Journal of Behavioral and Experimental Economics 102 (2023) 101956 Contents lists available at ScienceDirect Journal of Behavioral and Experimental Economics journal homepage: www.elsevier.com/locate/jbee Nudging low-carbon consumption through advertising and social norms Juana Castro-Santa a, b, *, Stefan Drews a, Jeroen van den Bergh a, c, d a Universitat Autònoma de Barcelona, Edifici Z, Carrer de Les Columnes s/n, Campus de la UAB, Institute of Environmental Science and Technology, Bellaterra 08193, Spain b Universidad de Loyola Andalucía, Sevilla, Spain c ICREA, Barcelona, Spain d School of Business and Economics & Institute for Environmental Studies, VU University Amsterdam, the Netherlands A R T I C L E I N F O A B S T R A C T Keywords: Widespread advertising of high-carbon goods challenges a shift to low-carbon consumption which is needed to Emissions reduction limit climate change. We test the pure and combined effects of advertising and communicating a social norm on Consumption low- and high-carbon consumption. This involved presenting to 2728 US citizens an imitation Facebook Advertising homepage containing green and non-green advertising as well as weak and strong social norms to nudge low- Social norms Experiment carbon consumption. In isolation, both green advertising and social norms were effective in promoting low- carbon choices. But when combined, advertising dominated choice and counteracted the positive effects of the social norm. We show that this result is due to advertising affecting more decision channels than the social norm. It suggests that low-carbon norms have a limited effectiveness in changing consumer preferences in a world dominated by advertising. 1. Introduction providing information to consumers with the aim to nudge their pur­ chases towards low-carbon options. Nonetheless, it is likely that Mitigating climate change requires that consumption patterns different types of information provision set in motion unique psycho­ change towards low-carbon. At the same time, consumers are being logical mechanisms which affect environmentally-relevant behaviors. swamped with product advertisements incentivizing high-carbon con­ Advertising influences choices by providing repeated information that sumption. It is estimated that the average American is exposed to around appeals to goals, self-concepts and desirable product qualities, in turn 4000 to 10,000 advertisements daily (Marshall, 2013), the majority of creating emotional associations and positive perceptions towards the which are for high-carbon products, that is, those which emit consid­ advertised product (Florack & Scarabis, 2006). erable CO2 emissions over their lifecycle (Annala & Vinnari, 2019). Similarly, ample research shows social-norm communication can be Moreover, targeted advertisements made possible by online platforms used to influence people’s choices (McDonald & Crandall, 2015). For such as Facebook and Google have dramatically increased the effec­ instance, social norms have been proposed an effective way to nudge tiveness of commercial advertising (Mochon et al., 2017). Studies indi­ certain pro-environmental behaviors by conveying information cate that advertising may explain up to 65% of the total variance in regarding what others do or think, in turn influencing people’s envi­ consumption patterns (Brulle & Young, 2007). Against this background, ronmental attitudes and behaviors (Byerly et al., 2018; Ölander & relevant questions are whether advertising for high-carbon products Thøgersen, 2014; Schultz et al., 2008). Both information instruments, i. reduces the effectiveness of climate policies aimed at controlling emis­ e. advertising and social norms, can make use of social media. In fact, sions associated with consumption, and whether green advertising, i.e. most advertising takes place today on such media. It is unclear which of advertising for low-carbon products, can promote consumption of these information mechanisms is more effective at influencing consumer low-carbon alternatives. behavior, and moreover, whether they interact, i.e., whether the pres­ Advertising can be regarded as a type of information provision ence of one influences the effectiveness of the other in influencing (Ekelund et al., 1995; Nelson, 1974). This allows it to be considered in a behavior. Here we examine their individual effectiveness on shifting broader setting of motivations, incentives and policies that involve consumption behavior towards low-carbon products, and whether * Corresponding author at: Universitat Autònoma de Barcelona, Edifici Z, Carrer de Les Columnes s/n, Campus de la UAB, Institute of Environmental Science and Technology, Bellaterra 08193, Spain. E-mail address: juana224@gmail.com (J. Castro-Santa). https://doi.org/10.1016/j.socec.2022.101956 Received 25 February 2022; Received in revised form 26 October 2022; Accepted 8 November 2022 Available online 18 November 2022 2214-8043/© 2022 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 nudging with norms for low-carbon consumption is an effective strategy environmental benefits (Hartmann, Ibáñez & Sainz, 2005). in the presence of advertising. Naturally, the extent to which consumer perceptions of products are To this end, we undertook an incentivized experiment presenting a influenced by cues and claims of green advertising is moderated by other choice between a high- and a low-carbon product. Our aim was to mimic dimensions, such as the consumers’ involvement with environmental a real-life scenario to test if consumer choice between low-and high- issues (Matthes et al., 2014; Petty & Cacioppo, 1979). Overall, the carbon products can be affected by information provision. We assess the reviewed evidence suggests that advertising strengthens consumer at­ separate and combined effects of product advertising and social-norm titudes towards a product by using a combination of green claims and communication on choice. In particular, we test for potential interac­ cues to create favorable perceptions of attributes such as quality or tion effects between the two information mechanisms, namely, through popularity, and in the case of green advertising, reduced environmental two scenarios: one in which advertising and the norm work in the same impact. direction (i.e., both favor a low-carbon option); and another where they work in opposite directions (i.e., advertising for a high-carbon option 2.2. The influence of social norms and a social norm for low-carbon consumption). Finally, we undertake a structural-equation-model analysis, which identifies the channels Within information policy research, social norm management has through which the two types of information (advertising and social been praised as being particularly effective at encouraging pro- norm) affect behavior. environmental behaviors, such as reducing home energy use and recy­ cling (Allcott & Kessler, 2019; Nyborg et al., 2016). The effectiveness of 2. Background and present study communicating a social norm depends on whether it can generate peer pressure through social information, or activate personal norms related 2.1. The influence of advertising to a moral behavior (Farrow et al., 2017). Norms can relate to the perceived appropriateness or moral nature of a particular behavior Psychologists have long established that repeated exposure to (injunctive norms), or to observations of what most people do inde­ advertising leads to a stronger preference for a product (Becknell Jr pendent of the appropriateness of the behavior (descriptive norms). The et al., 1963; Krugman, 1968; Zajonc, 1968). This exposure phenomenon percentage of people that follow a norm (i.e., how many people engage has been explained through the advertised products creating a sense of in the respective behavior) determines whether the norm is weak or familiarity that reduces the level of risk in choices, and exempts con­ strong, that is, followed by the minority or the majority (de Groot & sumers from evaluating too much information, thus avoiding overly Schuitema, 2012). complex decision making (Hansen & Wänke, 2009). Neurological evi­ Social norms only followed by a minority may fail to influence dence on this familiarity effect shows that exposure to known-brand others’ behaviors or even backfire (Cialdini, 2003). Nonetheless, recent logos activates emotional neural responses in consumers (McClure studies suggest that when accompanied by normative information, weak et al., 2004; Plassmann et al., 2012). Advertising typically combines norms can influence others’ behaviors either by conveying that a product information with pleasant cues, such as colors, music, or im­ behavior is desired, or even perceived as trending (Mortensen et al., ages, which elicit affective processing based on induced emotions 2019). For these reasons, it is not surprising that effective norm-based (Bagozzi et al., 2017). Highly advertised products are perceived as better interventions in the environmental domain have often combined both than non-advertised ones (Rossiter & Percy, 1980; Staats & Staats, descriptive and injunctive information (Schultz et al., 2007). Examples 1958), particularly in terms of having higher quality (Moorthy & include people reducing their energy use by 2% on average if they learn Hawkins, 2005; Moorthy & Zhao, 2000) and as being more popular (Kim that most of their neighbors consume less than they do (Allcott, 2011), & Min, 2014). Both of these perceptions serve as predictors of purchase re-using towels in hotel rooms if they are told that most other guests do intention. In this context it is difficult for new brands to enter the so (Goldstein et al., 2008), or increasing recycling by 25% after being market, which is the case for many low-carbon options. exposed to recycling campaigns transmitted via TV and radio stations, In the last decades, green advertising has emerged as a successful communicating that most local residents recycle (Cialdini, 2003). These marketing strategy to promote product or services with relatively little types of norm-based interventions, apart from describing others’ ac­ environmental impact (Henion & Kinnear, 1976). Consumption of tions, display to individuals that their behavior is relevant from an low-carbon products is often encouraged by advertisements making environmental perspective, which activates any latent environmental claims or using green color, labels or nature images, in pro-environmental preferences (Schultz et al., 2008). order to signal reduced environmental impact. Given the rapid rise of The activation of personal norms relevant to moral behaviors re­ environmental concern among the population, the association of prod­ quires that an individual (a) becomes aware of the consequences of their ucts with environmental benefits presents a topical opportunity for behavior on others’ welfare; (b) holds personal norms regarding these marketing firms (Pickett-Baker & Ozaki, 2008). It has been found that consequences; and (c) feels some capability to control the outcomes of consumers can be more strongly persuaded by advertisements which their behavior (Schwartz, 1973). Translated to emphasize green rather than non-green product attributes (Ku et al., environmentally-relevant behaviors, norm activation will be mediated 2012), and that such advertisements induce more positive attitudes to­ by the concern a person has for the environment, the extent to which wards brands (D’Souza & Taghian, 2005; Thorson et al., 1995). These they feel responsible, and their perception of an action’s effectiveness in positive attitudes generated through green advertising can further pre­ having an environmental impact (De Groot & Steg, 2009). These psy­ dict consumer purchase intention (Leonidou et al., 2011; Purohit, 2012). chological constructs can help explain when norm activation takes place Moreover, green advertising not only generates more positive atti­ for many environmentally-relevant behaviors (Oreg & Katz-Gerro, tudes towards a product through consumer perception of reduced 2006). For example, research has shown that social norms influence environmental impact, but also improves other consumer attitudes. For environmental behaviors only if individuals perceive that the particular example, some studies find that people rate an eco-labelled product as behavior (e.g., recycling) has a significant impact on reducing emissions having a better taste or performance than an identical non-labeled (Doherty & Webler, 2016). product, which has been suggested to be the result of a green-halo ef­ Social norms can induce behavioral change beyond the realm they fect (Sörqvist et al., 2015). Such effects on consumer attitudes are focus on. For example, it has been found that communicating social typically achieved using green cues. Not only do consumers associate norms regarding car-sharing can encourage recycling (Evans & Stano­ green color with a reduced environmental impact of products (Lim et al., vich, 2013). This is relevant to policy and raises the question whether 2020), but also research shows that images of nature in advertising have one should focus on explicit, concrete, social norms, such as “the ma­ higher positive effects on brand attitude than textual information about jority of consumers buy organic apples”; or implicit, general, social 2 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 norms. In the case of a social norm aimed at nudging consumption to­ order to create a realistic setting where both stimuli could be presented, wards low-carbon products, an implicit norm seems more practical for the experiment employs a simulated Facebook homepage where posts two reasons: (a) it will affect the majority of products in the market; and with green and non-green product advertising are presented. The social (b) it avoids the influence of conformity and reactance along the lines of norm takes the form of a Facebook poll showing most/few people “someone is trying to manipulate my choices”. So far, available exper­ consume low-carbon products nudging towards low-carbon consump­ imental evidence for social norms promoting a shift to lower-carbon tion. After being presented with the Facebook imitation page, partici­ options in consumption deals predominantly with explicit norm pants make an incentivized choice between a high-and a low-carbon communication. The findings of this research are, however, inconclu­ product. By including both green and non-green product adverts and a sive. While three studies report shifts to ‘greener’ options when in­ weak or strong social norm, we can assess the separate and combined dividuals become aware that most others consume green goods effects of the two types of information on consumer choice. (Demarque et al., 2015; Kim, Lee & Hur, 2012; Sparkman & Walton, Furthermore, we explore the underlying mechanisms behind the 2017), other studies report null results (DellaValle & Zubaryeva, 2019) observed effects from the social norms and advertising by undertaking a and even “boomerang effect” from norm presentation, shifting con­ structural-equation-model analysis for low-carbon choice, using sub­ sumption in the opposite direction (Richter et al., 2018). That is, when jects’ responses of a post-experimental survey. This allows the assess­ consumers were presented with a social norm nudging eco-labelled ment of the distinct decision channels suggested in the literature. In seafood in a supermarket (i.e., a message saying most consumers buy particular, we test whether green and non-green advertising effects on labelled seafood), they overall bought more seafood; both labelled and choice are mediated by product perceptions (e.g., quality, popularity unlabeled. Additionally, explicit norm communication in the way it is and environmental impact), and investigate if there are variations be­ applied in most studies so far is relatively unrealistic. In real life, it will tween the two types of advertising. For social norms we test whether its rarely be allowed for a policy to campaign for a given product since effects on low-carbon choice are mediated through norm activation there are countless products in the market and this could eventually go (environmental concern, responsibility and self-efficacy) or only against market regulation rules. Finally, the research to date lacks an through peer effects. Finally, we compare the decision channels used by important dimension of consumption, namely its extensive interaction social norms and advertising to interpret experimental results. with commercial advertisement. 3. The experiment 2.3. The present study 3.1. Experimental design and procedure In this study we investigate whether social norms nudging towards low-carbon consumption alongside green advertising can be effective at A 3 × 3 between-subjects survey experiment was designed by shifting consumption from strongly advertised, carbon-intensive prod­ manipulating product advertising (no advertising vs. green advertising vs. ucts towards less advertised, low-carbon options. We consider various non-green advertising) and the low-carbon social norm (no social norm vs. questions such as: Do social norms and advertising, both representing strong social norm vs. weak social norm) in an imitation Facebook cases of information provision, affect consumption decisions in a similar homepage (see Fig. 1). A total of 2728 US participants (51% women, way? Which is more effective, and why? And particularly relevant from mean age=38.7) were recruited via Amazon Mechanical Turk to a policy-making perspective are their interactions; for example, are so­ participate in an incentivized “observation task” and were randomly cial norms for low-carbon consumption effective in the presence of assigned to one of the nine treatments (approximately N = 300 by advertising for high-carbon alternatives? treatment, see Fig. 1). The participation fee was set at 1 dollar for a To address these questions, we run a controlled experiment in which seven-minutes task. In addition, participants could receive $0.20 for we manipulate the strength of “low-carbon social norms”, by suggesting each correct answer on five questions following the Facebook Home­ that a majority or minority of people consume low-carbon products, and page, that served as an attention test (SM10). The size of the sample per through the type of advertising, namely for either a low-carbon product treatment was set at 300, well above that of studies with similar designs (green advertising) or a high-carbon product (non-green advertising). In (see for example Sparkman & Walton, 2017), assuring sufficient Fig. 1. Experimental treatments. 3 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 statistical power to detect smaller effects and their magnitude. non-green advert for Iameco or Samsung respectively, displaying a The experimental sequence was as follows. Once participants agreed picture of the advertised tablet and the brand logo below. The slogans to participate in the “observation task”, they were instructed to look used in the simulated ads were taken from actual advertisements of both over an “image”, about which they would later answer some questions. brands. The Iameco advert had an environmental claim regarding the The image presented was the imitation Facebook homepage with tablet’s carbon emissions which read that “The Iameco touch screen different posts created for the purpose of the experiment. The questions computer has a carbon footprint that is 70% less than the average PC”, that followed were regarding the content of the posts (attention test). whereas the advert for Samsung did not use any environmental claims. After answering these questions, participants were asked to choose be­ In the social norm treatment, the post displayed a Facebook poll tween a low- and high-carbon tablet. The choice was presented in a where compliance with a low-carbon consumption behavior was either a screen containing a picture of both tablets accompanied by explicit de­ majority (strong social norm) or minority (weak social norm) of re­ scriptions of their technical characteristics (RAM memory, processor, spondents. The social norm used a low-carbon behavior correlated with etc.). In the description of the product, the low-carbon tablet had one the one we want to influence (see Kallgren et al., 2000) for a similar important additional characteristic, namely a certified eco-label. norm manipulation). The norm was framed as a poll asking, “Do you use The choice was incentivized, in the sense that one randomly chosen energy-saving light bulbs?” with two possible answers of “yes” or “no”. participant would win a $200 discount coupon for the tablet s/he had The number of respondents for each option was also displayed. This previously chosen. The likelihood of winning, approximately one over behavior was taken from a pre-survey (N = 87) that explored which 3000, was the same for every participant as it was independent of the low-carbon behaviors correlated with the one we examine, that is, choice. Finally, participants were asked about their perceptions of the buying low-carbon electronic appliances (see pre-survey results in S9 of high- and low-carbon products, their beliefs relating to climate change, the SM). In treatments involving both advertising and a social norm, and their basic socio-demographic information (for more details on the both types of posts appeared in the Facebook simulated page presented experimental procedures, see section S1, and for the full questionnaire to participants. and instructions see sections S10–12 of the Supplementary Material In previous studies, social norms have been manipulated by directly (SM)). stating to subjects that a certain percentage of people undertake a certain behavior (Rhodes et al., 2020). However, such norm manipula­ 3.2. Choice of product tion is prone to strong experimenter demand effects (Zizzo, 2010). In particular, when subjects are explicitly told that ‘most others chose Tablets were chosen as the product in our experimental setup for option x’ right before they make their decision, they are more likely to several reasons: (a) They are relevant to a large number of participants, choose x over the alternative(s) for reasons other than the induced social (b) they are likely to be advertised in Facebook, (c) their production is norm. For example, participants might try to adapt their behavior to associated to considerable carbon emissions, and (d) they represent a what they believe to be the experimenter’s hypothesis. Moreover, such non-trivial purchase decision given their high price. In the experiment an approach is difficult to implement in a realistic Facebook setting. the “low/high-carbon” product categorization refers to the extent of Instead, we use an innovative design that introduces the norm using a carbon emissions during lifecycle, particularly in production. As the aim poll feature that already exists in Facebook, thus allowing for a natural of the study was to test the efficacy of advertising and social norms in induction of the norm. shifting consumption towards lower-carbon products in the presence of The image Fig. 2 shows the Facebook posts used for advertising and a highly advertised brand, we used a well-known brand associated with social norm manipulations. In combined-effects treatments, both a post relatively high-carbon emissions, namely Samsung (Cook & Jardim, with advertising and a post with a social norm were added to the 2017). The low-carbon alternative (i.e., the Iameco tablet) was chosen as baseline Facebook homepage. its lifecycle emissions are 70% less, has a certified eco-label, and is ex­ pected to be less familiar to the participants. 3.4. Hypotheses While the two tablets (Iameco and Samsung) have similar functional characteristics and approximately the same market price, they differ To answer our research objectives, we formulate four hypotheses with respect to lifecycle emissions. In addition, they might differ in other based on the reviewed literature. First, we hypothesize that green and relevant dimensions (such as perceived quality, attractiveness, and fa­ non-green product advertising will shift average consumer choice to­ miliarity, which we elicit in a post-experimental questionnaire). Such wards the advertised product relative to the baseline condition. Since differences do, however, not affect the interpretation of the results, since the experiment is framed as a choice between a high- and low-carbon they remain the same across treatments. We decided against using the brand and the high-carbon brand is well-known by participants, we same brand for the alternative options, such as Samsung vs. green Sam­ expect choice in the baseline group to be concentrated in the non-green sung, since there is evidence that information about the environmental Samsung brand. Thus, we anticipate more low-carbon choice for par­ attributes of the same brand makes the “green” version always more ticipants assigned to the green advertising treatment compared to desirable (see Hartmann & Apaolaza-Ibáñez, 2015). With the chosen baseline, and less low-carbon choices if they were assigned to non-green setup of two brands, we are able to assess the real-word case of shifting advertising. choice from a well-known, carbon-intensive brand to a new and A second hypothesis is that low-carbon choice will be higher among less-known, low-carbon alternative. Crucially, all the important features participants assigned to the social norm treatments, relative to the that affect the subjects’ choice are present across all treatments. baseline condition. Given that the norm aligns what people do Therefore, any difference between treatments is attributable only to (descriptive norm) with what people typically approve of, and in line treatment effects and not to the specific differences of the two tablets. with the literature (Schultz et al., 2007), we do not foresee the weak For more details on the choice of the products please see S10–12 from social norm to backfire. We do, however, expect the weak norm to have the SM. a smaller positive effect than the strong norm. Third, we hypothesize that the effects of advertising and social norms 3.3. Manipulation of advertising and the social norm on participants’ choice change when the two instruments are presented together, that is, they interact. We expect participants presented with In the baseline condition, the Facebook simulated page contained social norms and green advertising to make more low-carbon choices five neutral posts. In the treatment conditions, the Facebook image was than those in the baseline since they both induce low-carbon choice. identical except for one or two posts that differed depending on the Specifically, we expect that their joint effect (green advertising + social treatment. In treatments involving advertising, the post was a green or norm) is larger than the sum of their single effects. Conversely, we 4 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 Fig. 2. Experimental stimuli. expect a negative interaction effect between a social norm and non- Results show that green advertising and strong social norm treatments green advertising as they enhance opposite choices. That is, we hy­ significantly affect low-carbon choice after controlling for socio- pothesize that low-carbon choice will be higher than in the case par­ demographic factors (see results of logistic regression in SM S3). In ticipants are presented with non-green advertising alone, but lower than the baseline condition 7.6% participants chose the low-carbon option, when presented with social norms alone. showing this represents a non-trivial decision where the vast majority Finally, our fourth hypothesis is about the mechanisms and medi­ prefer the non-green tablet from the familiar brand Samsung. However, ating channels of advertising and social norms to affect behavior. in the treatment with the strong social norm, the low-carbon choice Consistent with previous research, we predict that the mediating vari­ (Iameco tablet) is almost twice that of the baseline (13.1%, p = 0.016), ables underlying the influence of advertising and social norms on while the weak social norm has a smaller, yet still (marginally) signifi­ behavior are different. In particular, we expect advertising to influence cant effect (11.5%, p = 0.055). Nevertheless, the effects between strong behavior mainly by affecting consumers’ perceptions of the product and weak social norms are statistically indistinguishable (z = 0.601; p = advertised, while social norms to influence behavior by activating per­ 0.5479, Mann Whitney (MW) test), suggesting that behavioral effects sonal norms. may be present merely due to norm exposure, regardless of its strength. To our knowledge, this is novel evidence for the effect of implicit social 4. Experimental results norms regarding low-carbon consumption on other consumption choices. 4.1. Treatment effects Bars show the proportion of participants that chose the low-carbon option in baseline, alongside conditions with only the social norm Across treatments, most participants chose the well-known carbon- (weak/strong) or only advertising (green/non-green). Error bars indi­ intensive tablet (Samsung), while only 15.6% (419 out of 2728) opted cate 95% confidence intervals. for the low-carbon tablet (Iameco). This is not surprising given that Green advertising, however, almost quadruples low-carbon choice 99.9% of participants reported previous knowledge of the Samsung compared to the baseline (27.8%, p<0.001). Importantly, this increase brand, while only 8% reported previous knowledge of the Iameco brand is more than double that of the strong social norm treatment (z = 4.488; (further details on sample and demographics are provided in S2 from the p<0.001, MW test), revealing its high effectiveness at influencing SM). We first look at results for advertising-only and social-norm-only choice. Lastly, non-green advertising had virtually no effect (7.9%; p = treatments. Fig. 3 shows the difference in mean choices of the low- 0.833), something to be expected given the already very high percentage carbon product (Iameco) between treatment and baseline, indicating of participants choosing the high-carbon option (See the complete ma­ their effectiveness at shifting choice towards the low-carbon option. trix with MW tests in SM S4). 5 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 Fig. 3. Isolated effects of advertising and social norm on low-carbon choice. When analyzing treatments combining social norms and advertising, the 95% confidence intervals of the green and non-green advertisement the dominant role of advertising is further revealed. Fig. 4 shows the treatment, respectively, without any social norm manipulation. The percentages of participants choosing the low-carbon product across all figure reveals that there are no interaction effects between advertising nine treatments. Treatments involving green and non-green advertising and social norms. When they are combined, shares of low-carbon choice are depicted in green and red respectively, while treatments without any are almost exclusively determined by advertising. Combining green advertising are depicted in blue. The green- and red-shaded areas depict advertising with a strong social norm does not significantly increase the Fig. 4. Combined effects: Green and non-green advertising counteract social-norms. 6 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 share of the low-carbon choice compared with the green advertising- social norms posts. Even though both stimuli were of the same size in the only treatment. In fact, mean Iameco choice slightly decreases from Facebook image, it can still be the case that advertising triggers more 27.8% to 26.4% (z = 0.397; p = 0.692, MW test) and the same occurs attention, for reasons such as the presence of brighter colors or more when green advertisement is combined with a weak social norm attractive images, or simply because people are more used to check out (z=− 0.025; p = 0.980, MW test). ads than posts in Facebook. The dominance of the green advertisement Visually, this is depicted by low-carbon choice always remaining would then be on the grounds of attention. within the 95% confidence interval of green advertising only. Similarly, To test this, we asked subjects whether they remembered having seen combining non-green advertising with either a strong or a weak social the stimuli that they were presented with, i.e., either the post with the norm does not significantly change low-carbon choice compared with social norm or the advertisement. Correct answers were monetarily the non-green advertising-only treatment (z=− 0.668; p = 0.504, and z incentivized. Of respondents, 52% recall having seen the Iameco ad and = 0.779; p = 0.436, MW tests, respectively) and choice remains within 76.5% the Samsung ad. The post with the strong social norm was the 95% confidence interval of non-green advertising only. It is worth recalled by 55% of the subjects, and with the weak social norm by noting, however, that both green and non-green advertising have 42.7%. Therefore, the percentage of subjects aware of the stimuli in the greater dominance over the weak norm than over the strong norm; green advertising and the social norm treatments are statistically whenever combined with the weak norm, advertisement causes larger indistinguishable. changes towards what is advertised. To summarize, in the absence of any This suggests that, a similar percentage of subjects were “treated” advertising effects, social norms for low-carbon consumption might across treatments (which is comforting from an experimental design appear as effective in shifting choice towards low-carbon products. point of view: a significantly higher percentage of treated subjects might However, when social norms are communicated in the presence of point at an experimental artifact). To explore this further, we compare advertising, their effects are practically nullified. low-carbon choices between each subgroup within each treatment: The dots in the graph display the share of low-carbon choice for those who recall having seen the stimuli (the “treated” subjects) and advertising and social-norm nudges, and its combinations. Green and those who do not (the “untreated” subjects). One should expect the red colored dots display treatments presenting green and non-green probability of choosing the Iameco tablet to be the same among subjects advertising respectively, while treatments without any advertising assigned to the baseline and among those not remembering being pre­ (baseline and social norm only) are depicted in blue. The x-axis indicates sented with the stimuli. Likewise, the probability of choosing the low- the presence of a social norm nudging towards low-carbon consumption carbon tablet should be higher for those who were aware of being pre­ and its type (weak/strong). Shaded regions depict 95% confidence in­ sented with the advertisement or social norm. Fig. 5 presents the results: tervals for green and non-green advertising only treatments. low-carbon choice among participants who do not recall the social norms is not statistically different from the baseline (z = − 0.390; p = 5. Mechanisms and decision channels 0.6966, MW test). Yet, in the advertising treatment, even theoretically “untreated” subjects (those not recalling the Iameco ad) chose Iameco 5.1. Recall with higher probability than in the baseline treatment (20% vs 7,6%, z = − 7.280; p<0.001, MW test). These results suggest that green adver­ To explore the reasons behind the dominance of advertising over the tising might be affecting subjects via less conscious channels. social norms (and especially of green advertising), we first examine The figure compares low-carbon choice between participants who whether more subjects observed the product advertisements than the remember being present with the stimuli in the attention test and those Fig. 5. Low-carbon choice among participants that recall and do not recall the stimuli. 7 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 who did not, both relative to the baseline treatment. The percentages in in the case of low-carbon consumption are: (a) environmental concern the bars show the total proportion of participants in the treatment that (here specifically “climate concern”); (b) feeling of responsibility remember being presented with the stimuli. regarding the environmental impact of one´s consumption emissions (“climate responsibility”); and (c) the belief that consumption decisions are significant for the environment (“self-efficacy”). We measured these 5.2. Mediation analysis for advertising and social norms variables through a post-experimental survey and used them to assess all of the potential relationships they might have with the observed treat­ In this section we investigate the decision channels through which ment effects of advertising and social norms (see SM S5). advertising and social norm influence choice, i.e. the extent to which Our measurement model for low-carbon choice uses N = 1488 ob­ treatment effects are explained by participants’ (a) product perceptions servations from baseline, advertising and social norm treatments. and/or (b) psychological constructs triggering norm activation, as sug­ Treatments combining both instruments (norm and advertising) were gested in the literature. To analyze this, we use responses to the post- excluded due to the impossibility of disentangling their separate effects experimental survey, which measured each of these dimensions. We on choice. Overall, the structural model estimates: (a) effects of adver­ perform an estimation using a Structural Equation Model (SEM) rather tising and the social norm on participants’ product perceptions and than simple regression techniques, in order to properly establish norm activation variables, (b) effects of product perceptions and norm whether there are mediation effects (see Frazier et al., 2004). Can, for activation variables on low-carbon choice, and (c) direct effects of social example, an increased perception of the quality of the Iameco tablet, norms and advertising on choice. The results show that the model has a generated through green advertising, significantly predict participant good fit (χ2 =34.934 [df =9]***; RMSEA=0.044; SMRM=0.025; choice? Using a SEM estimation enables us to test simultaneously this CFI=0.985), explaining 22% of the variance in choice (see SM S6–8 for effect and other relationships between the proposed variables. measurement details). Fig. 6 shows the estimated relationship between We construct a behavioral model to identify the specific channels treatments, mediators and choice in the form of a path diagram. The through which advertising and the social norm affect low-carbon choice lines in the diagram in Fig. 6 indicate significant estimates (p<0.05) and in the experiment Fig. 6. According to the literature, advertising affects show the corresponding coefficients. choice via improved perceptions and attitudes towards the products The results from the SEM estimation help identifying the channels (Zajonc, 1968). Thus, we include product perceptions such as quality, through which each treatment affects behavior. Non-green advertising popularity and environmental impact as potential mediators of adver­ has no direct effect on low-carbon choice but, surprisingly, it negatively tising effects on low-carbon choice. Furthermore, social norm effects are influences the quality perception of the low-carbon product (β=− 0.18; attributed to peer effects and the activation of personal norms. While the p = 0.023), which among all mediators is the largest predictor of choice previous analysis shows that peer effects are significant, there are no (β=0.073; p<0.001). As for green advertising, it shows both direct differential effects between the strong and weak social norm in terms of (β=0.12; p<0.001) and indirect effects on choice through increasing a majority versus minority of people complying with the norm. Here we perceptions of all product characteristics as expected from the literature focus here on norm activation as the mechanism of influence. The (quality β=0.30, popularity β=0.35 and environmental impact β=0.70; relevant psychological constructs for norm activation (Schwartz, 1973) Fig. 6. Direct and Indirect Effects of Advertising and the Social Norm on Low-Carbon Choice. 8 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 p<0.001). Weak and strong social norms have no direct effects on activation or by peer effects. Instead, it influenced choice through choice, while both significantly increase the perceived environmental improving environmental perceptions of the product. To our knowledge, impact of the product (β=0.19; p = 0.014 and β=0.32; p<0.001, this is a novel result about how norms can affect consumption choices respectively). Although environmental impact predicts choice through influencing perceptions of product attributes. The finding that (β=0.053; p<0.001), it does so to a lesser degree than the other product both green advertising and social-norm communication influence characteristics that are influenced by advertising. Weak norm effects are behavior through changing participants’ perceptions of the products` limited to influencing environmental impact, whereas the strong social associated environmental impact might explain the lack of positive in­ norm also has a positive effect on climate responsibility (β=0.17; p = teractions when they are combined. 0.031). Nonetheless, this does not translate into increased low-carbon It must be considered when interpreting results that green adver­ choice since none of the variables for norm activation have a signifi­ tising was specific to a product (“buy Iameco”), while the social norm cant effect on low-carbon choice. nudging low-carbon consumption (buying energy saving lightbulbs), Lines indicate significant relations (p<0.05) between treatments, addressed a consumption behavior that was correlated but not specific mediators and choice. Covariances between-error terms (εi) included to to the target product. Using distinct behaviors was done to avoid de­ account for omitted-variable bias. mand effects. This can partly explain the smaller effect of the social Overall, the SEM analysis reveals the underlying reasons behind the norm compared to advertising. Moreover, we present a novel norm observed treatment effects. Particularly, it shows that advertising in­ manipulation that shows how a Facebook post can be used to increase volves the strongest channels for influencing behavior. While product conformity with low-carbon consumption. It should be noted, though, attitudes are the strongest predictors of choice and are affected by that different representations of a norm in a Facebook homepage might advertising, norm activation is not fully achieved by the social norm, affect the strength of its effects. Further research could test whether and thus cannot explain low-carbon choice. In fact, the effect achieved communicating a more explicit or specific norm results in stronger ef­ by norm communication in this case is through affecting a product fects on choice. Yet, for policy-making to communicate social norms attitude, namely the perceived environmental impact, which is the same separately for every product, or even brand, would be extremely costly as by advertising. Nevertheless, the influence of green advertising on the and thus practically impossible. It might also cause negative reactions environmental perception of the Iameco tablet is larger than that of the from consumers feeling manipulated. Indeed, the social norm used here social norm. This can explain the dominance of advertising over norm was implicit enough to expect norms related to low-carbon consumption communication and highlights a new channel, different to norm acti­ to have similar or even stronger effects on choice. vation and peer-influence, through which a social norm for low-carbon A complementary explanation for the dominant effect of advertising consumption can affect low-carbon choice. over norms derives from the notion of cognitive effort. This can be explained by short messages and graphic cues contained in advertising 6. Discussion requiring relatively little mental processing in comparison with social norms (Garbarino & Edell, 1997; Zajonc, 2001). This may also explain The experiment provides evidence on the effectiveness of advertising the non-conscious effects achieved by advertising, i.e. it was able to and social norm communication in shifting choice from a highly- influence choice even when participants were not aware of having seen advertised carbon-intensive product to a low-carbon alternative. The the ad. Nudge-like policies aimed at discouraging high-carbon con­ results reveal that both green advertising and low-carbon social norms sumption with social norms could also follow the principles of good presented through an imitation Facebook homepage were effective in advertising (Burchell et al., 2013). Copying advertising strategies such shifting choice towards lower carbon alternative. In isolation, green as use of images, repetition and framing will make messages in advertising is twice as effective at shifting consumption towards low- norm-communication capable of rapidly becoming familiar to people carbon product as the social norm for low-carbon consumption. No ev­ and effectively influencing their choices. This is in line with research idence of positive interactions between advertising and social norms was showing that consumers can reduce their carbon footprint when infor­ found. When combined, we observed a dominance of advertising in mation is simple (Camilleri et al., 2019; Kanay et al., 2021) and when discouraging (non-green adverts) or incentivizing (green adverts) low- choices involve low cognitive effort (Isley et al., 2016), carbon choice. This suggests that in the presence of advertising for Finally, while the obtained results could be interpreted as a sup­ highly known carbon-intensive brands, communicating a social norm is porting the belief that commercial green advertising by firms is an insufficient to induce relevant behavioral changes towards low-carbon effective way to encourage the consumption of low-carbon alternatives, consumption. Similarly, in the presence of green advertising, commu­ the risk of misinformation and false green claims remains a reason for nicating a social norm adds little. These results provide novel evidence skepticism (Schmuck et al., 2018). It is known that consumers tend to on social norms nudging low-carbon consumption in a more realistic perceive any product that claims to be “green” in some dimension as setting and in presence of advertising, which contributes to broader being less environmentally harmful in general (Gershoff & Frels, 2015; research on how social norms perform under different contextual factors Magnier et al., 2016). Even people that are well informed or have (Bergquist et al., 2019). pro-environmental inclinations may be sensitive to such confusion. The reasons underlying the dominance of advertising over social Therefore, governments should, in the interest of welfare and environ­ norms were explored by performing an analysis of the mechanisms and ment, regulate what can and may be advertised as low-carbon or green. decisions channels used by both stimuli to influence choice. Results show green advertising is able to influence choice even if participants do Funding not recall the ad, suggesting an unconscious mechanism. Moreover, the SEM analysis employed shows that, in line with theory, commercial This study has received funding through an ERC Advanced Grant advertising affects attitudes towards products. However, we discover from the European Research Council (ERC) under the European Union’s additional limitations to shifting choice from highly-advertised brands Horizon 2020 Research and Innovation Programme (grant agreement towards low-carbon ones. Participants exposed to Samsung’s advertising no. 741,087). not only perceived this brand better, but importantly, considering quality was the largest predictor of choice among product characteris­ Data availability tics, also perceived a lower quality of the low-carbon alternative (Iameco). Data will be made available on request. Communicating a social norm through a Facebook poll, on the other hand, did not achieve the expected influence on choice strictly by norm 9 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 Supplementary materials Isley, S. C., Stern, P. C., Carmichael, S. P., Joseph, K. M., & Arent, D. J. (2016). Online purchasing creates opportunities to lower the life cycle carbon footprints of consumer products. Proceedings of the National Academy of Sciences, 113(35), 9780. Supplementary material associated with this article can be found, in LP –9785. the online version, at doi:10.1016/j.socec.2022.101956. Kallgren, C. A., Reno, R. R., & Cialdini, R. B. (2000). A focus theory of normative conduct: When norms do and do not affect behavior. Personality and Social Psychology Bulletin, 26(8), 1002–1012. References Kanay, A., Hilton, D., Charalambides, L., Corrégé, J.-. B., Inaudi, E., Waroquier, L., et al. (2021). Making the carbon basket count: Goal setting promotes sustainable Allcott, H. (2011). Social norms and energy conservation. Journal of Public Economics, 95 consumption in a simulated online supermarket. Journal of Economic Psychology, 83, (9–10), 1082–1095. Article 102348. Allcott, H., & Kessler, J. B. (2019). The welfare effects of nudges: A case study of energy Kim, H., Lee, E.-. J., & Hur, W.-. M. (2012). The normative social influence on eco- use social comparisons. American Economic Journal: Applied Economics, 11(1), friendly consumer behavior: The moderating effect of environmental marketing 236–276. claims. Clothing and Textiles Research Journal, 30(1), 4–18. Annala, M., & Vinnari, M. (2019). Content analysis of TV food advertising using climate Kim, J.-. H., & Min, D. (2014). The effects of brand popularity as an advertising cue on impact and a nutritional impact index. Ecological Economics, 159, 68–74. perceived quality in the context of internet shopping. Japanese Psychological Bagozzi, R. P., Gopinath, M., & Nyer, P. U. (2017). The role of emotions in marketing. Research, 56(4), 309–319. Journal of the Academy of Marketing Science, 27(2), 184–206. Krugman, H. E. (1968). Processes underlying exposure to advertising. In American Becknell Jr, J. C., Wilson, W. R., & Baird, J. C. (1963). The effect of frequency of psychologist, 23 pp. 245–253). American Psychological Association. presentation on the choice of nonsense syllables. The Journal of Psychology, 56(1), Ku, H.-. H., Kuo, C.-. C., Wu, C.-. L., & Wu, C.-. Y. (2012). Communicating green 165–170. marketing appeals effectively. Journal of Advertising, 41(4), 41–50. Bergquist, M., Nilsson, A., & Schultz, W. P. (2019). A meta-analysis of field-experiments Leonidou, L. C., Leonidou, C. N., Palihawadana, D., & Hultman, M. (2011). Evaluating using social norms to promote pro-environmental behaviors. Global Environmental the green advertising practices of international firms: A trend analysis. International Change, 59, Article 101941. Marketing Review, 28(1), 6–33. Brulle, R. J., & Young, L. E. (2007). Advertising, individual consumption levels, and the Magnier, L., Schoormans, J., & Mugge, R. (2016). Judging a product by its cover: natural environment, 1900–2000. Sociological Inquiry, 77(4), 522–542. Packaging sustainability and perceptions of quality in food products. Food Quality Burchell, K., Rettie, R., & Patel, K. (2013). Marketing social norms: Social marketing and and Preference, 53, 132–142. the ‘social norm approach. Journal of Consumer behaviour, 12(1), 1–9. Marshall, R. (2013). Marketing survival in a digital world. Big RAM Publishing LLC. Byerly, H., Balmford, A., Ferraro, P. J., Wagner, C. H., Palchak, E., Polasky, S., et al. Matthes, J., Wonneberger, A., & Schmuck, D. (2014). Consumers’ green involvement and (2018). Nudging pro-environmental behavior: Evidence and opportunities. Frontiers the persuasive effects of emotional versus functional ads. Journal of Business in Ecology and the Environment, 16(3), 159–168. Research, 67(9), 1885–1893. Camilleri, A. R., Larrick, R. P., Hossain, S., & Patino-Echeverri, D. (2019). Consumers McClure, S. M., Li, J., Tomlin, D., Cypert, K. S., Montague, L. M., & Montague, P. R. underestimate the emissions associated with food but are aided by labels. Nature (2004). Neural correlates of behavioral preference for culturally familiar drinks. Climate Change, 9(1), 53–58. Neuron, 44(2), 379–387. Cialdini, RB. (2003). Crafting normative messages to protect the environment. Current McDonald, R. I., & Crandall, C. S. (2015). Social norms and social influence. Current Directions in Psychological Science, 12(4), 105–109. Opinion in Behavioral Sciences, 3, 147–151. Cook, G., & Jardim, E. (2017). Guide to greener electronics. https://www.greenpeace.or Mochon, D., Johnson, K., Schwartz, J., & Ariely, D. (2017). What are likes worth? A g/usa/reports/greener-electronics-2017/. facebook page field experiment. Journal of Marketing Research, 54(2), 306–317. De Groot, J. I. M., & Schuitema, G. (2012). How to make the unpopular popular? Policy Moorthy, S., & Hawkins, S. A. (2005). Advertising repetition and quality perception. characteristics, social norms and the acceptability of environmental policies. Journal of Business Research, 58(3), 354–360. Environmental Science & Policy, 19(20), 100–107. Moorthy, S., & Zhao, H. (2000). Advertising spending and perceived quality. Marketing De Groot, J. I. M., & Steg, L. (2009). Morality and prosocial behavior: The role of Letters, 11(3), 221–233. awareness, responsibility, and norms in the norm activation model. The Journal of Mortensen, C. R., Neel, R., Cialdini, R. B., Jaeger, C. M., Jacobson, R. P., & Ringel, M. M. Social Psychology, 149(4), 425–449. (2019). Trending norms: A lever for encouraging behaviors performed by the DellaValle, N., & Zubaryeva, A. (2019). Can we hope for a collective shift in electric minority. Social Psychological and Personality Science, 10(2), 201–210. vehicle adoption? Testing salience and norm-based interventions in South Tyrol, Nelson, P. (1974). Advertising as information. Journal of Political Economy, 82(4), Italy. Energy Research & Social Science, 55, 46–61. 729–754. Demarque, C., Charalambides, L., Hilton, D. J., & Waroquier, L. (2015). Nudging Nyborg, K., Anderies, J. M., Dannenberg, A., Lindahl, T., Schill, C., Schlüter, M., et al. sustainable consumption: The use of descriptive norms to promote a minority (2016). Social norms as solutions. Science, 354(6308), 42–43 (New York, N.Y.). behavior in a realistic online shopping environment. Journal of Environmental Ölander, F., & Thøgersen, J. (2014). Informing versus nudging in environmental policy. Psychology, 43, 166–174. Journal of Consumer Policy, 37(3), 341–356. Doherty, K. L., & Webler, T. N. (2016). Social norms and efficacy beliefs drive the Oreg, S., & Katz-Gerro, T. (2006). Predicting proenvironmental behavior cross- Alarmed segment’s public-sphere climate actions. Nature Climate Change 2016 6:9, 6 nationally: Values, the theory of planned behavior, and value-belief-norm theory. (9), 879–884. Environment and Behavior, 38(4), 462–483. D’Souza, C., & Taghian, M. (2005). Green advertising effects on attitude and choice of Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease advertising themes. Asia Pacific Journal of Marketing and Logistics, 17(3), 51–66. persuasion by enhancing message-relevant cognitive responses. Journal of Personality Ekelund, R. B., Mixon, F. G., & Ressler, R. W. (1995). Advertising and information: An and Social Psychology, 37(10), 1915. empirical study of research- Google Académico. Journal of Economic Studies, 22(2), Pickett-Baker, J., & Ozaki, R. (2008). Pro-environmental products: Marketing influence 33–43. on consumer purchase decision. Journal of Consumer Marketing. Evans, J. S. B. T., & Stanovich, K. E. (2013). Dual-process theories of higher cognition. Plassmann, H., Ramsøy, T. Z., & Milosavljevic, M. (2012). Branding the brain: A critical Perspectives on Psychological Science, 8(3), 223–241. review and outlook. Journal of Consumer Psychology, 22(1), 18–36. Farrow, K., Grolleau, G., & Ibanez, L. (2017). Social norms and pro-environmental Purohit, H. C. (2012). Product positioning and consumer attitude towards eco-friendly behavior: A review of the evidence. In Ecological Economics, 140, 1–13. labeling and advertisement: An analytical study. Journal of Management Research, 12 Florack, A., & Scarabis, M. (2006). How advertising claims affect brand preferences and (3), 153. category–brand associations: The role of regulatory fit. Psychology & Marketing, 23 Rhodes, N., Shulman, H. C., & McClaran, N. (2020). Changing norms: A meta-analytic (9), 741–755. integration of research on social norms appeals. Human Communication Research, 46 Frazier, P. A., Tix, A. P., & Barron, K. E. (2004). Testing moderator and mediator effects (2–3), 161–191. in counseling psychology research. Journal of Counseling Psychology, 51(1), 115–134. Richter, I., Thøgersen, J., & Klöckner, C. A. (2018). A social norms intervention going Garbarino, E. C., & Edell, J. A. (1997). Cognitive effort, affect, and choice. Journal of wrong: Boomerang effects from descriptive norms information. Sustainability, 10(8), Consumer Research, 24(2), 147–158. 2848. Gershoff, A. D., & Frels, J. K. (2015). What makes it green? The role of centrality of green Rossiter, J. R., & Percy, L. (1980). Attitude change through visual imagery in advertising. attributes in evaluations of the greenness of products. Journal of Marketing, 79(1), Journal of Advertising, 9(2), 10–16. 97–110. Schmuck, D., Matthes, J., & Naderer, B. (2018). Misleading consumers with green Goldstein, N. J., Cialdini, R. B., & Griskevicius, V. (2008). A room with a viewpoint: advertising? An affect–reason–involvement account of greenwashing effects in Using social norms to motivate environmental conservation in hotels. Journal of environmental advertising. Journal of Advertising, 47(2), 127–145. Consumer Research, 35(3), 472–482. Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2007). Hansen, J., & Wänke, M. (2009). Liking what’s familiar: The importance of unconscious The constructive, destructive, and reconstructive power of social norms. familiarity in the mere-exposure effect. Social Cognition, 27(2), 161–182. Psychological Science, 18(5), 429–434. Hartmann, P., & Apaolaza-Ibáñez, V. (2015). Green advertising revisited. International Schultz, P. W., Tabanico, J. J., & Rendón, T. (2008). Normative beliefs as agents of Journal of Advertising, 28(4). influence: Basic processes and real-world applications (pp. 385–409). Attitudes and Henion, K. E., & Kinnear, T. C. (1976). A guide to ecological marketing. Ecological Attitude Change. Marketing. Columbus, Ohio: American Marketing Association. 10 J. Castro-Santa et al. Journal of Behavioral and Experimental Economics 102 (2023) 101956 Schwartz, S. H. (1973). Normative explanations of helping behavior: A critique, Thorson, E., Page, T., & Moore, J. (1995). Consumer response to four categories proposal, and empirical test. Journal of Experimental Social Psychology, 9(4), of’’green’’television commercials. Advances in Consumer Research, xxii(22), 349–364. 243–250. Sörqvist, P., Haga, A., Langeborg, L., Holmgren, M., Wallinder, M., Nöstl, A., et al. Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and (2015). The green halo: Mechanisms and limits of the eco-label effect. Food Quality Social Psychology, 9(2p2), 1. and Preference, 43, 1–9. Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. Current Directions in Sparkman, G., & Walton, G. M. (2017). Dynamic norms promote sustainable behavior, Psychological Science, 10(6), 224–228. even if it is counternormative. Psychological Science, 28(11), 1663–1674. Zizzo, D. J. (2010). Experimenter demand effects in economic experiments. Experimental Staats, A. W., & Staats, C. K. (1958). Attitudes established by classical conditioning. The Economics, 13(1), 75–98. Journal of Abnormal and Social Psychology, 57(1), 37. 11