitt, 2016; Skitka et al., 2016). In this article, we focus on the first aspect, namely individuals. Dispositional interindividual differences were found to influence the perception of justice, i.e., differences in justice sensitivity (Lerner, 1977; Münscher, 2017; Preiser and Beierlein, 2017). Justice sensitivity is regarded as a personality trait (Schmitt et al., 1995), as it was found to be consistent across situations and contexts (Schmitt et al., 2009). Justice sensitivity is defined as the tendency to perceive injustice that results in negative emotional responses. Together, the perception of injustice and its negative evaluation form action tendencies to mitigate the injustice (Baumert and Schmitt, 2016). This process is in line with the appraisal theory developed by Arnold (1960), which states that emotions, i.e., action tendencies, are the result of two cognitive processes: the factual cognition (here: Is the situation just or not?) and the evaluative cognition (here: evaluating injustice as negative – or positive).

Individuals react with different intensity to injustices they have suffered themselves (as victims), to injustices they perceive in everyday situations (as observers), to injustices from which they passively benefit (as beneficiaries), and to injustices they have committed themselves (as perpetrators; Montada and Maes, 2016). Although justice sensitivity is inherent to all four perspectives, they differ in several ways. On the one hand, a high justice sensitivity as an observer, beneficiary, or perpetrator reflects the desire for justice for others and a feeling of social responsibility. On the other hand, a high score in justice sensitivity as a victim reflects a desire for justice for oneself (Preiser and Beierlein, 2017). More precisely, individuals who score highly in perpetrator or beneficiary sensitivity rather experience prosocial concerns (e.g., existential guilt, social responsibility, or solidarity with the disadvantaged; Gollwitzer et al., 2005). Observer sensitives were found to feel empathy and, as a result, take responsibility (Schmitt et al., 2005). Those who perceive injustice from the perspective of the beneficiary, observer, or perpetrator are more likely to feel responsible (Ehrhardt-Madapathi et al., 2018). A greater sense of responsibility, in turn, is a promising predictor to engage in pro-environmental behavior (Kollmuss and Agyeman, 2002). As already indicated, victim sensitivity relates differently and in part contrarily to the other three perspectives. In terms of behavioral outcomes, victim sensitivity was positively related to delinquent behavior (e.g., fare or tax evasion; Preiser and Beierlein, 2017). Victim sensitives are afraid of being disadvantaged and behave in anticipation of such an injustice. Therefore, they easily abandon their moral standards (Gollwitzer et al., 2009; Schmitt et al., 2009; Münscher, 2017; Preiser and Beierlein, 2017), are rather jealous, neurotic, socially mistrusting, and paranoid (Schmitt et al., 2005) and fear the exploitation of their investments (Rothmund et al., 2014). To summarize, we can distinguish three pro-social sensitivities (perpetrator, beneficiary, and observer) from one pro-self sensitivity (victim). We assume that the three pro-social sensitivities are positively associated whereas the pro-self sensitivity is negatively associated with pro-environmental intention and behavior.

These associations are mirrored in emotional dispositions as individuals with a higher score in beneficiary or perpetrator sensitivity are more prone to experience guilt than those with a lower score. In contrast, individuals with a higher score in observer sensitivity tend to more easily be outraged than those with a lower score. Finally, individuals with a higher score in the area of victim sensitivity are inclined to feel more anger than those with a lower score (Preiser and Beierlein, 2017).

Emotions – especially, moral emotions – affect environmentally relevant behavior (Kals and Maes, 2002; Hahnel and Brosch, 2018; Harré, 2018; Landmann, 2020), and, therefore, are of particular interest in the area of climate injustice. We refer to moral emotions, as they “inhibit socially undesirable behavior and foster moral conduct” (Tangney, 1995). For example, in a study by Roeser (2012), in climate change communication, emotions promoted a better understanding of the moral meaning of climate change in comparison to neutral communication. As a consequence, emotions are a reliable motivator for pro-environmental behavior (Roeser, 2012). Corral-Verdugo (2013) even found that emotions have a higher impact on pro-environmental behavior than intentions. More precisely, emotions and intentions together explained more than half of the variance in pro-environmental behavior. In addition, emotions play a vital role in initiating behavioral change (Stoll-Kleemann et al., 2022a). Kollmuss and Agyeman (2002) state in their behavior change model that “the stronger a person’s emotional reaction, the more likely that person will engage in a new behavior.”

When speaking about the influence of emotions on pro-environmental behavior, we can distinguish between (1) emotions that are currently experienced, (2) emotions that people believe their actions will elicit (anticipated emotions), and (3) emotions that are generalized due to multiple repetitions. Venhoeven et al. (2013) also address the emotions that pro-environmental behavior elicits, although their findings are not relevant to this study. The three aforementioned emotion groups cause different effects. First, experimentally induced current emotions, e.g., by confronting participants with environmental damages to elicit guilt (Reese and Jacob, 2015), affected the pro-environmental intention or behavior of these individuals. However, these effects were found to be rather small, vary individually, and not remain stable over a period of time (Landmann, 2020, p. 10). Second, anticipated emotions can motivate people to engage in pro-environmental behavior as they either believe it will make them experience pleasure (Rezvani et al., 2017) or prevent them from experiencing displeasure (Onwezen et al., 2013). Third, according to the model of affect generalization (Paolini et al., 2016), emotional episodes can generalize when experienced sufficiently often or with high intensity. In the context of pro-environmental behavior, Landmann (2020) states that emotions only affect behavioral intentions if they are generalized. In line with the model of affect generalization, the emotions used in this work are not used as an emotional episode (currently experienced or anticipated), but rather as an affective attitude, that is, the proneness to feel these emotions.

In this study, we investigate the role of the moral emotions (1) pride, (2) gratitude, (3) guilt, and (4) shame. Considering the seven categories of environmentally relevant emotions by Landmann (2020), pride is one of the self-praising emotions which occur as a result of positive norm deviations. Gratitude belongs to other-praising emotions which are triggered by observing positive norm deviations in others. Guilt and shame cover self-condemning emotions that arise due to individual violations of personal norms. These different emotion types lead to different action tendencies. Whereas self-praising emotions reinforce one’s own positive behavior, e.g., through in-group favoring pro-environmental intentions, other-praising emotions lead to support for the source of these emotions, e.g., protecting nature. Self-condemning emotions motivate people to correct their behavior, e.g., by repairing environmental damage (Landmann, 2020).

Rezvani et al. (2017) found that when people anticipated feeling pride due to their pro-environmental behavior, they had a higher pro-environmental intention than individuals without this anticipation. Additional to this direct effect, anticipated pride also mediated the effect of personal moral norms on pro-environmental intentions. Experienced pride in environmental behavior increased an individual’s engagement in pro-environmental behavior (Bissing-Olson et al., 2016). However, two types of pride can and should be distinguished: authentic and hubristic pride. Authentic pride, on the one hand, focuses on the “self-in-action,” e.g., pro-environmental behavior, and is therefore tied to specific situations. Hubristic pride, on the other hand, focuses on the “self-as-actor” and generalizes across situations. This difference impacts behavior: whereas authentic pride positively predicted moral behavior, hubristic pride was negatively related to it (Krettenauer and Casey, 2015).

Gratitude is benefit-triggered and can be considered a trait and therefore may serve as a second positive emotion. Tam (2022) found that gratitude correlated with pro-environmental intention and behavior. However, manipulations of gratitude toward nature did not have robust effects on pro-environmental behavior change.

Guilt and shame due to environmental damage were consistently found to influence pro-environmental intention and behavior (Carrus et al., 2008; Ferguson and Branscombe, 2010; Gausel and Brown, 2012; Harth et al., 2013; Rees et al., 2014). When individuals are highly perpetrator- or beneficiary sensitive, they are also prone to feel guilty (Montada et al., 1986). Guilt results from regretting a specific behavior afterward but does not affect one’s self-identity. Shame, however, affects one’s self-identity as it goes along with a feeling of worthlessness and powerlessness (Tangney, 1995). Therefore, guilt tends to be tied to a specific behavior (similar to authentic pride), whereas shame is linked to self-identity (similar to hubristic pride). Rees et al. (2014) found that both emotions predicted positive behavioral intentions, but it was shame and not guilt that transformed the intention into actual behavior. However, Carrus et al. (2008) showed that anticipated guilt strongly affected an individual’s desire to use public transportation and engage in household recycling. The relationship between negative anticipated emotions and the pro-environmental intention was even stronger than that between attitudes and pro-environmental intention. Most research on emotions in pro-environmental behavior focused on guilt and pride. In direct comparison, pride and guilt were both significantly related to pro-environmental behavior and intention to engage in pro-environmental behavior. Shipley and van Riper (2022) found that anticipated pride and guilt had equal effects on pro-environmental behavior, but that experienced guilt had stronger effects on pro-environmental behavior than experienced pride. However, negative emotions must be dealt with consciously, as they result in aversive action tendencies. This is evident in the case of cognitive dissonance (Festinger, 1957). According to the theory of cognitive dissonance, emotional stress occurs when actions and beliefs are inconsistent. This emotional stress is highly uncomfortable and needs to be resolved. Based on Festinger’s theory, Bandura (2007, 2016) developed the concept of moral disengagement. Moral disengagement can be seen as a cognitive restructuring of a situation to justify inaction or subjectively wrong action (Bandura et al., 1996; Bandura, 2007, 2016) and does not only occur in individuals, but also in groups of individuals when harmful behavior is collectively morally justified (Bandura, 2002). These justification mechanisms include denial and diffusion of responsibility. Euphemistic labeling, advantageous comparison, in addition to minimizing, ignoring, or misconstruing the consequences are also forms of moral disengagement. Engaging in these strategies can be prevented by a high degree of moral self-regulation, that allows individuals to remain consistent with their standards of justice. Such a higher moral self-regulation and, hence, less moral disengagement, was found in perpetrator and beneficiary sensitive individuals in a longitudinal study by Maltese and Baumert (2016). Bandura (2007) already theorized how moral disengagement selectively contributes to increasing environmental damage. These mechanisms prevent individuals from engaging in pro-environmental behavior, even if they have already recognized the injustice of their behavior. Some of the mechanisms were investigated in a qualitative study (Stoll-Kleemann and O’Riordan, 2020) while the association of moral disengagement with high carbon behavior was found in a quantitative study (Stoll-Kleemann et al., 2022b). It is assumed that moral disengagement in high carbon behavior correlates positively with victim sensitivity, as both tendencies are self-serving and prevent an individual from taking responsibility for the environment. Therefore, moral disengagement is also assumed to correlate negatively with pro-environmental intention and behavior.

As argued above, pro-environmental behavior change is necessary to achieve climate justice. We hypothesize that climate justice is not only the ultimate goal of pro-environmental behavior but also its cause as justice functions as a motivator of behavior. To our knowledge, justice sensitivity perspectives have not yet been associated with pro-environmental intention and behavior. Based on the reported findings, the aim of this study is to investigate the extent to which perceptions of injustice predict pro-environmental intention and, in turn, might even predict pro-environmental behavior. More precisely, we assume that the pro-social justice sensitivities (perpetrator, observer, and beneficiary) are positively associated with pro-environmental intention (H1a). The pro-self justice sensitivity (victim) and the tendency to morally disengage are assumed to be negatively associated with pro-environmental intention (H2a). Furthermore, we assume an influence of moral emotions (guilt, shame, authentic and hubristic pride, and gratitude) on pro-environmental intention (H3b).

We also checked whether the three hypotheses also apply for pro-environmental behavior (H1b, H2b, and H3b). However, one must note that pro-environmental intentions might not be implemented in pro-environmental behavior and therefore may not account for all of the variance in pro-environmental behavior. This issue is known as the intention-behavior gap (Sheeran, 2002). If the predictions of Hypotheses 1–3 apply for pro-environmental intention, they might not necessarily apply to pro-environmental behavior.

As guilt results from both high perpetrator and beneficiary sensitivity, and the action tendency of guilt is to correct one’s behavior, we assume that guilt mediates the relationship between beneficiary (H4a) and perpetrator (H4b) sensitivity and pro-environmental intention. Moreover, we hypothesize an association between moral disengagement and victim sensitivity (H5). Hypotheses 1-3 can be found in Figure 1.

FIGURE 1

Figure 1. Hypotheses 1–3 (a and b) of the current Study.

Study 1 was conducted to address Hypotheses 1 (a and b), 2 (a and b), and 5. Albeit the fact that pro-environmental intention cannot be regarded as a promising predictor for (future) pro-environmental behavior (Sheeran, 2002), past pro-environmental behavior was found to be a strong predictor for pro-environmental intentions (e.g., Bamberg et al., 2003; Carrus et al., 2008; Manca et al., 2020). As pro-environmental behavior was assessed retrospectively (as “past pro-environmental behavior”) in Study 1, we also checked if this finding can be replicated with our data (see Figure 1).

Study 2 addressed Hypotheses 3 (a and b) and 4 (a and b) and aimed to replicate the findings of Study 1 with a larger sample and a potentially more precise measure of (current) pro-environmental behavior, i.e., the “carbon footprint.”

Study 1

Materials and Methods

Participants

We recruited 220 German panelists on meinungsplatz.de. Meinungsplatz.de is a German survey platform offering a pool of 250,000 active participants from all social strata in Germany and Switzerland for market and opinion research purposes. All participants were compensated for their expenses following the policy of the panel service.

Quotas were defined concerning age and gender to ensure representativeness regarding these socio-demographic dimensions. The scheduled age and gender proportions were based on the Federal Statistical Office Germany (2019). Three panelists responded that they were 16 or 17 years old. We initially planned to only include data from participants that were at least 18 years old as these were requested from the panel provider. However, we decided to retain the data from these participants for the analyses as their participation is in accordance with the German legislation and we did not expect their response pattern to differ substantially from the participants that were at least 18 years old. The gender distribution in Germany is nearly equal (50.7% female, 49.3% male; Federal Statistical Office Germany, 2019) and therefore was assumed to be 50:50. Due to an unknown distribution of individuals who identify themselves as non-binary, these were not included (see Table 1).

TABLE 1

Table 1. Proportions of German age groups (in 2018) and the derived sampling goals along with the final sample proportions in Study 1.

As exclusion criteria, we consulted the relative speed index (RSI; Leiner, 2019) and a social desirability score (Satow, 2012). The RSI is computed by dividing the median page completion time of the sample by the page completion time of the respective individual. A relative speed index of 2 indicates that an individual was twice as fast as the median of the total respondents. According to the recommendation of Leiner (2019), we removed 46 panelists with RSI > 2 as we assumed them to not have worked conscientiously. Additionally, we excluded 12 participants due to socially desirable response tendencies as they scored 7 or higher on the social desirability scale (see Instruments; Satow, 2012).

Therefore, the final sample for Study 1 comprised 174 participants: 91 females and 83 males (age M = 49.9, SD = 17.5, range = 16–80 years). Two χ²-tests for goodness of fit for age groups and gender, respectively, found no significant deviation from the proportions set for the recruiting procedure, either for gender, χ²(1) = 0.37, p = 0.544 or for age group, χ²(5) = 2.67, p = 0.751. The resulting sample therefore still represents the distribution of age and gender in the German population as a whole.

Sensitivity analyses using G*Power (Faul et al., 2007) indicated that with the given sample size of N = 174, α = 0.05, and a power of 1 – β = 0.80, in a linear multiple regression with five predictors (four justice sensitivities, and moral disengagement in high carbon behavior), small effect sizes of R² = 0.071 and with a power of 1 – β = 0.95 small effect sizes of R² = 0.105 could be detected.

Instruments

To test our hypotheses, we measured two dependent variables, pro-environmental behavior, and pro-environmental intention. The independent variables were the justice sensitivity perspectives and moral disengagement in high carbon behavior. The descriptive statistics and internal consistency measures of all variables of Study 1 can be found in Table 2.

TABLE 2

Table 2. Descriptive statistics and internal consistency of all measures in Study 1.

Past Pro-environmental Behavior

To measure the first dependent variable, 10 pro-environmental behaviors were queried for the last year (i.e., 2019 as the survey was conducted in 2020). Due to the pandemic state of emergency during the survey period, we did not ask about the current behavior as this may have been affected by the COVID-19 pandemic. Therefore, we measured past pro-environmental behavior in Study 1. The participants were asked to indicate whether they pursued the specified behavior (“yes,” “no,” or “do not know”). To prevent an acquiescence bias (the tendency to answer with “yes” regardless of the question), some items were framed pro-environmentally and others from an environmentally damaging perspective. An example of an item is “In 2019, I purchased green electricity.” Based on the idea of the Campbell Paradigm (Kaiser et al., 2013), we addressed behaviors of varying “difficulty.” For example, we expected it to be more difficult to use green electricity (because it is more expensive) than to lower the washing temperature. The inverted items were subsequently reversed, and the relative frequency of pro-environmental behavior was computed for each person. The internal consistency of the scale was Cronbach’s α = 0.41. As the measure can be regarded as a formative measure (the latent construct is formed by the items) instead of a reflective measure (the latent construct is causal for the item responses), internal consistency estimates for reliability coefficients should not be interpreted and not considered for item selection (e.g., Bollen and Lennox, 1991; Rossiter, 2002). The scale can be found in Appendix A.

Pro-environmental Intention

To measure the second dependent variable, the same 10 items from the past pro-environmental behavior scale were re-formulated as intentions for the following year (i.e., 2021). For example, one item taken from the section above reads: “In 2021, I would like to (continue to) purchase green electricity.” As we asked for behavior intentions rather than actual behavior, answers were given on 6-point scales that expressed the participant’s approval of the items (ranging from “totally disagree” to “totally agree”). The internal consistency of the scale was Cronbach’s α = 0.67. The questionnaire can be found in Appendix B.

Justice Sensitivity (USS-8)

The injustice sensitivity scales (USS-8) of Beierlein et al. (2012) measure the four perspectives of justice sensitivity (victim, perpetrator, beneficiary, and observer). Each perspective is measured by two items, answered on a 6-point scale (ranging from “does not apply at all” to “fully applies”). An item example of beneficiary sensitivity is: “I feel guilty when I am undeservedly better off than others.” The reliabilities of perpetrator (Cronbach’s α = 0.82), beneficiary (Cronbach’s α = 0.89), and victim sensitivity (Cronbach’s α = 0.81) were good or even excellent; however, the reliability of observer sensitivity was questionable (Cronbach’s α = 0.68).

Moral Disengagement in High Carbon Behavior Scale

For Study 1, based on the questionnaire of Moore et al. (2012) on moral disengagement in general, we developed a German questionnaire that measures an individual’s propensity to morally disengage in the context of high carbon behavior (MD-HCB; Stoll-Kleemann et al., 2022b). The MD-HCB covers nine mechanisms of moral disengagement following Markowitz and Shariff (2012) and Bandura (2016) with two questions each (18 items in total). An exemplary item is “In terms of my carbon emissions, I don’t want to give much consideration to people who live very far away and whom I will never meet.” The reliability was excellent, with Cronbach’s α = 0.96, and the measure was found to be valid in terms of factorial, criterion, convergent, and discriminant validity (see Stoll-Kleemann et al., 2022b).

Social Desirability

To detect test falsification due to positive self-presentation and socially desirable response tendencies, the short version of social desirability scale (Satow, 2012) was used. The participants answered two items on a 4-point scale and thus the sum score ranged from 2 to 8. A value of 7 or higher indicated a high tendency to distort self-presentation and it is thus recommended to exclude participants scoring 7 or higher. Based on this criterion, we had to exclude 12 participants prior to the analysis. The final sample consisted of N = 174 participants, as described above.

Sociodemographic variables such as gender, age, education level, and working situation were queried directly.

Design and Procedure

We used SoSci Survey (Leiner, 2020) to prepare and host the online survey. The questionnaires analyzed in this article were part of a larger data collection comprising further questionnaires on moral constructs (see Stoll-Kleemann et al., 2022b). Starting with the sociodemographic variables, the questionnaires were answered in the following order: moral competence¹, past pro-environmental behavior, moral disengagement in high carbon behavior, pro-environmental intention, moral disengagement in general¹, idealism and relativism¹, justice sensitivity, moral pride¹, empathy and perspective taking¹, moral identity¹, and Machiavellianism¹. Answering all questions took the participants between 10 and 34 min (Mdn = 19 min). The findings were considered significant in two-sided testing when p < 0.05. All analyses were conducted using R (R Core Team, 2020). In addition to established packages, the following were used: bda (Wang, 2021), car (Fox and Weisberg, 2019), dplyr (Wickham et al., 2022), skedastic (Farrar, 2020), and lm.beta (Behrendt, 2014).

Results

To test the assumed positive prediction of the pro-social justice sensitivities and the assumed negative prediction of victim sensitivity and moral disengagement according to Hypotheses 1 and 2, we conducted two regression analyses with (a) pro-environmental intention and (b) past pro-environmental behavior as dependent variables, respectively. We included perpetrator, beneficiary, observer, and victim sensitivity as well as moral disengagement in high carbon behavior as predictors (Model A: on pro-environmental intention, Model B: on past pro-environmental behavior).

The prerequisites of no autocorrelation and normal distribution of residuals were given in both models. Model A showed multicollinearity, and Models A and B heteroscedasticity. As a linear model can be considered robust (Cohen, 2013), it was nevertheless used in this study. The regression tables of Models A and B are provided in Table 3.

TABLE 3

Table 3. Regression analyses of pro-environmental intention and past pro-environmental behavior on justice sensitivity and moral disengagement in high carbon behavior in Study 1.

Model A significantly predicted pro-environmental intention, F(5, 168) = 20.08, p < 0.001, with a large effect size (according to Cohen, 2013), R² = 0.37 (adjusted R² = 0.36). Observer sensitivity (positive) and moral disengagement (negative) were significant predictors. All effects were in the hypothesized direction and, therefore, Hypotheses 1a and 2a were partially supported.

Model B predicted past pro-environmental behavior significantly, F(5, 168) = 7.46, p < 0.001, with R² = 0.18 (adjusted R² = 0.16), which is indicative of a medium effect size (Cohen, 2013). The only significant predictor was moral disengagement (negative). Except for beneficiary sensitivity, the effects were at least in the hypothesized direction, although they were not significant. As a result, Hypothesis 2b received only partial support.

Additionally, as past behavior was shown to be a strong predictor of pro-environmental intention (e.g., Bamberg et al., 2003; Carrus et al., 2008; Manca et al., 2020), we exploratively added past pro-environmental behavior in a second step in the regression analysis on pro-environmental intention. The standardized regression weights are shown in Table 3. Together, the predictors of this model (justice sensitivities, moral disengagement, past pro-environmental behavior) significantly predicted pro-environmental intention, F(6, 167) = 66.78, p < 0.001, with a large effect size (Cohen, 2013), R² = 0.71 (adjusted R² = 0.70). Adding past behavior as a predictor explained an additional 33% of the variance of pro-environmental intention, incremental F(1, 167) = 188.35, p < 0.001. While, perpetrator, beneficiary, and victim sensitivity remained non-significant predictors, moral disengagement in high carbon behavior was still significant. Observer sensitivity no longer significantly predicted pro-environmental intention. Past behavior was not only a significant, but also the strongest predictor in this model, followed closely by moral disengagement. Moreover, the inclusion of past behavior also reduced the explanatory power of moral disengagement.

In line with Hypothesis 5, moral disengagement in high carbon behavior was positively correlated to victim sensitivity, r(172) = 0.21, p = 0.005. As both variables were not normally distributed, Spearman’s rank correlation was additionally computed. This supported the result obtained with the Pearson correlation, ρ = 0.12, p < 0.001.

Study 2

To replicate the findings of Study 1 (H1, 2, and 4) in a larger sample and to additionally test Hypotheses 3 and 5, we again measured the two dependent variables pro-environmental behavior and pro-environmental intention. However, to provide a potentially more accurate measure for environmentally friendly behavior, we used a carbon footprint calculator in Study 2. The independent variables again were the justice sensitivity perspectives and moral disengagement in high carbon behavior. We included moral emotions as predictors and guilt as a mediator.