Why Going Green Feels Good Venhoeven, Leonie A.; Bolderdijk, Jan Willem; Steg, Linda

University of Groningen

Why going green feels good Venhoeven, Leonie A.; Bolderdijk, Jan Willem; Steg, Linda

Published in: Journal of Environmental Psychology

DOI: 10.1016/j.jenvp.2020.101492

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 2020

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Venhoeven, L. A., Bolderdijk, J. W., & Steg, L. (2020). Why going green feels good. Journal of Environmental Psychology, 71, [101492]. https://doi.org/10.1016/j.jenvp.2020.101492

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 30-09-2021 Journal of Environmental Psychology 71 (2020) 101492

Contents lists available at ScienceDirect

Journal of Environmental Psychology

journal homepage: www.elsevier.com/locate/jep

Why going green feels good

Leonie A. Venhoeven a,c, Jan Willem Bolderdijk b,*, Linda Steg c a Dijksterhuis & Van Baaren, St. Anthoniusplaats 9, 6511, TR Nijmegen, Groningen, the Netherlands b Department of Marketing, Faculty of Business and Economics, University of Groningen, Nettelbosje 4, 9747AE, Groningen, the Netherlands c Department of Psychology, Faculty of Behavioral and Social Sciences, University of Groningen, Grote Kruisstraat 2/1, 9712TS, Groningen, the Netherlands

ARTICLE INFO ABSTRACT

Handling Editor: Sander van der Linden Recent research found that when asked explicitly, people associate (future) sustainable actions with positive instead of negative emotions. This empirical finding implies that policy makers could harness people’s intrinsic Keywords: motivation to promote sustainable actions. It is however not clear where this association between sustainable Meaning actions and positive emotions stems from. Why would people report that going green feels good, given that such Sustainable behavior actions often require more effort or inconvenience? We argue and show that the previously found relationship Positive emotions between sustainable actions and anticipated positive emotions is not merely a matter of social desirability, but rather a matter of meaning: acting sustainably is often perceived as a moral choice and thus as a meaningful course of action, which can elicit positive emotions. Specifically, we found that participants also associate sus tainable actions with positive instead of negative emotions when using an implicit association measure – one that reduces the likelihood of socially desirable responses (Study 1). Moreover, participants anticipated more intense emotions when those actions were meaningful to them: they felt more positive about sustainable actions, and more negative about unsustainable actions, when they chose to engage in those actions and when those actions are considered personally relevant to the individual (Study 2). Together, these studies explain why acting green feels good, and support the notion that sustainable actions are indeed expected (not merely reported) to be intrinsically motivating.

1. Why going green feels good communication. However, it is currently not completely clear where the association Environmentally-sustainable actions oftentimes involve a certain between sustainable actions and positive emotions stems from. Why degree of discomfort or inconvenience. Taking the bike instead of the car would people report that sustainable actions (e.g. recycling) would to work, for instance, saves emissions, but requires physical effort, and make them feel good, given that those actions often require some level of longer travelling time. Washing clothing at lower temperatures saves sacrifice or effort? energy, but may leave stains intact. Recycling requires more effort than One more cynical explanation is that people do not actually have a putting all garbage into the same bin. Nevertheless, people anticipate positive association with sustainable actions, but simply report that sustainable actions will elicit positive emotions and a ‘warm glow’: they sustainable actions would make them feel good when explicitly asked report that sustainable actions will make them feel good about them about it. There are several reasons why explicit measures (i.e. self- selves, suggesting that they find it intrinsically rewarding to act sus reports) could allow for such response bias (Thomas & Walker, 2015). tainably (Pfister& Bohm,¨ 2008; Smith, Haugvedt & Petty, 1994; Taufik, First, people may consider it socially desirable to self-report that sus Bolderdijk, & Steg, 2016; Van der Linden, 2018). Given anticipated tainable actions would make them feel good: people may want to appear emotions often guide our decisions (Loewenstein & Lerner, 2003; Mel to themselves (i.e. self-deception) and to others (i.e. impression man lers & McGraw, 2001; Pfister & Bohm,¨ 2008; Schwarz, 2000; Taufik agement) as being concerned about the environment (Ewert & et al., 2016; Zeelenberg, Nelissen, Breugelmans, & Pieters, 2008), this Galloway, 2009). So, when explicitly asked to report their feelings to association between sustainable actions and positive emotions offers a wards sustainable actions, people may present themselves as being more hopeful message to policy makers: they could harness consumers’ intrinsically motivated than they really are. Second, explicit measures intrinsic motivation to act sustainably in their messages and could act as ‘prompts’ (Hafner, Walker & Verplanken, 2017). By asking

* Corresponding author. E-mail address: [email protected] (J.W. Bolderdijk). https://doi.org/10.1016/j.jenvp.2020.101492 Received 30 January 2020; Received in revised form 10 August 2020; Accepted 17 August 2020 Available online 22 August 2020 0272-4944/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492 participants to explicitly report whether sustainable actions would make IAT measures people’s automatic emotional association with a concept them feel good, self-reports make the possibility of positive feelings via response times. If people also implicitly associate sustainable salient in participants, making them report more positive feelings than behavior with positive instead of negative emotions, it would be they would experience spontaneously. Thus, it is conceivable that peo cognitively easier for participants to pair sustainable actions with pos ple (unintentionally) inflate the claim that they would feel good about itive instead of with negative emotions. Thus, they would respond sustainable actions in self-reports – the method used in prior empirical quicker when sustainable actions (e.g. ‘energy saving’) are presented in work to examine people’s emotional association with sustainable ac conjunction with the word ‘positive’ instead of with the word ‘negative’. tions (Taufik et al., 2016; Van der Linden, 2018). In other words, the IAT assumes participants respond quicker when We however argue that the previously found association between the pairing task is set up to be consistent with their pre-existing, auto sustainable actions and positive emotions is unlikely just the result of matic associations. On the other hand, participants would respond response bias. We tested this empirically in Study 1, using a method that slower when the pairing task requires them to override any pre-existing minimizes the influenceof response bias. Instead, we argue people may associations. By using response times to detect the valence of people’s actually anticipate positive emotions, as sustainable actions can elicit automatic associations, the IAT makes it difficult for participants to tell meaning: acting sustainably is often perceived as a morally-good choice what researchers are actually interested in, or to exert control over their and thus as a meaningful course of action, which can elicit positive responses, thus reducing the influenceof socially desirable response bias emotions. Due to meaning, people may associate sustainable behaviors - (De Houwer, 2006). In environmental psychology, the IAT has been used even those that involve a certain degree of effort or sacrifice - with to test people’s implicit attitude towards nuclear energy (Siegrist, Keller, positive instead of merely negative emotions. We tested the underlying & Cousin, 2006), their connection with natural versus built environ role of meaning in shaping positive emotional associations in Study 2. ments (e.g. Bruni & Schultz, 2010), as well as their implicit environ The notion that acting sustainably can elicit positive emotions can be mental attitudes and values (Thomas & Walker, 2015). Building on this traced back to Aristotle, who proposed that well-being can be found in work, we used the IAT for a new question: do people also associate the expression of virtue or ‘doing the right thing for the right reasons’ sustainable actions with positive instead of negative emotions when such (Ryan, Huta, & Deci, 2008). This implies that having pleasant hedonic associations are measured implicitly? experiences is not the only route to positive emotions (Reis, Sheldon, In addition to examining the valence of people’s implicit associations Gable, Roscoe, & Ryan, 2000; Ryan & Deci, 2000; Sheldon, Elliot, Kim, with sustainable actions (Study 1), we also tested impact of meaning & Kasser, 2001); having meaningful experiences can feel good as well more directly in Study 2. Because protecting the environment is typi (Petersen, Park & Seligman, 2005). Indeed, morally-good behaviors cally seen as a morally-relevant course of action, people may come to (such as caring for others) are generally seen as important and mean associate environmentally-sustainable actions with positive emotions (e. ingful behaviors that can elicit positive emotions (Aknin, Dunn, & g. pride). Vice versa, they may associate unsustainable actions with Norton, 2012; Andreoni, 1989, 1990; Batson & Powell, 2003; Dunn negative emotions (e.g. guilt). Importantly, our reasoning implies that Aknin & Norton, 2008; Grant & Sonnentag, 2010). the emotions people associate with both sustainable and unsustainable Sustainable actions are prototypical of ‘morally-good’ actions. actions would become more intense when they perceive those actions as Indeed, many people relate sustainable actions to the prevention of being more meaningful (H2). harm, and care for others (Jia, Soucie, Alisat, Curtin, & Pratt, 2017), and We tested this hypothesis in two ways. We firstreasoned that people thus perceive it as a manifestation of moral behavior (Feinberg & Willer, are likely to feel that their actions say more about them (i.e. are more 2012; Howell, 2013; Pandey, Rupp, & Thornton, 2013; Thøgersen, meaningful) when they choose to engage in those actions. After all, one 1996). Americans, for example, strongly agreed that nature has intrinsic can hardly claim to have done the ‘good’ or ‘bad’ thing if one is not value and that humans have moral duties and obligations to animals, responsible for having initiated that action (Heberlein, 1972; Venho plants, and non-living nature such as rocks, water, and air (Leiserowitz, even, Bolderdijk & Steg, 2016; Zeelenberg, van der Pligt, & de Vries, Kates, & Parris, 2005). Likewise, many people in the UK agreed that 2000). Thus, the emotions people associate with both sustainable and people have personal, social and moral responsibilities to address unsustainable actions would become more intense when those actions climate change (Lorenzoni, Nicholson-Cole, & Whitmarsh, 2007). In as are perceived as more volitional: instigated by them personally (H2a). far as people consider acting sustainably to be the morally-right, and Specifically, our reasoning implies we would find interactions between thus a meaningful thing to do, they may come to associate actions that the level of volition and the sustainability of the behavior: people would promote sustainability with positive, instead of just negative emotions. report feeling more positive about sustainable actions when they voli So how can we test this account, given that asking people explicitly tionally chose to engage in those ‘good’ behaviors. Vice versa, people would about their associations with sustainable actions may bias response report feeling more negative about unsustainable actions when they patterns? We employed a two-step strategy. First, we examined whether chose to engage in those ‘bad’ behaviors (as compared to having no the association between sustainable actions and positive emotions still choice). We tested these interactions in Study 2, by systematically exists when measuring associations implicitly, thus reducing the influ varying whether the action was positioned as being sustainable or not (i. ence of response bias. Second, we examined the underlying role of e., conducive vs. harmful to environmental quality), and varying meaning in shaping people’s emotional associations with (un-)sustain whether or not participants chose to engage in those actions volitionally. able actions. Second, we reasoned that how positive people expected to feel about If the previously found association with positive emotions indeed sustainable actions (and how negative they would feel about unsus stems from the personal meaning people attach to sustainable behaviors, tainable actions) would be magnified for those individuals for whom and is not the result of a (socially desirable) response bias, the associa acting sustainably is a particularly meaningful course of actions (H2b). tion between sustainable actions and positive emotions would also be Specifically,acting in an (un-)sustainable fashion is likely to mean more found when associations are measured implicitly, without participants to individuals who perceive environmental quality as an important being aware that their associations are measured (H1). Therefore, in value (Steg, Perlaviciute, Van der Werff, & Lurvink, 2014), and more to contrast to questionnaires that directly ask for people’s explicit associ those who feel morally obliged to protect the environment. On the basis ations with sustainable actions (e.g. Taufiket al., 2016; Van der Linden, of this, we expected to findanother interaction: individuals with strong 2018), in Study 1, we tested the valence of people’s implicit associations biospheric values and a strong personal norm to act sustainably would with various sustainable actions. Specifically, we used an implicit as feel most positive about sustainable actions, and most negative about sociation test (the “IAT”, Greenwald, McGhee, & Schwartz, 1998; unsustainable actions. Greenwald, Nosek, & Banaji, 2003) to examine the valence of partici pants’ overall implicit association with various sustainable actions. The

2 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

1.1. Current research Next, participants took part in the main part of this study - a computerized Implicit Association Test (IAT; Greenwald et al., 1998). In two studies, we empirically examined the feelings that people Various words were shown on a computer screen (See Appendix). Par associate with various environmentally-relevant actions. In Study 1, we ticipants had to assign these words as quickly as possible to either of two zoomed in on sustainable actions, and examined whether people also categories that appeared left or right in the top of the screen. Specif associate sustainable actions with positive instead of negative emotions ically, using the keyboard keys “Z” (assign word to the category dis when these associations are measured implicitly – via an implicit asso played left) and “M” (assign word to the category displayed right), ciation test (H1). In Study 2, in addition to sustainable actions, we also participants had to assign sustainable actions such as ‘recycling’ to the examined participants’ associations with unsustainable actions. We category ‘sustainable’ instead of ‘neutral’2, and assign neutral words tested whether the emotions people explicitly associate with sustainable such as ‘mirror’ to the category ‘neutral’ instead of ‘sustainable’. and unsustainable actions become more intense when those actions are Moreover, they had to assign negative emotions such as ‘unhappy’ to the perceived as more meaningful: when they are volitional (H2a; Study 2) category ‘negative’ instead of ‘positive’ and assign positive emotions and considered personally relevant to the individual (H2b; Study 1 and such as ‘proud’ to the category ‘positive’ instead of ‘negative’. We 2). In other words, we isolated the underlying role of meaning by strived for the negative emotions to mirror the positives (e.g. proud vs. examining interactions between the type (sustainable vs. unsustainable) guilt, happy vs. unhappy). We logged response times. and meaningfulness (volitional or not, personally relevant or not) of the Importantly, across blocks, we systematically varied whether or not behavior. Thus, rather than relying on mediation to test a specific pro the category ‘sustainable’ was presented in the conjunction with words cess (see Fiedler, Schott & Meiser, 2011), we used interactions to isolate from the category ‘positive’ (congruent block) versus ‘negative’ (dis the underlying role of meaning (i.e. testing a process by experimentally congruent block). If people indeed also implicitly associate sustainable interrupting vs. amplifying the hypothesized process; see Spencer, actions with positive instead of negative emotions, assigning words such Zanna, & Fong, 2005; Jacoby & Sassenberg, 2011). as ‘recycling’ to the corresponding category ‘sustainable’ would be cognitively easier, and thus quicker, when the category ‘sustainable’ is 2. Study 1 presented in combination with the category ‘positive’ (‘sustainable/ positive’; congruent block) instead of with the category ‘negative’ We expected participants would still associate sustainable actions (‘sustainable/negative’; discongruent block). Similarly, our theory im with positive instead of negative emotions when associations are plies that assigning positive emotions such as ‘proud’ to the category measured implicitly, without participants being aware that their associ ‘positive’ would be easier when the category ‘positive’ is presented in ations are measured (H1). We tested the valence of participants’ auto conjunction with the category ‘sustainable’ (congruent block) instead of matic associations with various sustainable actions, using an implicit ‘neutral’ (discongruent block). association test (Greenwald et al., 1998). Informed consent was ob We examined the valence of participants’ implicit association with tained from all participants. The study was assessed by an Ethical sustainable actions by testing whether response times would be lower Committee and was approved. during congruent blocks (see Appendix, Blocks 3 and 4) than during discongruent blocks (Blocks 6 and 7). In coding response times, we 2.1. Participants followed standard IAT procedures (Greenwald et al., 2003). To minimize the influence of outliers and error, we excluded trials where reaction times were slower than 10.000 ms (0 cases) and faster than 400 ms (38 First year psychology students (N = 76; 67 female, 9 male; Medianage = 19.0 years) took part in a laboratory experiment in exchange for cases). Incorrect responses (e.g. participant incorrectly assigning the ’ ’ course credits. Power analysis indicated that a sample size of 34 subjects word ‘mirror to the category ‘sustainable ) were penalized and ’ was required for the study to have 80% power to detect a medium-sized substituted by participants mean response time for correct responses in + effect (f = 0.25) at an α level of 0.05 for a repeated measure within that block 2 SD (238 cases). We used both the practice and test blocks participants with two levels (congruent vs. discongruent blocks, see to calculate response times for congruent vs. discongruent blocks. See below), using the default value of 0.50 for the correlation among Appendix for a more elaborate description of the order of the IAT blocks repeated measures. We stopped collecting data after two weeks in the and the exact stimuli used. lab, having reached more than double this number of subjects. 2.3. Results and discussion 2.2. Materials During congruent blocks (when the category ‘sustainable’ was pre ’ Participants firstcompleted a value questionnaire (Steg et al., 2014), sented in conjunction with the category ‘positive ), participants = which included four items gauging individual differences in the extent to responded faster (M 795.24 ms) than during discongruent blocks ’ which people consider caring about nature and the environment (i.e., (when the category ‘sustainable was presented in conjunction with the ’ = = < η 2 = respecting the earth, unity with nature, protecting the environment and category ‘negative ; M 902.82 ms); F(1,75) 11.86, p .01, p = = preventing pollution) to be important. Respondents indicated how 0.14; IAT effect 107.58 ms; IAT D 0.24. Thus, it appears to be important these values are as guiding principles in their life; scores on cognitively easier for participants to pair sustainable words with posi value items could range from 1 = opposed to my principles, 0 = not tive emotions than with negative emotions. This implies that, also when measured implicitly, sustainable actions are more strongly associated important, to 7 = extremely important; Mbiopheric = 3.60, SDbiospheric = 1.46)1. with positive than with negative emotions (H1). Additionally, we tested whether the implicit associations with sus tainable behavior would be more positive for people with stronger – 1 The questionnaire also included a measure of individual differences in biospheric values for those who are likely to perceive protecting the regulatory focus (Van Stekelenburg, 2006) and two questions on meat con environment as being a particularly meaningful course of action (H2b). sumption (“How many days a week on average do you eat meat with your main dish?” and “How many grams meat do you on average eat with your main dish, when it contains meat?”). Results on these last two constructs are not reported 2 We contrasted the category ‘sustainable’ with the category ‘neutral’, here, as they are not relevant for the goal of the current paper. The value because the goal of current study was to test whether people associate sus questionnaire was asked in between the regulatory focus measure and the two tainability words with positive instead of negative emotions, not to test their questions on meat consumption. association with unsustainable words (as in Study 2).

3 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

The difference in response times between congruent and discongruent washing clothes at a low temperature when it is the only available option blocks was more pronounced for people with stronger biospheric values, on the machine, lights are automatically turned off when leaving an empty but this effect was modest; r(74) = 0.19, p = .10). This could be due to room, buying organic products in the cafeteria when they are the only the specific sample used - we used a relatively homogenous, liberal available option, separating waste at work because your employer requires sample (psychology students). it). Half of the participants evaluated sustainable actions (e.g. using the Summing up, it seems that most participants, regardless of the extent bike for short distances, turning lights off in an unused room), while the to which they endorsed biospheric values, tended to also implicitly other half evaluated the unsustainable counterparts of the same be associate sustainability with positive instead of negative emotions. Since haviors (e.g., using the car for short distances, leaving the light on in an we measured automatic associations between sustainable actions and empty room). Hence, all participants evaluated ten behaviors in total, positive emotions, and participants were unaware of the purpose of this which varied in terms of whether or not they were volitional (within study, it is unlikely that these results are strongly affected by social subjects), and whether they were sustainable versus unsustainable (be desirability (participants reporting they would feel good about sus tween subjects). tainable actions out of self-representation concerns; De Houwer, 2006) – or by response bias (explicit self-reports prompt participants to report 3.3. Measures emotions they would not consider spontaneously; Thomas & Walker, 2015). Participants reported to what extent they expected that engaging in While demonstrating that the previously found association between each action would make them feel ‘proud’, ‘satisfied’ and ‘cheerful’ ′ sustainable actions and positive emotions is unlikely to be the result of a (averaged to represent positive emotions; α s for the ten behaviors response bias, this study does not answer the question: so why do people ranged from 0.87 to 0.93; αoverall = 0.97) and to what extent they ex associate sustainable actions with positive emotions when asked pected to feel ‘disappointed’, ‘frustrated’ and ‘guilty’ (averaged to ′ explicitly? Is it due to the fact that such actions provide meaning? Study represent negative emotions; α s for the ten behaviors ranged from 0.87 3 2 was conducted to study test this. to 0.94; αoverall = 0.97) . Scores could range from 1 ‘not at all’ to 5 ‘very strongly’, with 3 reflecting ‘neutral’. The emotions chosen here are 3. Study 2 consistent with the emotions typically measured to gauge people’s as sociation with sustainable actions (see Krettenauer, Wang, Jia, & Yao, We conducted a questionnaire study in which we experimentally 2019; Rees, Klug, & Bamberg, 2015; Rezvani, Jansson, & Bengtsson, manipulated the type of action that participants envisioned (sustainable 2017). We decided to use a slightly different set of emotions relative to vs. unsustainable, and voluntary or not, respectively), and asked par the ones used in Study 1, as to ensure that the results are not limited to a ticipants to report how they anticipated to feel after engaging in those specific set of emotions positive and negative emotions. actions. We now used a broader, non-student sample, and now asked for Next, participants completed the same values questionnaire as in people’s explicit associations. We hypothesized participants would Study 1 to gauge individual differences in biospheric values (M = 4.18, report anticipating more intense emotions when the same actions were SD = 1.47, α = 0.874). Additionally, participants indicated the extent to more meaningful (H2): when participants chose to engage in those ac which they felt morally obliged to engage in each of the fivesustainable tions (H2a), and when those actions were considered meaningful to the behaviors (1 = not at all to 5 = very strongly; M = 3.49, SD = 0.90, α = specific participant (H2b; so among individuals with strong biospheric 0.81). The scores on biospheric values and feelings of moral obligation values and a personal norm to act sustainably). Thus, we tested the role to act pro-environmentally were mean centered prior to the analysis5. of meaning via examining interactions (Jacoby & Sassenberg, 2011; Spencer et al., 2005). Informed consent was obtained from all partici 3.4. Results and discussion pants. The study was assessed by an Ethical Committee and was approved. We firstexamined whether the emotions (both negative and positive) participants associated with their sustainable and unsustainable actions, 3.1. Participants respectively, become more intense when those actions are perceived as volitional (H2a). = = Participants (N 132; 60 female, 65 male, 7 unknown; Medianage Fig. 1 shows the results for positive emotions. We ran a mixed 30.5 years) completed a questionnaire that was distributed door-to-door analysis of variance (with volitional as the within-subject factor, and in diverse neighborhoods in a city in the Netherlands, and recollected sustainability as the between-subject factor). The results show partici after approximately half an hour. Power analysis indicated that a sample pants who imagined engaging in the sustainable behaviors (e.g. using size of 128 subjects was required for the study to have 80% power to the bike for short distances) anticipated feeling more positive (Mpos = detect a medium-sized effect (f = 0.25) at an α level of 0.05 with two 2.99, SDpos = .80, so around the midpoint of the scale) than participants groups; thus, we stopped collecting data when we reached a bit over this number of subjects. The data of 9 participants were not included in the analyses because they had missing values on at least one of the depen 3 For each participant, we calculated aggregate scores for both negative and dent or independent variables. positive emotions across both voluntary and involuntary behaviors. 4 For half of the participants the value scale was included as the firstmeasure 3.2. Materials in the questionnaire, and for the other half the value scale was included as the last measure in the questionnaire. The order in which the questions were asked We asked participants to report what emotions they expected to did not affect any of our results. 5 experience after engaging in different types of (un-)sustainable actions. Participants also indicated to what extent they thought each of the five These behaviors are widely known and prototypical examples of sus sustainable behaviors (so using the bike for short distances, turning the light off in an empty room, buying organic products in the cafeteria, separating waste at tainable actions to our (Dutch) participants. Participants were first work, washing clothes at a low temperature) had a positive impact on envi prompted to imagine five behaviors that they engaged in out of their ronmental quality (1 = not at all to 5 = very strongly; M = 4.23, SD = 0.65, α = own volition (i.e., using the bike for short distances, turning the light off 0.78). Moreover, participants indicated to what extent they intended engaging in an empty room, buying organic products in the cafeteria, separating in the sustainable version of each of the fivevoluntary behaviors (1 = not at all waste at work, washing clothes at a low temperature) and next imagined to 5 = very strongly; M = 3.96, SD = 0.74, α = 0.76) before they rated the five equivalent behaviors they engaged in out of situational constraints environmentally-friendliness and moral obligation towards these behaviors. (i.e., using the bike for short distances while your car is being serviced, These results are beyond the scope of this paper but are available upon request.

4 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

Fig. 1. Anticipated positive emotions after engagement in volitional and non- Fig. 2. Anticipated negative emotions after engagement in volitional and non- volitional (un-)sustainable behavior. Error bars denote standard errors. volitional (un-)sustainable behavior. Error bars denote standard errors. who imagined engaging in the unsustainable counterparts of those be more intense among participants who consider such actions personally 7 haviors (e.g. using the car for short distances; Mpos = 1.85, SDpos = .68; relevant (H2b). Regression analyses showed that participants expected 2 6 Fpos(1,121) = 72.29, ppos < .001, ηp pos = .37 ). Importantly, we found feeling relatively more positive about sustainable (and least positive 2 the predicted interaction; F(1,121) = 31.92, p < .001, ηp = 0.21. about unsustainable) actions when they more strongly endorsed Specifically, participants who considered the sustainable actions antic biospheric values (Bbehaviortype x values = 0.34, t(119) = 3.71, p < .001, see ipated feeling even more positive when they engaged in those behaviors Fig. 3 for a visual illustration) and when they felt more morally obliged = out of their own volition (e.g. using the bike for short distances, M to engage in those sustainable behaviors (Bbehaviortype x obligation = 0.44, t 3.19, SD = 0.90) than out of situational constraints (e.g. using the bike (119) = 3.00, p < .01, see Fig. 3). Thus, we findthe expected interaction: for short distances when your car is being serviced, M = 2.79, SD = 0.81, sustainable actions are anticipated to elicit more positive emotions Mdifference = 0.40, t(60) = 5.40, p < .001). Thus, people associate the among participants for whom such actions are particularly meaningful. same sustainable actions more strongly with positive emotions when Vice versa, participants who strongly endorsed biospheric values such actions are more meaningful: the result of their own choosing (e.g. (Bbehaviortype x values = 0.12, t(119) = 1.50, p = .14, see Fig. 4 for an ‘using the bike for short distances’ as opposed to ‘using the bike for short illustration) and who felt more morally obliged to act sustainably (Bbe distances when your car is being serviced’). haviortype x obligations = 0.37, t(119) = 2.90, p < .01, see Fig. 4) reported As can be seen in Fig. 2, the overall pattern of results for negative feeling relatively more negative about unsustainable actions (and least emotions seems the exact opposite of the pattern for positive emotions. negative about sustainable actions). In other words, we again find sup Participants who imagined engaging in the unsustainable actions port for the predicted interactions (H2b): the negative emotions people anticipated feeling more negative (Mneg = 2.50, SDneg = 0.81, slightly associate with unsustainable actions became more intense among in below the midpoint of the scale) than those who anticipated engaging in dividuals for whom such actions are particularly meaningful. the sustainable actions (Mneg = 1.48, SDneg = 0.45); Fneg(1,121) = 73.61, In sum, consistent with our reasoning, we found that participants’ 2 pneg < .001, ηp neg = 0.38. Importantly, we again findthe hypothesized self-reported emotions (both positive and negative) become more 2 interaction (F(1, 121) = 26.71, p < .001, ηp = 0.18): participants who intense when the actions are perceived as more meaningful: when the engaged in the unsustainable actions felt more negative when they chose (un-)sustainable action is a consequence of one’s own volition (H2a), (e.g. not separating waste at work; M = 2.58, SD = 0.87), rather than and when those actions are considered personally more meaningful to were forced (e.g. not separating waste at work due to a lack of containers; the individual (H2b). M = 2.42, SD = 0.83) to engage in those ‘bad’ behaviors (Mdifference = 0.16, t(60) = 2.26, p = .03). In other words, people anticipate feeling 4. General discussion relatively less bad about the same unsustainable actions when such ac tions are not particularly meaningful: not the result of their own Previous research found that people associate sustainable actions choosing. with positive emotions instead of negative emotions (e.g. Taufik et al., In sum, consistent with H2a, we find the expected interactions. The 2016; Van der Linden, 2018). But why would people report that sus emotions (both positive and negative) people associate with their ac tainable actions would make them feel good, given that such actions tions become more intense when those actions are more meaningful, i.e. often involve some degree or discomfort or effort? Is this positive as when participants imagine they choose to engage in those sustainable sociation perhaps an artefact produced by the explicit nature of the and unsustainable actions, respectively. self-reports used in previous research? We proposed people actually Next, we tested whether the emotions (both negative and positive) anticipate feeling good about sustainable actions, because such actions participants associate with sustainable and unsustainable actions are provide meaning. Consistent with this reasoning, we found that people also associate sustainable actions with positive instead of negative emotions when associations are measured implicitly, instead of

6 When we looked at the emotions separately, the same pattern of results was found for all emotions. 7 For each participant, we calculated one score, so averaging emotions across the voluntary and involuntary behaviors. We first regressed positive emotions onto type of behavior (sustainable vs. unsustainable), individual differences in biospheric values (mean-centered), and their interaction term. Next, we repeated this, now including feelings of moral obligation as the moderator. Finally, we repeated both analyses for the negative emotions. Thus, in total, we conducted four regression analyses. No control variables were included.

5 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

Fig. 3. Individual differences in biospheric values (plotted at 1 SD below, at, and 1SD above the mean, respectively) and perceived moral obligation determine how positive people feel about sustainable and unsustainable actions. Error bars 95% confidence intervals.

6 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

Fig. 4. Individual differences in biospheric values (plotted at 1 SD below, at, and 1SD above the mean, respectively) and perceived moral obligation determine how negative people feel about sustainable and unsustainable actions. Error bars 95% confidence intervals.

7 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492 explicitly (Study 1), thus reducing the likelihood of (socially desirable) Yet, it has been argued that people can ‘fake’ IAT tests when instructed response bias affecting the results (supporting Hypothesis 1). Study 2 (Fiedler & Bluemke, 2005). However, participants need to know how to examined why people would explicitly report that sustainable actions do this – they need to be able to tell apart congruent from discongruent would make them good (and why unsustainable actions would make blocks, and intentionally slow down their responses in congruent blocks them feel bad). We find that the emotions people explicitly associate (Fiedler & Bluemke, 2005). Therefore, we did not inform participants with both sustainable and unsustainable actions become more intense about the goal of the study (i.e. testing people’s implicit association with when participants see those actions as being more meaningful (Study 2, sustainable vs. neutral words using response times), making it unlikely H2). Specifically, as expected (Hypothesis 2a), participants felt more that participants ‘caught on’ and intentionally slowed down their re positive about sustainable actions (and more negative about unsus sponses during congruent blocks. Yet, future research could verify this tainable actions) when they chose to engage in these actions. Moreover, and control for any participants attempted to ‘rig’ the results. these reported emotions were more intense among participants for Study 2 examined the underlying role of meaning. In addition to whom acting sustainability is more meaningful: those who care more experimentally varying whether the actions were volitional and sus about environmental protection (i.e., those strongly endorsing tainable (H2a), we explored the moderating role of individual differ biospheric values) and those who feel more morally obliged to act ences (H2b). Here, we findthat the emotions associated with sustainable pro-environmentally (supporting Hypothesis 2b). actions are more intense for participants with stronger biospheric Together, these results support our theorizing that the previously values, and those feeling more morally obliged to act pro- found relationship between sustainable actions and positive emotions is environmentally. While these findings support the role of meaning and merely a matter of social desirability, but rather a matter of meaning: thus are consistent with our experimental results, future research could acting sustainably is perceived as a moral choice and thus as a meaningful an activate rather than measure individual differences in feelings of course of action, which can elicit positive emotions. moral obligation and values, and aim to replicate these findings experimentally. 4.1. Practical implications Our results could be affected by the particular selection of emotions, samples and type of sustainable actions. We used different emotions and It is often assumed that people associate sustainable actions mainly samples across Studies 1 and 2, but nevertheless find similar results: with negative emotions and thus would decrease well-being, as acting participants seem to associate sustainable actions with positive (instead sustainably oftentimes requires a certain degree of effort or inconve of just negative) emotions. These findings are in line with prior work nience. The previous finding that people also associate sustainable ac (Taufiket al., 2016; Van der Linden, 2018), and suggest that the positive tions with positive emotions paints a more optimistic picture, and forms associations we found is unlikely an artefact of the particular samples the basis for an alternative view on environmental policy: rather than and words chosen for our studies. However, future research should exclusively using extrinsic incentives to motivate sustainable actions, examine whether similar results are found when including other types of policies and communication campaigns could also be geared toward emotions, more diverse, non-Western samples. For Dutch participants facilitating individuals to act on their pre-existing intrinsic motivation to (our sample) for instance, using the bike to cover short distances is quite act sustainably (Steg, 2016; Van der Linden, 2018). Our work solidifies common. The same action may elicit more meaning and thus more the empirical basis for this conclusion: the association between sus positive emotions other countries, where cycling is less normalized. tainable actions and positive emotions seems not just the result of par Moreover, we tested our reasoning using prototypical sustainable ac ticipants showing (socially desirable) response bias. Instead, the tions (e.g. taking the bike for shorter distances, recycling), but one could association between sustainable actions and positive emotions appears argue that these behaviors only involved minor discomfort or effort. to be rooted in a fundamental psychological need: humans striving for Future research should examine whether people still associate sustain meaning. able actions with positive emotions when the behavioral costs (and Finding ways to increase the perceived meaning of sustainable ac therefore also the meaning) of the sustainable behavior increases (e.g. tions could thus be a fruitful way to further increase the positive not flying anymore). emotional association people have with acting sustainably, and thus boost people’s intrinsic motivation to act sustainably. Stressing the moral nature of sustainable behavior in general, for instance, may make 4.3. Conclusion any sustainable behavior feel more meaningful. Campaigns could for instance consider highlighting the societal, collective advantages of Using a combination of methods and materials, our studies suggest sustainable actions (e.g. reducing climate change), rather than solely that the previously found relationship between sustainable actions and stressing the individual benefits (e.g. saving money by conserving en anticipated positive instead of negative emotions is not merely a matter ergy). Moreover, our results suggest that stressing and facilitating the of social desirability, but rather a matter of meaning: acting sustainably voluntary aspect impact of a particular sustainable behavior may in is often perceived as a moral choice and thus as a meaningful course of crease the meaning of that behavior specifically. For instance, rather action, which can elicit positive emotions, particularly when people than legally requiring people to recycle, policy makers could consider choose to engage in those actions, and feel that acting sustainably is an praising individuals who go beyond the call of duty, and recycle out of important personal goal. Vice versa, acting unsustainably, despite that it their own volition. Moreover, strengthening or activating biospheric is often more convenient, can elicit negative emotions, particularly values and feelings of moral obligation to act pro-environmentally may when such actions are more meaningful: when people choose to engage enhance the extent to which pro-environmental actions are perceived to in those actions, and when people feel that acting sustainably is an be meaningful, and thus intrinsically rewarding (cf. Steg, 2016). Future important personal goal. In conclusion, our results support one expla research could test different means to increase the perceived meaning of nation as to why people report that going green would make them feel sustainable behavior, and test whether this indeed increases the positive good: these actions offer meaning. emotions people associate with sustainability. Acknowledgements 4.2. Limitations and future research This project is part of the projects CReating Innovative Sustainability We used the IAT in Study 1 to reduce the likelihood that people’s Pathways (CRISP). We thank R. De Masi for her help with collecting data association with sustainability would be biased by social desirability. for Study 2.

8 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

Appendix

IAT Procedure in detail

Fig. 1. Example choices from the IAT.

Participants completed seven blocks, each consisting of 20 Trials, in which they had to place the target words in the corresponding category, as illustrated above. Words could be placed in the category that appeared on the left by pressing the “Z” key on the keyboard, or be placed in the category on the right by pressing the “M” key on the keyboard. The categories used in the different blocks were the following:

• Block 1: Positive—Negative • Block 2: Sustainable—Neutral • Block 3: Sustainable/Positive—Neutral/Negative (Practice congruent block) • Block 4: Sustainable/Positive—Neutral/Negative (Test congruent block) • Block 5: Neutral—Sustainable • Block 6: Neutral/Positive—Sustainable/Negative (Practice congruent block) • Block 7: Neutral/Positive—Sustainable/Negative (Test congruent block)

The words were presented in randomized order within each of the seven blocks and the order of the blocks was counterbalanced between par ticipants: all participants started with Block 1, after which half of the participants were firstpresented with blocks Block 2, 3 and 4 while the other half of the participants were first presented with blocks Block 5, 6 and 7. The exact words chosen for the sustainable and neutral category were chosen based on a pilot study: they reflected prototypical sustainable concepts that require a certain degree of inconvenience or cost on behalf of those who execute those actions – i.e. “solar power”, “energy saving”, “environmentally-conscious”, “reuse” and “recycling”. The neutral words – i.e. “mirror”, “striped”, “alphabetical”, “elastic”, “figurative” were random concepts that, according to the results of the pilot study, were not related to sustainability, and did not evoke any emotional response (See Table 1).

9 L.A. Venhoeven et al. Journal of Environmental Psychology 71 (2020) 101492

Categories and Words Used in the IAT

Table 1 The IAT was conducted in Dutch. English translations of the used words are included in brackets. The sustainable words, on average, counted more (4.2) syllables than the neutral words (2.8). The positive emotions counted 2.0 syllables on average, the negative 2.6.

Positive Negative Sustainable Neutral

Gelukkig (Happy) Ongelukkig (Unhappy) Zonne-energie (Solar power) Spiegel (Mirror) Trots (Proud) Schuldig (Guilty) Energiebesparing (Energy saving) Gestreept (Striped) Plezier (Fun) Naar (Miserable) Milieubewust (Env. Conscious) Alfabetisch (Alphabetical) Voldaan (Fulfilled) Ontevreden (Dissatisfied) Hergebruik (Reuse) Elastisch (Elastic) Prettig (Contented) Schaamte (Shame) Recycling (Recycling) Figuurlijk (Figurative)

Appendix D. Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.jenvp.2020.101492.

Author statement

This project was funded by the European Commission (FP7), Grant Agreement 265310 (CRISP). Leonie Venhoeven collected and analyzed the data for Studies 1 and 2, under the supervision of Jan Willem Bolderdijk and Linda Steg. LV, JWB and LS conceptualized the work and contributed to the writing and editing of the manuscript.

References A. H. Huffman, & S. R. Klein (Eds.), Green organizations: Driving change with IO psychology (pp. 69–92). New York: Routledge. Pfister, H., & Bohm,¨ G. (2008). The multiplicity of emotions: A framework of emotional Aknin, L. B., Dunn, E. W., & Norton, M. I. (2012). Happiness runs in a circular motion: functions in decision making. Judgment and Decision Making, 3, 5–17. Evidence for a positive feedback loop between prosocial spending and happiness. Rees, J. H., Klug, S., & Bamberg, S. (2015). Guilty conscience: Motivating pro- Journal of Happiness Studies, 13, 347–355. environmental behavior by inducing negative moral emotions. Climatic Change, 130 Andreoni, J. (1989). Giving with impure altruism: Applications to charity and ricardian (3), 439–452. equivalence. Journal of Political Economy, 97, 1447–1458. Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoe, J., & Ryan, R. M. (2000). Daily well- Andreoni, J. (1990). Impure altruism and donations to public goods: A theory of warm- being: The role of autonomy, competence, and relatedness. Personality and Social glow giving. The Economic Journal, 100, 464–477. Psychology Bulletin, 26, 419–435. Batson, C. D., & Powell, A. A. (2003). Handbook of psychology. John Wiley & Sons, Inc. Rezvani, Z., Jansson, J., & Bengtsson, M. (2017). Cause I’ll feel good! an investigation Dunn, E. W., Aknin, L. B., & Norton, M. I. (2008). Spending money on others promotes into the effects of anticipated emotions and personal moral norms on consumer pro- happiness. Science, 319, 1687–1688. environmental behavior. Journal of Promotion Management, 23(1), 163–183. Ewert, A., & Galloway, G. (2009). Socially desirable responding in an environmental Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions context: Development of a domain specific scale. Environmental Education Research, and new directions. Contemporary Educational Psychology, 25, 54–67. 15(1), 55–70. Ryan, R. M., Huta, V., & Deci, E. L. (2008). Living well: a self-determination theory Feinberg, M., & Willer, R. (2012). The moral roots of environmental attitudes. perspective on eudaimonia. Journal of Happiness Studies, 9, 139–170. Psychological Science, 24, 1–7. Schwarz, N. (2000). Emotion, cognition, and decision making. Cognition & Emotion, 14, Grant, A. M., & Sonnentag, S. (2010). Doing good buffers against feeling bad: Prosocial 433–440. impact compensates for negative task and self-evaluations. Organizational Behavior Sheldon, K. M., Elliot, A. J., Kim, Y., & Kasser, T. (2001). What is satisfying about and Human Decision Processes, 111, 13–22. satisfying events? Testing 10 candidate psychological needs. Journal of Personality Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual and Social Psychology, 80, 325–339. differences in implicit cognition: The implicit association test. Journal of Personality Siegrist, M., Keller, C., & Cousin, M. E. (2006). Implicit attitudes toward nuclear power and Social Psychology, 74, 1464–1480. and mobile phone base stations: Support for the affect heuristic. Risk Analysis, 26(4), Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the 1021–1029. implicit association test: I. An improved scoring algorithm. Journal of Personality and Smith, S. M., Haugtvedt, C. P., & Petty, R. E. (1994). Attitudes and recycling: Does the Social Psychology, 85, 197–216. measurement of affect enhance behavioral prediction? Psychology and Marketing, 11, Howell, R. A. (2013). It’s not (just) “the environment, stupid!”. Values, motivations, and 359–374. routes to engagement of people adopting lower-carbon lifestyles. Global Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why Environmental Change, 23, 281–290. experiments are often more effective than mediational analyses in examining Jacoby, J., & Sassenberg, K. (2011). Interactions do not only tell us when, but can also psychological processes. Journal of Personality and Social Psychology, 89(6), 845. tell us how: Testing process hypotheses by interaction. European Journal of Social Steg, L. (2016). Values, norms and intrinsic motivation to act proenvironmentally. Psychology, 41(2), 180–190. Annual Review of Environment and Resources, 41, 277–292. https://doi.org/10.1146/ Jia, F., Soucie, K., Alisat, A., Curtin, D., & Pratt, M. (2017). Are environmental issues annurev-environ-110615-085947. moral issues? Moral identity in relation to protecting the natural world. Journal of Steg, L., Perlaviciute, G., Van der Werff, E., & Lurvink, J. (2014). The significance of Environmental Psychology, 52, 104–113. hedonic values for environmentally relevant attitudes, preferences, and actions. Krettenauer, T., Wang, W., Jia, F., & Yao, Y. (2019). Connectedness with nature and the Environment and Behavior, 46, 163–192. decline of pro-environmental behavior in adolescence: A comparison of Canada and Taufik, D., Bolderdijk, J. W., & Steg, L. (2016). Going green? The relative importance of China. Journal of Environmental Psychology, 101348. feelings over calculation in driving environmental intent in The Netherlands and the Leiserowitz, A. A., Kates, R. W., & Parris, T. M. (2005). Do global attitudes and behaviors United States. Energy Research & Social Science, 22, 52–62. support sustainable development? Environment: Science and Policy for Sustainable Thøgersen, J. (1996). Recycling and morality A critical review of the literature. Development, 47, 22–38 (2005). Environment and Behavior, 28, 536–558. Loewenstein, G., & Lerner, J. S. (2003). The role of affect in decision making. In Van Stekelenburg, J. (2006). Promoting or preventing social change. Instrumentality, R. Davidson, K. Scherer, & H. Goldsmith (Eds.), Handbook of affective science (pp. identity, ideology and groups-based anger as motives of protest participation, VU 619–642). New York: Oxford University Press. University Amsterdam. Doctoral Dissertation. Amsterdam: Ridderprint. Lorenzoni, I., Nicholson-Cole, S., & Whitmarsh, L. (2007). Barriers perceived to engaging Van der Linden, S. (2018). Warm glow is associated with low- but not high-cost with climate change among the UK public and their policy implications. Global sustainable behaviour. Nature Sustainability, 1, 28–30. Environmental Change, 17, 445–459. Zeelenberg, M., Nelissen, R. M. A., Breugelmans, S. M., & Pieters, R. (2008). On emotion Mellers, B. A., & McGraw, A. P. (2001). Anticipated emotions as guides to choice. Current specificity in decision making: Why feeling is for doing. Judgment and Decision Directions in Psychological Science, 10, 210–214. making, 3, 18–27. Pandey, N., Rupp, D. E., & Thornton, M. A. (2013). The morality of corporate Zeelenberg, M., van der Pligt, J., & de Vries, N. K. (2000). Attributions of responsibility environmental sustainability: A psychological and philosophical perspective. In and affective reactions to decision outcomes. Acta Psychologica, 104(3), 303–315.