THE IMPACT OF GROUP UNANIMITY ON THE EXPRESSION OF MORAL BEHAVIOR

HOW TO ENCOURAGE LATENT DO-GOODERS TO ACT ON THEIR

MORAL INCLINATIONS

THE IMPACT OF GROUP UNANIMITY ON THE EXPRESSION OF MORAL BEHAVIOR

by

SANDER HOLWEG

20

of June, 2016

HOW TO ENCOURAGE LATENT DO-GOODERS TO ACT ON THEIR

MORAL INCLINATIONS

THE IMPACT OF GROUP UNANIMITY ON THE EXPRESSION OF MORAL BEHAVIOR

by

SANDER HOLWEG

Master Thesis 20^th of June, 2016

University of Groningen Faculty of Economics and Business

Department of Marketing MSc Marketing Management

Author: 1^stsupervisor: 2^nd supervisor:

Sander Holweg dr. J.W. Bolderdijk prof. dr. ir. K. van Ittersum s.r.c.holweg@student.rug.nl j.w.bolderdijk@rug.nl k.van.ittersum.rug.nl

Ambonstraat 7B Nettelbosje 2 Nettelbosje 2

9715HA Groningen 9747 AE, Groningen 9747 AE, Groningen

(+31) 64 268 6224 (+31) 50 363 9086 (+31) 50 363 66

student number 2805715

I

Abstract

Building upon research on the moral do-gooder derogation phenomenon, I propose in this paper that a unanimous opposition hinders latent do-gooders to express their intentions, particularly when their behavior carries a moral charge.

In an experiment, based on the conformity experime nts of Asch in the early 1950’s, 102 vegetarians and vegans were offered the opportunity to sign a petition that was aligned with their interests – getting more sustainable alternatives into the supermarket. Prior to signing themselves, the petition was declined by fellow participants (in actuality instructed confederates). In half of the situations however, the group unanimity was broken by having one confederate (the ‘dissenter’) endorsing the petition. This was supposed to create a

‘liberating effect’ for the vegetarian/vegans participants to show their sincere moral intentions.

The petition was manipulated in such a way that there were two versions: a morally framed and a selfishly framed petition. This created a two by two between-subjects experimental design.

The results of this design showed significant positive effects for the influence of the dissenter on deviation from group norms; the percentage of vegetarians/vegans signing the petition increased from 52.3% to 84.4% when a dissenter was present. Even though the experiment did not provide statistical evidence for the other hypothesized relationships, an analysis of the verbal comments provided some remarkable insights: vegetarians tried to minimize the resentment by the group majority by downplaying or hiding their moral reasons for not eating meat.

Keywords: Moral do-gooder; Unanimity; Group pressure; Anticipated resentment;

Undercover altruism; Vegetarians; Sustainable consumption

II

Preface

In front of you lies my master thesis as a completion of the Master of Science program in Marketing Management offered by the University of Groningen.

The motivation for the subject of this paper stems from my interest in consumer psychology and consumer wellbeing. My personal objective for writing this thesis was to extend my knowledge in marketing and develop a better understanding of the motives and cognitive capacities that determine consumers’ (sustainable) behavior. The process of writing this thesis helped me to improve my writing and research skills. Moreover, it confirmed my belief that I want to pursue a career as marketer, specializing in consumer behavior.

The thesis is written independently, but not without the support of others. I would like to express my profound gratitude to the people who helped me during the course of this study.

First of all, I would like to thank my supervisor dr. J.W. Bolderdijk for guiding me throughout the whole process. He was always available whenever I had a question or ran into a problem.

He consistently allowed this thesis to be my own work, but provided me with grounded feedback when necessary. I would also like to acknowledge prof. dr. ir. K. van Ittersum as my second supervisor, and I am thankful for the time and effort he spent in thoroughly reviewing my thesis. I would also like to thank my fellow master student Julia Storch with whom I worked alongside during the data collection. We combined the experiments of both our theses, and together we recruited the participants for our studies. Without her passionate participation and input, the execution of the experiment would not have been so successful as it was.

Furthermore, I want to thank the confederates of the experiment, Jasper Hidding, Ingeborg Zimmer and Tjeerd Havinga for their time and dedication. My gratitude also goes to prof. dr.

J.C.J. Hoeks for providing the rooms and the necessary materials that were needed to carry out the experiment. Finally, an honorable mention goes out to the world famous psychologist S.E. Asch, who masterminded the conformity experiments of which this research is based upon.

Groningen, 20

of June, 2016

Sander Holweg

III

Table of contents

Abstract ... I Preface ... II

1. Introduction ... 1

2. Theoretical framework ... 3

2.1 Conforming to group behavior ... 3

2.2 Group unanimity ... 4

2.3 Morality in public settings ... 5

2.4 Group unanimity in moral settings ... 7

2.5 Conceptual model ... 7

3. Research design ... 9

3.1 Operationalization of the latent do-gooder ... 9

3.2 Participants & data collection ... 10

3.3 Operationalization of the group setting ... 11

3.4 Manipulation of the research variables ... 12

3.5 Experimental procedure ... 14

4. Results & discussion ... 16

4.1 Results of the unanimity effect ... 16

4.2 Results of the morality effect ... 17

4.3 Results of the interaction effect ... 17

4.4 Results of the group discussion analysis ... 18

5. General discussion ... 20

6. Limitations & future research ... 22

7. Managerial implications ... 24

Literature ... 25

Appendices ... 28

List of appendices ... 28

1

1. Introduction

Sustainability is becoming increasingly important. People all over the world are thinking about alternative ways to generate energy, to reduce carbon emissions, and to keep the ecosystems of our planet in balance. Also the food industry is coming up with more sustainable products, like tofu burgers and insects.

That marketing can play an important role in encouraging more sustainable behavior and consumption is highlighted by several researchers, such as Peattie and Peattie (2009), Jones, Clarke-Hill and Hillier (2007), and Sheth and Parvatiyar (1995). But so far the performance of marketing in this matter is disappointing (Peattie & Crane, 2005; Pickett-Baker, 2008). One novel explanation for the slow uptake of sustainable products may be that people feel that buying sustainable products would make them unpopular among their peers.

This notion is based o n the ‘moral do-gooder derogation’ theory, which explains why people put down morally-acting individuals (Minson & Monin, 2012): people try to preserve their self-image as moral human beings. They do this, for example, by derogating others who challenge their moral self-concept. That self-concept is challenged by moral do-gooders, because they chose a moral path that other people could have chosen as well but did not (Bolderdijk, Brouwer, & Cornelissen, 2016). So buying sustainable products, where others chose not to, might lead to resentment by other people.

I argue that people anticipate this resentment, and thus strategically avoid acting morally in social situations where others did not choose morally. Thus, the presence of a non-moral majority creates group pressure to not act morally as well. One of the most famous studies that shows how people’s behavior is affected by others, is the conformity study by Asch (1951). He uncovered how big of an impact group pressure has on people. In this paper I propose that group pressure is particularly high in moral situations (due to the accompanied anticipated resentment); people do not want to be seen as the moral outlier of the group.

Hence, there is a high threshold to show moral intentions under group pressure.

I argue that this threshold can be lowered by a break in the unanimity of a group. A break can

be created by adding one dissenter into the group – someone who deviates from the group

norm. Similar to Asch’s experiments, the dissenter breaks the group unanimity and thus can

serve as a ‘liberator’ for people to express their sincere personal inclinations. In this research,

2

the presence of a dissenter is expected to work especially liberating when moral arguments are in question, because group pressure is higher in those morally-charged situations.

So to find out more on how to encourage sustainable behavior in the most effective way, it is important to understand what hinders latent do-gooders

to publically act on their moral intentions. Gaining insights into the differential impact of group unanimity on the behavior of these latent do-gooders can be of particular importance for marketers and companies that produce sustainable products.

I aim to discover if moral do-gooder aversion can help to explain why people do not want to choose sustainable products in public settings. This paper intends to fill a part of the void in research on moral do-gooder derogation and undercover altruism. This is done by providing insights on how moral do-gooders behave when being confronted with a unanimous opposition under different circumstances regarding the degree of morality.

The following research question is central to the empirical research of this study: To what extent does group unanimity hinder latent do-gooders to express their moral inclinations, therewith deviating from group norms, and how is this affected by the degree of morality of the public setting?

1 Do-gooders who are morally motivated, but did not express themselves publically as such.

3

2. Theoretical Framework

2.1. Conforming to Group Behavior

It is known that people are aware of the social consequences that may follow from their actions (Holmes, Miller & Lerner, 2002). Linking this to the actor’s perspective of the moral do-gooder derogation phenomenon – it has been found that prospective do-gooders may strategically regulate their behavior to avoid resentment by others (Cornelissen & Bolderdijk, 2013). This tendency to avoid resentment by others can be explained by the social needs of humans. According to Maslow (1943), humans need to feel a sense of belonging and acceptance among social groups. One way to achieve this is by conforming to the behavior of the group.

One of the most famous experiments on conformity is the line judgment task by Asch from the 1950’s. This task had one clearly correct answer – participants were supposed to pick out the shortest of three lines, the shortest one was easy to see. So if the participant gave an incorrect answer it was obvious that this was due to the social pressure of the confederates, rather than for example visual impairment (Asch, 1955). The results of the experiment were that about one third of the participants conformed to the obvious wrong answer given by the confederates. Without this social pressure from the confederates, not even one percent of the participants answered incorrectly.

The research by Asch concluded that there are two main motives for conformity:

informational social influence and normative social influence. Informational influence occurs

when someone accepts information from others as evidence about reality – believing they are

better informed (Deutsch & Gerard, 1955). Normative influence is defined by Aronson,

Wilson and Akert (2005) as the influence of other people that lead us to conform in order to

be liked and accepted by them. In the case of the experiment of Asch it was unlikely that the

participants actually saw it wrongly (because of the evident wrong answer). The conformity

that took place was therefore a consequence of normative social influence rather than

informational social influence. Thus, in order to be liked and to be accepted by a group,

people conform to the behavior of their peers through normative influence (Aronson et al.,

2005). This normative influence is applicable to the actor’s perspective of moral do-gooder

derogation, because the actors tend to change their moral behavior or reasoning to avoid

4

resentment and to be accepted by others. Bolderdijk and Cornelissen (2015) showed that people sometimes want to appear less moral than they actually are. They reduce their observable levels of morality and increase the level of conformity to avoid socially awkward situations.

Over time, researchers have established a wide range of factors that affect conformity, including group size (Asch, 1956; Wilder, 1977), fear (Darley, 1966), group cohesiveness (Back, 1951), and the in-group status (Berkowitz & Macaulay, 1961).

2.2. Group Unanimity

Another factor that affects conformity is unanimity (Asch, 1956). When group members are unanimous in their support of a norm, an individual feels more pressure to follow that same norm. However, according to Aronson, Wilson and Akert (2007), even a small break in unanimity by a dissenter can lead to a decrease in the power of normative influence and lowers the need to conform to the majority of the group. Hence, a dissenter has a ‘liberating’

effect on people to show their sincere personal inclinations. The presence of a dissenter therefore also increases the likelihood that someone deviates from group norms, because the amount of peer pressure decreases.

In a variation of the experiment with the line judgment task by Asch (1955) a dissenter was inserted into the group to break the group unanimity. The presence of the dissenter as a minority member gave the participants the confidence to exert their independence to a larger extent. Asch (1956) showed that the presence of just one dissenter in the group could reduce conformity from 32% to 5%. In 1968, these results were replicated and extended in another version of the line judgment experiment by Allen and Levine. They introduced a dissenter who was wearing thick-rimmed glasses, suggesting he was visually impaired. So the dissenter was not a very credible source when it comes to judging line lengths. However, the study of Allen and Levine (1968) revealed that this did not matter – the participants still conformed less to the majority of the group just because there was a dissenter present, no matter how credible the dissenter was. The presence of one dissenter gave the participants the confidence to deviate from the norms of the group.

The importance of group unanimity in relation to other conformity affecting factors, like

group size, is displayed by several researchers. For example, Myers (2013) showed, in terms

5

of social influence, that a large group with a non-unanimous majority is usually less successful in eliciting conformity than a small unanimous group. This affirms the importance of group unanimity in conforming behavior and in the confidence to exert one ’s independence.

In this study, moral do-gooders (e.g. vegetarians), like the participants in Asch’s experiment, often face an opposing majority (e.g. meat eaters are dominant in the case of vegetarians).

This creates group pressure that may prevent moral do-gooders to act on their personal inclinations, similar to the Asch experiments. In this research, the unanimity of a group is an explanatory variable in whether someone deviates from the group norm or not. While in the experiments of Asch ‘deviating from group norms’ meant that the participant gave the actual right answer, it is manifested in this study by whether someone expresses their private moral inclination while being confronted with an opposing majority. This explanatory variable contains two item levels, namely unanimity and non-unanimity. Non-unanimity is achieved by placing a dissenter into the group. This dissenter is hypothesized to ‘liberate’ the moral do- gooders from the need to conform, making it therewith easier for them to deviate from the behavior of the majority of the group. In other words, the absence of group unanimity lowers overall conformity as the moral do-gooder feels less need for social approval from the group.

Consequently, he or she is more likely to deviate from group norms in this situation. Based on the previous reasoning the following hypothesis is formulated:

H1: Moral do-gooders will feel more liberated, and thus deviate more often from group norms, when confronted with a non-unanimous group than moral do-gooders who are confronted with a unanimous group.

2.3. Morality in Public Settings

Another factor that can affect the tendency to conform to the group norm in public settings is

the degree of morality in that particular setting. To stimulate more sustainable consumption,

marketers often create a moral setting by framing sustainable products as the morally superior

option. For example, the main selling point of tofu burgers is the lower environmental impact

compared to burgers made from beef. Although some people assume that humans are

primarily motivated by economic self-interest, Bolderdijk, Steg, Geller, Lehman and Postmes

(2013) showed that people also care about maintaining a favorable view of themselves. They

want to maintain a ‘positive self-concept’. Jordan, Mullen and Murnighan (2011)

demonstrated something similar, namely people’s desire to maintain a moral self-image. So

6

the morally superior framing could be an effective marketing strategy because it creates a moral appeal, and acting on such an appeal satisfies people’s desire to maintain a moral self- image.

There are also alternative ways of framing sustainable products, apart from moral framing, as was shown by Thøgersen (2011). He found that consumers buy ‘green’ products, such as organic foods for unselfish/moral reasons (i.e. the common good), but also for selfish reasons (e.g. status, pleasure, health). Some of the alternative ways of sustainable product framing might pose less of a threat to bystanders. Cramwinckel, Van Dijk, Scheepers and Van den Bos (2013) illustrated, for example, that moral arguments pose a bigger threat to bystanders than non-moral arguments. Participants were asked to taste meat and were afterwards confronted with other participants who refused to taste meat. Cramwinckel et al. (2013) showed that people who were asked to taste meat were more likely to dislike someone who refused to do so, if that refusal was for moral reasons rather than non-moral reasons. Participants (i.e. non- refusers) disliked the moral refuser more than the non-moral refuser, because the moral refuser posed a threat to their moral self-image. The reasoning of the non-moral refuser is different because it is based on personal preferences, whereas moral statements tend to apply to everyone. Täuber, van Zomeren and Kutlaca (2015) also suggested how non-moral framing of persuasive messages could be used as a means to avoid defensive responses.

The second explanatory variable of this study is the degree of morality in the public setting.

The two types of settings that are utilized during the experiment are accomplished by framing

the behavior in question as a moral or selfish decision. Moral framing is expected to hinder

moral do-gooders more in showing their sincere moral inclinations, compared to selfish

framing. I expect this because selfish arguments are personal and therewith do not implicate

anything about the morality of others, unlike moral arguments that tend to apply to everyone

and therewith judge bystanders who do not act on the moral appeal. Thus, I argue that in the

moral setting people are more inclined to conform to the majority of a group than in the

selfish setting, due to higher anticipated resentment. In comparison, in the selfish setting

people are less inclined to conform to the majority of a group, and are therewith expected to

deviate more often from group norms than in the moral setting. Based on this reasoning the

following hypothesis is derived:

7

H2: Moral do-gooders will deviate more often from group norms in a selfishly framed setting than in a morally framed setting.

2.4. Group Unanimity in Moral Settings

Based on the paper by Bolderdijk and Cornelissen (2015) on undercover altruism; there are people who want to act in a moral manner, but at the same time do not want to be seen as a moral do-gooder. They will be somewhat afraid to deviate from the unanimous norms of a group based on moral grounds. Thus, the presence of one dissenter could have a positive effect in taking away some of that fear, which makes it easier for moral do-gooders to show their sincere moral inclinations. This effect is especially present in a morally framed setting, because the fear of being resented by others is expected to be higher when moral arguments are in question, in comparison to selfish arguments. Hence, when a setting is framed in a moral way, the inclinations to conform to the behavior of a unanimous group will be stronger than when a setting is framed in a selfish way. The presence of a dissenter would therefore have a more liberating effect when group pressure is high, which is the case with moral framing.

Consequently, the degree of morality in public settings also serves as a moderator in the relationship between group unanimity and deviation from group norms. The former reasoning means for the interaction effect that non-unanimity will likely lead to a more liberating feeling for moral do-gooders to show their sincere moral behavior, when being confronted with moral framing, in comparison to selfish framing. This greater feeling of liberation will subsequently lead to more deviation from group norms in the particular situation. In more abstract terms, the increase in the amount of moral do-gooders that deviates from group norms when confronted with a non-unanimous group (with one dissenter present), in comparison to a unanimous group, will be larger with moral framing than with selfish framing. The following hypothesis is formulated based on the theoretical framework and the previously mentioned relationships between the research variables:

H3: One dissenter in a group, causing non-unanimity, has a greater liberating effect on deviation from group norms in morally framed settings than in selfishly framed settings.

2.5. Conceptual Model

8

The conceptual model in Figure 1 graphically displays the expected relationships between the explanatory variables and the dependent variable of this study. The model matches the hypotheses disclosed in the theoretical framework.

+--

–

+

–

H1

Figure 1. Moderated impact of the degree of morality in public settings on the relationship between group unanimity and deviation from group norms.

Group unanimity (unanimous vs.

non-unanimous)

Degree of morality in public settings (moral vs. selfish)

Deviation from

group norms

(signing vs. not signing)

9

3. Research Design

3.1. Operationalization of the Latent Do-gooder

The previously mentioned hypotheses are tested in the context of vegetarians as being the latent do-gooders. Vegetarians can be seen as a good example of moral outliers because they are part of a minority and a lot of vegetarians adopted a vegetarian diet out of moral reasoning.

These moral motives are illustrated by a study on vegetarianism in America by Vegetarian Times (2008). The study indicates that 47 percent of vegetarians eat vegetarian because of environmental concerns. Another moral reason for vegetarians to abstain from eating meat is animal welfare with 54 percent (note: one person can have multiple reasons for being vegetarian). Another study, by Fox and Ward (2008), also explored the motivations of vegetarians. They found that health and the ethical treatment of animals were the main motivators for vegetarianism. The common denominator in the motivations for sustainable consumption in these two studies is the degree of morality. ‘Health’ can be seen as selfish reasoning (low degree of morality) because it involves personal preferences/benefits, whereas

‘ethical treatment of animals’ can be seen as moral reasoning (high degree of morality) because it involves the common good. Health framing (selfish values) and ethical framing (moral/altruistic values) can therefore be seen as antonyms of each other.

Vegetarians also often face an opposing majority in the form of meat eaters. Minson and Monin (2012) showed how meat eaters demonstrate negative responses towards vegetarians.

Meat eaters used negative wordings to describe vegetarians because they believed that vegetarians feel morally superior. In this case vegetarians are moral outliers who pose a threat to the moral self-concept of people who do eat meat. Meat eaters therefore engage in defensive behavior, manifesting in associating vegetarians with negative wordings. The defensive behavior by meat eaters can obstruct vegetarians to express their moral inclinations in public settings, because they might already anticipate the resentment by meat eaters. This was confirmed by Greenebaum (2012) who demonstrated

how vegetarians and vegans anticipate resentment from omnivores. They engaged in ‘face-saving’ techniques, such as

2 Using a qualitative study to explain how the participants engage in promoting vegetarianism and veganism to omnivores in daily interactions.

10

avoiding confrontation or focusing on health benefits, to manage the tension that arises when being confronted with omnivores.

3.2. Participants & Data Collection

An experiment took place in Groningen, The Netherlands, from the 18

of April until the 3

of May 2016. To gather the required data for this experiment, 102 vegetarians and vegans were recruited via Facebook groups, flyers and vegetarian/vegan restaurants. Vegans were also included in the study to broaden the reach for recruitment

. The participants were randomly assigned to one of the four experimental situations of Table 1 and received financial compensation for their participation.

The experiment was partly combined with a study of another researcher who asked the participants to fill out an online questionnaire (see Appendix A) prior to the experiment.

From the 102 participants that took part in the study, the first nine participants were excluded because they served as a pilot test for the final experiment. Another participant was excluded because she did not pass the screening question in the questionnaire (she indicated that she was not vegetarian). Three other participants were excluded because they correctly guessed the purpose of the study – they figured out the objectives of the study as indicated in the comment section of the preceding questionnaire

. Due to a programming error in the preceding questionnaire, five participants were exposed to two manipulations instead of one.

This could have made them suspicious about the actual objectives of the experiment. That is why extra analyses were performed, with and without these participants, to check if the results were significantly different. This was not the case, so the five participants were included in the study. Altogether, 89 participants were included in the final analyses.

In total, 83.1% of these 89 participants were female (and 16.9% were male). The study consisted for 66.3% of vegetarians and for 33.7% of vegans. Most of the participants were students (85.4%) and in the age category of 18 to 24 years (73.0%). The nationalities of the participants were mostly German (37.1 %) and Dutch (28.1 %)

.

3For reasons of convenience, ‘vegetarians and vegans’ will be referred to as ‘vegetarians’ in the rest of this paper.

4Two participants understood that the study was about anticipated resentment, and another participant suspected that the other discussants were confederates.

5The nationalities of the remaining participants (28.9 %) existed of a wide range of other nationalities.

11

All the participants, who signed up for the debriefing at the end of the questionnaire, were debriefed on the actual objectives of the experiment one month after the experiment had ended.

3.3. Operationalization of the Group Setting

Inspired by Asch’s conformity experiments, the idea was to expose participants to a situation where the majority of the group makes a different choice than the participant would make in a private situation.

Vegetarians were asked to fill out a questionnaire, as part of another study, which would be followed up by a group discussion. The participants did not know that only vegetarians were participating in the experiment. The questionnaire implied that people with other eating habits were present as well

, creating a more hostile environment. This made it more likely for the vegetarians to anticipate possible resentment from meat eating discussion members (in actuality instructed confederates). This resentment is important for the study. Without any possible resentment, the moral do-gooder would not feel any pressure to conform to the behavior of the group. The experimental procedure will be elaborated upon in paragraph 3.5.

The discussion groups constantly consisted out of four confederates (including the researcher) and one actual participant. These numbers were based on the optimum conformity effects

established by Asch (1951), Gerard, Wilhelmy and Conolley (1968), and Barnard (1991).

This way the majority of the group would always consist out of (at least) three persons. In order to neutralize other normative influences, such as group cohesiveness, confederate participants were instructed to act as if they did not know each other. To prevent possible confounds that could stem from differences in the personalities of the confederates, the same confederates were used (in the same roles) throughout the entire study

.

6An exemplary sentence from the questionnaire that implied that participants with other eating habits were present as well: “If you are a vegetarian, please continue to question five”.

7 Group size influences conformity in a way that the bigger the majority, the more people conform, but only up to a certain point. The optimum conformity effects were found by Asch (1956) at a majority of three persons.

Increasing the majority beyond three did not significantly increase conformity. Researchers do not all agree on the exact number for this optimum point. Taking into account several studies, the most robust finding is that conformity reaches its full extent at a 3-6 person majority (Asch, 1951; Gerard et al., 1968; Barnard, 1991).

8In 9 out of the 89 discussions, participants encountered a slightly different group composition. In these 9 cases, confederate 3 was replaced by another confederate. The results were not significantly affected by this, and the gender distribution among the confederates was kept equal during the entire study. The group constantly existed out of two male and two female confederates.

12

3.4. Manipulation of the Research Variables

The experiment had a two by two between-subjects design, with categorical data types, creating four experimental situations by varying the framing of the petitions and the unanimity of the group’s decisions (Table 1).

Table 1

Overview experimental situations

Situation Group unanimity Public setting Number of participants

1 Unanimous Selfish 22

2 Unanimous Moral 22

3 Non-unanimous Selfish 22

4 Non-unanimous Moral 23

The explanatory variable, the degree of morality in public settings, was manipulated by offering participants the opportunity to sign a petition that was aligned with their interest – getting more sustainable alternatives into the supermarket. Following Asch’s approach, where the correct answer was obvious to participants, the petitions were set up in a way that vegetarians were motivated to sign it and had no real incentive to decline signing it.

Consequently, because the confederates were instructed to constantly not sign the petition, it could be attributed to peer pressure when participants did not sign the petition as well.

The degree of morality was manipulated in such a way that there were in fact two different petitions – a morally framed petition and a selfishly framed petition

(see Figures 2 & 3, and Appendix B for enlarged versions). Half of participants were offered to sign the ‘moral petition’, the other half were offered to sign the ‘selfish petition’. Choices for these framing types were based on the previously mentioned studies of Thøgersen (2011), Fox and Ward (2008) and Greenebaum (2012).

9The petition versions were randomly assigned to the participants.

13

A short pilot study was held on the authenticity of the two petitions. Ten randomly selected people were asked to take a look at the petition (five per framing)

. All ten people thought the petitions were highly authentic. Although they were not vegetarian, six people still claimed that they would have signed the petition because they did not see any harm in signing it. It was therefore assumed that the petitions were set up properly and that a big majority, of vegetarians in particular, would sign these petitions in a private setting. This way, when participants did not sign the petition, it would be seen as unusual behavior due to group pressure.

Figure 2. The morally framed petition. Figure 3. The selfishly framed petition.

The explanatory variable group unanimity was also manipulated. Either the group was unanimous in its decision not to sign the petition or the group contained one dissenter who did sign the petition and therewith broke the group unanimity. The researcher in the experiment was used to operationalize the non-unanimity condition by serving as the dissenter; she handed out the petition without an endorsement: “... but don’t feel obliged to sign it, I didn’t

10 They were told afterwards that the petitions were fake.

14

sign it as well.” (unanimous condition), or with an endorsement: “... I signed it, but don’t feel obliged to sign it as well.” (non-unanimous condition).

3.5. Experimental Procedure

When a participant was done filling out the questionnaire, the researcher told that participant that there was one spot left at a group discussion in another room. The researcher then took the participant to the discussion room where he/she joined three confederates of whom the participant believed were participants as well.

The researcher made the participant sit down at the end of the row and thanked the participants for coming (see Figure 4 for the experimental setting)

. Before the discussion actually started, the researcher presented either of the two petitions (moral or selfish) to the group, and said: “Before we start the discussion, I would like to ask you to take a look at this petition. It’s a petition of a friend of mine who wants to get more vegetarian and vegan alternatives in the assortment of supermarkets. Here, have a look.”. This introduction was adapted to the experimental situation in question (Table 1).

Figure 4. The experimental setting of the group discussion. From left to right: researcher, confederate 1, confederate 2, confederate 3, participant (white shirt).

11 She points out the presence of a vegan sample product (vegan chorizo) for the group to try if they want.

Confederate number mentions that he tasted the product while he was waiting, but that he still prefers the real chorizo. This implies that he is a meat eater which creates extra anticipated resentment.

15

The framings of the different petitions were emphasized upon by the researcher who mentioned the motive for the particular petition. In the moral situation the researcher added to her introduction: “It’s a petition of a friend of mine who wants to get more vegetarian and vegan alternatives in the assortment of supermarkets, because he thinks it’s the ethically right thing to do regarding animal rights and environmental issues .” In the selfish situation the researcher added to her introduction: “It’s a petition of a friend of mine who wants to get more vegetarian and vegan alternatives in the assortment of supermarkets, because he believes plant-based diets are healthier and it reduces risks on diseases .”

After the researcher introduced the particular petition, she handed it over to confederate number 1, who passed it on to number 2, 3, who subsequently passed it on to the participant.

All three confederate participants refused to sign the petition in every situation, creating a situation where the participants faced an opposing majority. Whether the participant decided to sign the petition or not in the given situation was the crucial question. Not signing the petition was seen as a sign of conformity to the majority of the group, and signing the petition was seen as deviating from the group norm. This is how the dependent variable, deviation from group norms, was measured. The participant’s choice was monitored and documented by confederate 1 on a computer directly after the participant left the discussion room

.

12 The experimental procedure is described in full detail in Appendix C.

16

4. Results & Discussion

Before testing the hypotheses it was checked if the manipulations of the preceding questionnaire significantly influenced the outcomes of the experiments. This was not the case

. Because participants were randomly assigned to the conditions of the questionnaire, and there was no significant correlation with (not) signing the petition, the questionnaire’s manipulation has been disregarded for the rest of the analyses. All analyses were performed using a 95% confidence level.

4.1. Results of the Unanimity Effect

I hypothesized (H1) that the participants would sign the petitions more often (therewith deviating from group norms) when confronted with a non-unanimous group than with a unanimous group. Table 2 illustrates that group unanimity and deviation from group norms are related in the expected direction. The percentage of vegetarians signing the petition increased from 52.3% to 84.4% after inserting one dissenter into the group (X

(2, N = 89) = 10.679, p = .001). In other words, deviation from group norms seemed to increase significantly after inserting a dissenter. These results are in line with hypothesis 1.

Table 2

Cross tabulation: Group unanimity vs. Signing petition

Group unanimity

Total Unanimous Non-unanimous

Signing petition

No Count

% within Signing petition

21 47.7%

7 15.6%

28 31.5%

Yes Count

% within Signing petition

23 52.3%

38 84.4%

61 68.5%

Total Count

% within Signing petition

44 100.0%

45 100.0%

89 100.0%

13 To check whether the questionnaire’s (public/private) conditions would influence the results of this study, there has been controlled for these conditions. This was done via a Chi-square test and a correlation test. The Chi-square test was done to check if the manipulation of the questionnaire was randomly assigned among the manipulations of the experiment. This was the case, because no significant relationships could be found between the public/private condition and group unanimity or public settings; X² (2, N = 89) = .280, p = .597.

14 A bivariate correlation test was executed to check whether there was a significant correlation between the questionnaire’s manipulation and the dependent variable of this study. The test turned out to have a non- significant correlation of r(89) = -.041, p = .704.

17

4.2. Results of the Morality Effect

The percentage of people signing the petition in the selfish setting was lower (59.1%) than in the moral setting (77.8%) (

(2, N = 89) = 3.603, p = .058). Additionally, people seemed to conform more often to group norms in the selfish setting (40.9%) than in the moral setting (22.2%). According to the results of the Chi-square test of a contingency table, the variables

‘degree of morality in public setting’ and ‘deviation from group norms’ were not significantly related. These results are not in line with hypothesis 2. Moreover, the analysis tends to show signs of a relationship in the opposite direction of what hypothesis 2 initially proposed.

Table 3

Cross tabulation: Public setting vs. Signing petition

Public setting

Total Selfish Moral

Signing petition

No Count

% within Signing petition

18 40.9%

10 22.2%

28 31.5%

Yes Count

% within Signing petition

26 59.1%

35 77.8%

61 68.5%

Total

Count

% within Signing petition

44 100.0%

45 100.0%

89 100.0%

4.3. Results of the Interaction Effect

For the interaction between group unanimity and the degree of morality in the public setting, I hypothesized (H3) that the dissenter would have a greater liberating effect on deviation from group norms in morally framed settings than in selfishly framed settings. I used logistic regression, with ‘Unanimity’, ‘Framing’ and ‘Framing by Unanimity’ as predictors of

‘Signing’, to test this (see Appendix E for the elaborated SPSS output).

Figure 5 visualizes the extent to which participants deviated from the group norm by signing the accompanying petition in the given experimental situation

. Looking at Figure 5, the dissenter had indeed a liberating effect on deviation from group norms in the moral setting, as

15The complete overview of the data of the experiment is presented in table 4 of Appendix D, including the participants who declined signing a petition.

18

well as in the selfish setting (as was tested significantly in paragraph 4.1). Unlike expected, the liberating effect of the dissenter seemed to be higher in the selfish setting compared to the moral setting, but this difference was not statistically significant – the logistic regression analysis revealed a non-significant interaction effect between the two explanatory variables; β

= -.200, Wald X

(1) = 0.034, p = .854.

Figure 5. Deviation from group norms per experimental situation.

4.4. Results of the group discussion analysis

An extra analysis was done on the comments that the participants made during the discussions (Appendix F). The results of this analysis showed that there were signs of participants strategically altering their motivations for becoming vegetarian in order to appear less moral.

For example, seven participants expressed downplaying behavior by stating that they were not fanatic in eating vegetarian. Five out of those seven participants seemed to derogate vegans, for instance by saying: "But I am not vegan or something, that is too much for me, I am just a vegetarian". Another interesting finding was that only 8% of the participants mentioned moral motivations. This is surprising because statistics on vegetarian motivations say that more than half of vegetarians have at least one moral reason for eating vegetarian (Vegetarian Times, 2008). It was evident that the participants were not very keen on disclosing their moral motivations for becoming vegetarian, and that they were hiding the altruistic motivations

41%

64%

77%

91%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Selfish setting Moral setting

% people signing the petition

Unanimous Non-unanimous

19

regarding their vegetarian lifestyle. One participant for instance checked if anyone signed the (moral) petition by looking at the back of the petition. She subsequently said multiple times while signing: “I will help the poor guy”, hereby referring to the supposed friend of the researcher who drafted the petition and the fact that nobody signed. Later on, she turned to the fellow discussants in particular and repeated that she only signed to help out the researcher.

Obviously there is no hard evidence that this participant was lying, but circumstances made it

likely that she was hiding her sincere moral intentions to prevent getting resented by the group.

20

5. General Discussion

This paper introduced and tested a conceptual model that aimed to discover how moral do- gooders are hindered in publically acting on their moral inclinations. Central to this research is the impact of group pressure and how it inhibits moral do-gooders to deviate from group norms. With the use of a two by two between-subjects experimental design, four situations were created in which vegetarians were exposed to either a unanimous or a non-unanimous group in either a morally or a selfishly framed setting. The experiment sought to find out if the liberation effect (created by adding a dissenter into a unanimous group) gives latent do- gooders the confidence to act on their private beliefs, and it examined how much impact this would have in moral versus selfish settings.

In accordance with the initial expectations, there was a significant relationship between group unanimity and participant’s deviation from the group norm. Vegetarians deviated more often from group norms when confronted with a non-unanimous group (including one dissenter) than with a unanimous group. Thus, vegetarians seem less afraid to be seen as a moral outlier in a group when there is at least one other person who deviates from the group norm as well.

Because vegetarians are a sustainable minority they will relatively often be confronted with a hostile situation wherein the majority of the people eats meat. The presence of a dissenter can therefore be of great value to vegetarians, in such a way that it boosts their confidence and enables them to express their moral intentions.

Contrary to the initial expectations, the results did not indicate a significant relationship between the degree of morality in the public settings and deviation from group norms – vegetarians did not deviate more often from group norms when being confronted in a selfishly framed setting than in a morally framed setting. Although not significant, the results did show signs of a relationship in the opposite direction. Vegetarians deviated more often from group norms in the moral setting than in the selfish setting. An explanation can be derived from the more conventional line of thinking that latent do-gooders are likely to act on a moral appeal to maintain a moral self-image. It could be that this desire to maintain a moral self-image was stronger than the fear of being resented by the group.

However, the analysis of the group discussion made clear that the participants did fear

resentment. Several participants showed clear signs of downplaying their moral motivations

21

or hiding their altruistic behavior. It seems plausible that the moral appeal was too strong for most vegetarians to not act upon. This way they could maintain their moral self-image, but it increased resentment from the fellow discussants as well. Vegetarians therefore downplayed their moral behavior or hid their altruistic inclinations probably to weaken the increased resentment.

People signed the petition more often when there was a dissenter present. But this liberating

effect was not bigger in moral settings than in selfish settings. Other than expected, the

liberating effect seemed to be larger in the selfish setting than in the moral setting (although

not significant). The initial expectation that the liberating effect would be larger in moral

settings was based on the higher anticipated resentment in those settings. In the end it looks

like the participants did not really succumb to the group pressure in the moral setting as

expected. Instead they figured out a way to act on the moral appeal and at the same time

minimize the anticipated resentment from the group by downplaying moral behavior or hide

altruistic intentions during the discussion afterwards.

22

6. Limitations & Future Research

Some limitations have to be kept in mind when interpreting the findings of this paper, but new future research opportunities can be derived from these limitations as well. During the experiment the researcher functioned as the dissenter in the group. This choice was driven by practical considerations, because it implicated that one person less was needed to act as a confederate. This facilitated the gathering of the desired number of people to act as confederates in the experiment. However, this might have had a negative impact on the internal validity of experiment, because the researcher can also be seen as an authority figure during the group discussion. Because people tend to conform more to people with a high in- group status (Berkowitz & Macaulay, 1961), which is the case with the researcher, more conformity towards the researcher could be expected. The authoritative aspects were tried to be held to a minimum by assigning a young person to the role of the researcher, who communicated in an informal way with the participants, and wore casual clothing.

Nevertheless, it is possible that the results of this study, like the liberating effect, might have been slightly enhanced and/or distorted due to the authoritative aspects. Future research could copy the experimental design, but with an extra confederate added to the group who functions as the dissenter. This way the researcher can focus solely on his/her role as the discussion leader.

Another way to improve the internal validity of this study is by testing the petitions that were used more thoroughly on equal authenticity. Although a short pilot test on the authenticity of the petitions showed that both petitions were highly authentic, this pilot test did not check whether they were equally authentic or not. Furthermore, despite the authenticity, one of the petitions might still have been more convincing than the other one. In the end more people signed the morally framed petition, but this could have been the case because that petition was simply more convincing than the selfishly framed petition.

It needs to be noted that this research focused on vegetarians and vegans only as morally

motivated people and the discussion topic was meat substitutes. It is questionable whether the

same results will be generated when choosing other morally motivated people and/or other

types of sustainable products. The experimental design of this study lends itself well for

adapting to different target groups and different topics. This way the external validity could be

explored in future research. Moreover, it is also worthwhile to make use of qualitative data, as

23

illustrated by Greenebaum (2012). Complementing the quantitative results with the qualitative data, derived from the discussion, will provide a more comprehensive representation of the participants’ behavior.

Another idea for future research is investigating different options regarding the liberating effect of a dissenter. More specifically, whether a dissenter has to be physically present or not in order to have a liberating effect on a morally motivated person. It might be possible that a poster of a person (next to the product), endorsing a specific sustainable product, gives morally motivated persons the confidence to act on their moral inclinations.

Finally, as was concluded in the discussion section, it looked like the participants figured out

a way to act on the moral appeal and at the same time minimize the anticipated resentment

from the group by downplaying their moral behavior afterwards. It might be interesting for

future research to find out if this was really the reason for their downplaying behavior and if

this is a common course of events among moral do-gooders.

24

7. Managerial Implications

The findings of this paper offer some novel practical implications. The first one is of experimental nature and builds upon the future research idea of the non-physical presence of the dissenter. The idea is to prevent morally motivated customers from ending up in a situation with a (potential) hostile opposition. One way of preventing this could be by hanging up posters next to the sustainable products with people who endorse that particular product. It could work liberating if someone tells you that he or she uses the product on a regular basis, preferably someone with a high ‘in-group’ status – an affiliated authority figure or celebrity.

The main finding of this paper is that morally acting people are undoubtedly motivated to act

on their moral inclinations, but at the same time are looking for ways to minimize anticipated

resentment from bystanders. If these ways cannot be found, it is plausible that they will

abstain from acting on their moral inclinations after all (e.g. buying a sustainable product). It

is therefore important for marketers to p rovide these people with ‘a way out’. This can be

done by dual positioning the sustainable product in question. By morally framing the product

you enable the customer to act on the moral appeal and satisfy their desire to maintain a moral

self-image. Simultaneously, they want to minimize the anticipated resentment from

bystanders, and marketers can help with this by adding a second ‘positioning’ to the product

which is based on non-moral arguments. The group discussion results give some clear

examples of arguments that can be used for this dual positioning, such as health or taste

arguments. A moral appeal can be completed with sentences like: “Better for the environment,

but also better for yourself because it strengthens your immune system and slows down the

aging process ”, or: “Better for the environment, but at the same time does not compromise on

the authentic taste ”. This way, latent do-gooders are able to act on the moral appeal, while not

having to anticipate the same amount of peer resentment as before, because the reasoning

behind their product choice is not as clear for bystanders anymore.

25

Literature

Allen, V. L., & Levine, J. M. (1968). Social support, dissent and conformity. Sociometry, 31(2), 138-149.

Aronson, E., Wilson, T. D., & Akert, R. M. (2005). Social Psychology (5th ed.). Upper Saddle River, NJ: Prentice Hall.

Aronson, E., Wilson, T. D., & Akert, R. M. (2007). Conformity: Influencing Behavior. Social Psychology. Pearson.

Asch, S. E. (1951). Effects of group pressure upon the modification and distortion of judgments. Groups, leadership, and men, 222-236.

Asch, S. E. (1955). Opinions and social pressure. Readings about the social animal, 193(5), 17-26.

Asch, S. E. (1956). Studies of independence and conformity: I. A minority of one against a unanimous majority. Psychological monographs: General and applied, 70(9), 1-70.

Back, K. W. (1951). Influence through social communication. The Journal of Abnormal and Social Psychology, 46(1), 9-23.

Barnard, W. A. (1991). Group influence and the likelihood of a unanimous majority. The Journal of Social Psychology, 131(5), 607-614.

Berkowitz, L., & Macaulay, J. R. (1961). Some effects of differences in status level and status stability. Human Relations, 14, 135-147.

Bolderdijk, J. W., Brouwer, C., & Cornelissen, G. (2016). When do moral pioneers elicit irritation instead of imitation? Manuscript in preperation.

Bolderdijk, J. W., & Cornelissen, G. (2015). The Undercover Altruist: When Doing Good Is Socially Awkward. Association for Consumer Research, 43, 491-491.

Bolderdijk, J. W., Steg, L., Geller, E. S., Lehman, P. K., & Postmes, T. (2013). Comparing

the effectiveness of monetary versus moral motives in environmental

campaigning. Nature Climate Change, 3(4), 413-416.

26

Cornelissen, G. & Bolderdijk, J. W. (2013). Social selfishness. Manuscript in preparation.

Cramwinckel, F. M., van Dijk, E., Scheepers, D., & van den Bos, K. (2013). The threat of moral refusers for one's self-concept and the protective function of physical cleansing. Journal of Experimental Social Psychology, 49(6), 1049-1058.

Darley, J. M. (1966). Fear and social comparison as determinants of conformity behavior.

Journal of Personality and Social Psychology, 4(1), 73-78.

Deutsch, M., & Gerard, H. B. (1955). A study of normative and informational social influences upon individual judgment. The Journal of Abnormal and Social Psychology, 51(3), 629.

Fox, N., & Ward, K. J. (2008). You are what you eat? Vegetarianism, health and identity. Social Science & Medicine, 66(12), 2585-2595.

Gerard, H. B., Wilhelmy, R. A., & Conolley, E. S. (1968). Conformity and group size. Journal of Personality and Social Psychology, 8(1), 79-82.

Greenebaum, J. B. (2012). Managing Impressions “Face-Saving” Strategies of Vegetarians and Vegans. Humanity & Society, 36(4), 309-325.

Holmes, J. G., Miller, D. T., & Lerner, M. J. (2002). Committing altruism under the cloak of self-interest: The exchange fiction. Journal of Experimental Social Psychology, 38(2), 144-151.

Jones, P., Clarke-Hill, C., Comfort, D., & Hillier, D. (2008). Marketing and sustainability. Marketing Intelligence & Planning, 26(2), 123-130.

Jordan, J., Mullen, E., & Murnighan, J. K. (2011). Striving for the moral self: The effects of recalling past moral actions on future moral behavior. Personality and Social Psychology Bulletin, 0146167211400208.

Maslow, A. H. (1943). A theory of human motivation. Psychological review, 50(4), 370-396.

Minson, J. A., & Monin, B. (2012). Do-gooder derogation disparaging morally motivated

minorities to defuse anticipated reproach. Social Psychological and Personality

Science, 3(2), 200-207.

27

Myers D. G. (2013). Social Psychology (11th ed.). New York: Freeman Press.

Peattie, K., & Crane, A. (2005). Green marketing: legend, myth, farce or prophesy?. Qualitative Market Research: An International Journal, 8(4), 357-370.

Peattie, K., & Peattie, S. (2009). Social marketing: A pathway to consumption reduction?. Journal of Business Research, 62(2), 260-268.

Pickett-Baker, J., & Ozaki, R. (2008). Pro-environmental products: marketing influence on consumer purchase decision. Journal of Consumer Marketing, 25(5), 281-293.

Sheth, J., & Parvatiyar, A. (1995). Ecological imperatives and the role of marketing. In M.

Polonsky, A. Mintu-Wimsatt (Eds.), Environmental Marketing: Strategies, Practice, Theory and Research, (3-20). New York: Haworth Press.

Täuber, S., van Zomeren, M., & Kutlaca, M. (2015). Should the moral core of climate issues be emphasized or downplayed in public discourse? Three ways to successfully manage the double-edged sword of moral communication. Climatic Change, 130(3), 453-464.

Thøgersen, J. (2011). Green shopping for selfish reasons or the common good?. American Behavioral Scientist, 55(8), 1052-1076.

Vegetarian Times. (2008). Vegetarianism in America. Harris Interactive Service Bureau.

Retrieved from http://www.vegetariantimes.com/article/vegetarianism-in-america.

Wilder, D. A. (1977). Perception of groups, size of opposition, and social influence. Journal

of Experimental Social Psychology, 13(3), 253-268.

28

Appendices

LIST OF APPENDICES:

Appendix A: Questionnaire ... 29

Appendix B: Petitions ... 37

Appendix C: Experimental procedure (elaborate) ... 39

Appendix D: Final data experiment ... 43

Appendix E: Logistic regression clarification (SPSS output) ... 44

Appendix F: Qualitative analysis of discussion commentary ... 46

29

APPENDIX A: Questionnaire

30

31

32

33

34

35

36

37

APPENDIX B: Petitions

The morally framed petition

38

The selfishly framed petition

39

APPENDIX C: Experimental procedure (elaborate)

The researcher gets one participant from the waiting area and brings him/her to the questionnaire room, and tells the participant:

“Welcome to my study on vegetarian and vegan products! In this study you are going to join a short group discussion together with other participants. Before the discussion starts, I need you to fill in some preliminary background questions at the computer. Once you have finished answering the questionnaire, I will pick you up for the group discussion.”

The participant starts the questionnaire and a few minutes later the researcher gets the second participant and brings him/her to the survey room as well (and gives the same explanation).

The first person should be nearly finished with the survey and he/she will be asked to join a group for the discussion:

“We have one spot left for a group discussion in another room. Please come with me to fill up the group and join the discussion.”

The other participant is told to wait in the survey room when he/she is finished filling out the questionnaire. The researcher mentions that she will pick him/her up in a couple of minutes when there is another group available for a discussion. Then she takes the first participant to the discussion room.

The researcher takes the participant from the survey room and guides him/her to the discussion room. Here are three other ‘participants’ (confederates) waiting for the group to be completed in order to get the discussion started. The researcher makes the participant sit down at the end of the row and introduces the next part of the experiment:

“Okay, thank you all for coming, and thank you guys for waiting (towards the confederates).

Over here (points to the table), I have some vegan chorizo for you to try, in case you never tried a meat substitute before. Feel free to taste.”

Before the discussion actually starts, the researcher tells the group one more thing, namely to

sign a petition (which is placed on the desk behind the researcher) on getting more veggie

alternatives in the assortment of supermarkets. Depending on the assigned conditions, she will

ask the group to sign a morally framed petition which she signed herself as well (moral + non-

unanimous), or a selfishly framed petition which she did not sign herself (selfish +