Corporate Social Irresponsibility and the influence of a firm’s
prior CSR reputation
The influence of the valence of the CSR reputation and the congruency between the
domain of the corporate social irresponsible behaviour and the domain of the CSR
reputation on consumers’ attitude
Author: Berbel Pietersma (10736360) Under supervision of: Drs. M. Vock
Second assessor: Frank Slisser
Final version -‐ January 30, 2015
University of Amsterdam, Faculty of Economics and Business Msc Business Administrations – Marketing track
2
Acknowledgements
I would like to express my deepest appreciation to all those who have been of great help during this study. Firstly, I want to thank my supervisor Ms. Marlene Vock for her support, feedback and calmness during the process. Secondly, I want to thank Wieteke Kluiwstra and Jessie Helmond for their cooperation during the first phase of this study. Finally, I want to thank my family and friends for their support and search for respondents. This master thesis is the final stage of my master Business Administrations. It has been a great journey in which I developed my research skills and my knowledge about business in general, marketing and CSR tremendously.
3
Table of Contents
Acknowledgements ... 2
Abstract ... 5
List of abbreviations ... 5
1. Introduction ... 6
1.1 Theoretical contribution and managerial implications ... 8
2. Theoretical framework ... 10
2.1 Corporate Social Responsibility ... 10
2.2 Corporate Social Irresponsibility ... 11
2.3 The valence of the CSR reputation ... 15
2.4 The relatedness of CSI ... 18
2.4.1 Hypocrisy ... 19
2.4.2 Expectancy ... 21
2.5 Consumers’ attitude to CSR and CSI ... 23
2.6 Conceptual model ... 24
3. Method ... 24
3.1 The research design ... 25
3.2 Pre-tests ... 26
3.3 Independent variables and stimuli ... 30
3.4 Dependent variable ... 31
3.5 Mediators ... 31
3.6 Controls ... 31
3.7 Other variables ... 32
3.8 The participants ... 32
3.9 The procedure ... 33
4. Results ... 33
4.1 Descriptives ... 33
4.2 Controls ... 35
4.4 Hypothesis 1 ... 35
4.5 Hypothesis 2 and 4 ... 37
4.6 Potential mediation effects ... 41
4.7 Demographic results ... 45
5. Discussion ... 47
5.1 The valence of the CSR reputation ... 47
5.2 Relatedness of the CSI ... 49
6. Limitations and recommendations ... 51
Appendix I: The CSR reputation scenario’s ... 61
Appendix II: The CSI scenarios ... 62
Appendix III: Questionnaire for pre-test II and III ... 64
Appendix IV: Dutch version of the full final scenarios ... 65
Appendix V:
English version of the full scenarios ... 68
Appendix VI: Output pre-tests ... 71
Appendix VII: Demographics ... 77
Appendix VIII: The questionnaire (in Dutch) ... 79
Appendix IX:
The questionnaire (in English) ... 82
Appendix X: Histograms dependent variables ... 85
Appendix XI:
SPSS output correlation analysis ... 89
Appendix XII: SPSS output of the reputation manipulation check ... 90
Appendix XIII: SPSS output of hypothesis 1 ... 91
Appendix XIV: SPSS output of hypothesis 2 and 4 ... 92
Appendix XV: SPSS output hypothesis 3 and 5 ... 100
Appendix XVI: SPSS output of the demographics ... 105
Appendix XVII: Realism and credibility ... 109
5
Abstract
This study consists of an online experiment, to investigate the influence of Corporate Social Irresponsibility (CSI) on consumers’ attitude toward a firm, by investigating (a) the influence of the valence of a company’s prior CSR reputation (positive or negative) and (b) the congruency between the domain of the CSR reputation and the domain of the company’s CSI (related or unrelated). A significant interaction effect is found between the valence of the prior CSR reputation of a company with the relatedness of the transgression with the firm’s prior CSR domain. The results revealed that, regardless the congruency between the domain of the CSI and the domain of the company’s CSR reputation, corporate irresponsible behaviour will conduct significantly more negative consumer attitudes toward a company if this company already had a bad CSR reputation than toward a company that had a prior good CSR reputation. Putting effort in building a good CSR reputation could in some cases pay off in terms of buffering harm of CSI, but only occasionally if the transgression is unrelated with the CSR domain of the CSR reputation. More important appeared the avoidance of a bad CSR reputation since CSI is found even more harmful when a company already has a prior bad CSR reputation than when there is no CSR reputation at all.
List of abbreviations
CSR = Corporate social responsibility CSI = Corporate social irresponsibility
6
1. Introduction
“Though CSR tends to be seen as a moral matter, it comes down to the hard practicalities of companies’ reputations and risk registers.”
- The Economist, May 2013, following the collapse of an eight-storey garment factory in Rana Plaza on the outskirts of Dhaka
Due to the increasing availability of public information about firm’s activities and the growing evidence that consumers are willing to pay more for products from corporations that are socially responsible, Corporate Social Responsibility (CSR) activities have been widely adopted. And while CSR is becoming increasingly important to companies, corporate social irresponsibility (CSI) becomes harder to conceal from consumers (Webb & Mohr, 2005; Sen & Bhattacharya, 2001). It seems that there is often a gap between stated CSR business practices and actual CSR business behaviour (Wagner et al., 2009). For example, the apparel company Gap Inc, has recently been praised for their eco-friendly policies and it has even been chosen Worlds’ Most Ethical Company in the apparel category for eight years in a row (by the Ethisphere Institute) (Bogen, 2014). However, in may 2013, Gap Inc. which works with almost 80 factories in Bangladesh, refused to sign a legally binding international accord to protect Bangladeshi garment workers that was initiated after the Rana Plaza building in Bangladesh collapsed which killed hundreds of garment workers (Engel, 2014). This has caused a lot of distrust and negative media attention for the company and calls into question whether “doing good” while also “doing harm” leads to reputational consequences that might in some cases be worse than simply doing nothing at all in the field of CSR.
The use of the Internet and social media has increased transparency and negative messages are shared in particular (Popescu, 2013). Companies need to ensure that they do everything they can to enhance their reputation and avoid anything that might damage it. As Warren Buffet once said: “It takes 20 years to build a reputation and 5 minutes to ruin it. If you think about that, you’ll do things differently” (Helm et al., 2011). Due to the increasing transparency and inconsistent CSR messages
7 appearing in the media, more insights in the effect of contradictory CSR messages on consumer response are needed. Companies are frequently criticized for socially irresponsible business practices. Consequently, companies frequently make an effort to demonstrate socially responsible behaviours to be perceived as good corporate citizens. Managers might believe that a positive CSR reputation can shield the company from ethical mishaps, but is this actually the case?
In the last couple of years CSR has shifted from a niche activity to a mainstream business focus across all sectors of industries and companies of all sizes and it has been acknowledged as a key component of a firm’s image, having a substantial impact upon consumer behaviour (Brown and Dacin, 1997). CSR does not only affect purchase decisions, but also a consumers’ evaluation of the company. It appears to be an important driver of customer satisfaction, the propensity of consumers to buy a brand’s products or services, and eventually the CSR-firm market value (Xueming & Bhattacharya, 2006).
Most prior research investigated when and why CSR has a positive effect on consumers (e.g. Brown & Dacin, 1997; Sen & Bhattacharya, 2001; Du, Bhattacharya and Sen, 2007). However, managers might ask themselves whether CSR always leads to a positive image and customer satisfaction. Indeed, there seems to be a risk for having a certain CSR reputation if consumers find out that the company is actually not fully behaving like their stated CSR, when the firm shows CSI (Wagner et al., 2009). Consumers’ perceptions of CSI often influence the formation of their attitudes toward the company, and therefore potentially contribute to a more negative company image and reputation (Grappi, 2013; Sen & Bhattacharya, 2001). Furthermore, corporate irresponsible behaviour has been found to have a greater impact on consumers’ purchase intention than responsible behaviour (Bhattacharya & Sen, 2004). For example, H&M has recently been known for organic cotton, which is good for the environment. When the consumer finds out that they are actually harming the environment in some other way, the consumer will probably perceive them as inconsistent and hypocrite (Wagner, 2009).
Nowadays companies often engage in multiple CSR activities and mostly put emphasis on one CSR topic in particular. While prior studies have focused on corporate ethical transgressions, none of them have investigated the impact of the relatedness of the domain of the transgression with the
8 domain of the company’s prior CSR reputation. Therefore, it is not yet clear whether a transgression in the prior domain of the firm’s positive CSR reputation is considered worse than if the transgression is not related with this domain. For example, how would consumers respond if they knew a firm for their environmentally friendly policies, but then find out that this company does not treat their employees well? Consumers might respond differently to such an ‘unrelated’ transgression than they would in a situation in which the transgression conflicts with the prior ‘related’ CSR reputation. In other words, the firm’s CSR reputation and the type of corporate ethical transgression, in terms of whether it is related or unrelated to the firm’s prior CSR reputation, may moderate the relationship between unethical corporate behaviour and consumers’ attitude toward that firm.
The next research question will therefore be examined: How are consumers’ attitudes with regard to corporate social irresponsible behaviour influenced by the valence of the company’s CSR reputation (positive or negative) and the congruency between the domain of the CSR reputation and the CSR domain of the company’s irresponsible behaviour (related or unrelated)?
1.1 Theoretical contribution and managerial implications
This study contributes to scientific research in several ways. Although CSR is an established topic within academic research and in the business community, little is known about corporate ethical transgressions in combination with a specific CSR reputation. The studies of Yoon et al., (2006), Wagner et al. (2009), Grappi et al. (2013) and Effron & Monin (2010), all already brought attention to the effect of (corporate) ethical transgressions and CSR reputations in a specific way. However, none of them investigated the combination of corporate social irresponsibility with different types of CSR reputations and the issue of the relatedness of the CSI with the domain of the prior CSR reputation
Yoon et al. (2006) have done research on companies with bad reputations and the relatedness of their CSR activities. They conclude that the relatedness of CSR activities with the firm’s reputation has a significant impact on consumer responses. However, they do not pay attention to CSR transgressions and they do not compare their findings with companies with positive reputations. The
9 valence of the company’s CSR reputation, in terms of a positive or negative reputation, may influence the relationship between a CSI activity of a firm and consumers’ attitude toward that firm.
Wagner et al. (2009) investigated CSR inconsistencies of concrete and abstract CSR statements, but they only paid attention to CSIs that were related to the prior CSR domain of the company. They did not pay attention to unrelated CSIs. The research of Grappi et al. (2013), studied how CSR transgressions affect consumers, but they did not take the prior CSR reputation of the company into account. Effron & Monin (2010) examined when and why an actor’s prior good deeds make observers more willing to excuse his or her morally dubious behaviour. They demonstrated that good deeds can license suspicious behaviour of a person. However, their research is only based on people and not on companies. People might be more suspicious towards companies engaging in CSR than towards people. Additionally, they only looked at ethical transgressions in combination with good reputations.
Thus, prior studies did not yet combine the aspects of the relatedness of the domain of CSI’s with the domain of the firm’s prior CSR reputation and the valence of the CSR reputation. Therefore there are questions still unanswered about the harmfulness of CSI’s. What if the company already has a bad reputation and the transgression is either related or not with the prior CSR domain of the company? And how is this different from a company that has a good CSR reputation? In what situation is a corporate ethical transgression more harmful to a company? In order to fill this gap, this study will make a theoretical contribution to this line of research by not only examining the influence of CSI by companies with good and bad CSR reputations, but also examining this in combination with the influence of the congruency of CSI of companies with specific CSR reputations.
This research gap needs to be filled because CSI can have a huge impact on firm’s reputations and consequently, its profits. It is therefore important to know when CSI is the most and the least harmful, to limit potential reputational damage. Managers might believe that a positive CSR reputation can shield the company from ethical mishaps, but is this actually the case? Should managers really aim to build up a positive CSR reputation on a particular issue? And are transgressions less harmful when a company already has a bad CSR reputation? This study will provide insights into the conditions
10 under which ethical transgressions are most harmful. Managers will therefore be able to minimize reputational damage and consequently, the impact on their profits.
In the next section, the theoretical framework will be presented. Subsequently, the methods of this study will be discussed. Thereafter are the results, the discussion, limitations and future research directions. Finally, the conclusions will be presented.
2. Theoretical framework
2.1 Corporate Social Responsibility
Nowadays there are many different definitions of CSR and the concept of corporate social responsibility has often been characterized as broad and complex (e.g. Öberseder et al., 2013). Since the definition is wide-ranging, this study will continue with the most recent definition of the European Commission. The European Commission refers to CSR as “the responsibility of enterprises for their impacts on society” (European Commission, 2011, p. 6). This definition is complemented by Campbell (2007), by stressing the stakeholders. He suggests that companies “must not knowingly do anything that could harm their stakeholders [...] and if corporatist stations do cause harm to their stakeholders, they must then rectify it when ever the harm is discovered and brought` to their attention” (Campbell, 2007, p. 951).
A strategic approach to CSR is increasingly important to the competitiveness of enterprises. More than ever before, companies engage in CSR initiatives to make a positive contribution to society and to support their strategic goals. According to Öberseder et al. (2013), companies make a clear distinction between different CSR domains and different areas of responsibility that often relate to multiple stakeholders. Skarmeas & Leonidou (2013) found that attributions of just egoistic- and stakeholder-driven motives elicit consumer scepticism toward CSR, while value-driven attributions inhibit scepticism. Since there are different domains of CSR that are related to different stakeholders, many companies are active in multiple domains of CSR.
11 Prior research suggests that there is a positive relationship between a company’s CSR actions and consumers’ attitudes toward a company and its products or services (Brown and Dacin, 1997; Creyer and Ross, 1997; Ellen, Mohr, and Webb, 2000). Additionally, Sen & Battacherya (2001) found that, not surprisingly, consumers’ evaluations of a company are more sensitive to negative CSR information than to positive CSR information.
While CSR is often thought of as engaging in “doing good”, a less visible dimension of CSR is “not doing harm”. Minor & Morgan (2011) found a synergistic effect on reputation for firms who are good in both doing good and avoiding harm. However, they found that doing good while also doing harm leads to reputational consequences that are worse than simply doing nothing at all in the field of CSR. The next section will elaborate on corporate social irresponsibility.
2.2 Corporate Social Irresponsibility
Nowadays, CSI becomes harder to conceal from consumers. The use of the internet and social media has increased transparency and negative messages are shared in particular (Popescu, 2013). Companies are frequently criticized for socially irresponsible business practices through social media, mass media and consumer advocacy groups (Wagner et al., 2008). Because of numerous reported incidents of corporate misbehaviour, many people doubt the extent to which companies actually live up to their professed CSR standards, and consumer scepticism toward corporate social involvement grows (Skarmeas & Leonidou, 2013).
The study and examination of CSI in the academic literature was introduced by Armstrong (1977, p. 185), who defined CSI as follows: “A socially irresponsible act is a decision to accept an alternative that is thought by the decision maker to be inferior to another alternative when the effects upon all parties are considered. Generally this involves a gain by one party at the expense of the total system”. Yet, despite an increasing emphasis on CSR, still little is known about consumers’ responses to CSI. Although prior studies have shown a positive relationship between a company’s CSR behaviour and consumers’ attitudes toward that company (Brown and Dacin, 1997; Ellen, Mohr and Webb, 2000), there also seems to be a risk of having a certain CSR reputation if consumers find out that the company is actually not fully behaving in accordance with their stated CSR (Wagner et al.,
12 2009). Indeed, Bhattacharya & Sen (2004) found that CSI has a greater impact on consumers’ purchase intention than CSR. Additionally, Sweetin et al. (2013) found that consumers are more likely to punish and less likely to reward a company that shows CSI than a company that does not show CSI. Thereby, Wagner et al. (2009) found that a company that claims to be socially responsible but subsequently shows CSI, is perceived to be hypocrite.
According to Osgood and Tannenbaum’s congruity theory (1995), inconsistent information could lead to a sceptical attitude and negative consumer responses. Incongruity is unpleasant and motivates people to change their attitudes (Osgood and Tannenbaum, 1995). Therefore, when a CSR statement appears to be incongruent with a firm’s CSR behaviour, consumers will often be motivated to change their attitudes toward this firm and therefore potentially contribute to a negative company image and reputation (Grappi et al., 2013; Wagner et al. 2009).
Table 1 shows an overview of the research that is done in the field of consumers’ responses to CSI. All of these studies underline the negative effect of CSI on consumer responses. Most of these studies investigated the effect of CSI on the level of negative consumer responses, determining negative emotions, attitudes, uncertainty level and other negative responses.
Table 1: Overview of research in the field of consumers’ responses to social irresponsibility.
Year Author Title Publication
2008 Wagner, T., Pelin
B., and Zachary R.H.
The dark side of retailing: towards a scale of corporate social irresponsibility.
International Journal of Retail & Distribution Management, 36(2),
124-142.
2009 Wagner, T., Lutz,
R.J., and Weitz, B.A.
Corporate Hypocrisy: Overcoming the Threat of Inconsistent Corporate Social Responsibility Perceptions. Journal of Marketing, 73(6), 77-91. 2013 Grappi, S., Romani, S. and Bagozzi, R.P.
Consumer responses to corporate irresponsible Behaviour: moral emotions and virtues.
Journal of business research.
66(10) , 1814-1812.
2013 Sweetin, V.H.,
Knowles, L.L., Summey, J.H. and McQueen K.S.
Willingness to punish the corporate brand for corporate social irresponsibility
Journal of business research.
13
2014 Kim, S. The Role of Prior Expectancies and
Relational Satisfaction in crisis.
Journalism & Mass
Communication, 9(1), 139-158.
In their quantitative research, Wagner et al. (2008) provided a comprehensive overview of the different CSI domains, such as natural environment, foreign labour and dishonesty. See figure 1. They state that companies will need to concentrate on the dimensions that are of relevance to their particular business activities and their stakeholders.
14 Grappi et al. (2013) defined two categories of moral transgressions, ethical transgressions and social transgressions, which have been proposed in psychology and anthropology literature. The first category, ethical transgressions, concerns harm done to other people, such as to employees or to consumers. By doing harm, they mean violate the freedom or human dignity of individuals. The second category of transgressions, social transgressions, concerns harm done to a community in some way. For example violating norms or expectations of respect of, and loyalty to a community. They investigated consumer responses to corporate irresponsible behaviour in combination with other related virtues, negative word of mouth and protest toward the corporation. They demonstrated how both types of transgressions instigate negative word of mouth and protest toward the company and both could therefore be extremely harmful for companies.
Likewise, Kim (2014) investigated the harmfulness of CSI. He found that stakeholders’ prior expectancies play a vital role in their evaluation of CSIs of a specific company. He states that, just as transgressions violate a partner’s expectancies in personal relationships, companies’ transgressions can violate the public’s expectancies in company–public relationships. It was found that stakeholders’ expectancies are significant predictors determining negative valence, uncertainty level, and other negative responses toward an organization showing CSI.
Thus, all of these studies underline the negative effect of CSI on consumer responses and thereby on the companies’ eventual CSR reputation. Therefore, we know that CSI is harmful for a companies’ reputation. But what about the influence of a companies’ prior CSR reputation in case of an ethical transgression? Accept for Wagner et al. (2009), none of these studies take the valence of the prior CSR reputation of firms into account when they study the effect of CSI. Therefore there are questions still unanswered about the harmfulness of CSIs. What if the company already had a bad CSR reputation and subsequently shows CSI. Will consumers still be negatively surprised? This study will contribute to these prior studies by firstly examining the impact of the valence of a firm’s prior CSR reputation on the harmfulness of CSI. The valence of a firm’s prior CSR reputation will be described in the following section.
15
2.3 The valence of the CSR reputation
According to a study by the Reputation Institute, a private global consulting firm, consumers’ willingness to buy, recommend, and invest in a company is driven 60% by consumers’ perceptions of the company – or it’s reputation, and only 40% by consumers’ perceptions of the product or services it sells (Smith, 2013). It is therefore not very surprising that in particular, companies with bad reputations seem to be interested in changing their negative image through CSR activities. However, while some companies successfully changed their image by stressing their CSR initiatives, the same strategy has backfired on others.
Based on the theories of attribution and suspicion, Yoon et al. (2006), provide an explanation for the circumstances under which CSR activities may or may not achieve their desired effects in the context of companies with bad reputations. In their research they specify conditions under which consumers become suspicious of the true motives behind a CSR activity. The authors found that when consumers feel that the CSR activity is only used to improve the company’s image, for example when a company spends more on advertising the CSR features than they spend on their actual CSR activities, CSR is not only inefficient, but may even backfire on the company. However, Yoon et al. (2006) only looked at the influence of CSR activities on companies with bad reputations. They did not study the effect of CSI activities on companies with bad reputations. Additionally, they only did research in the context of companies with bad reputations; they do not compare their findings with companies with positive reputations.
According to Effron & Monin (2010) a good reputation, in terms of previous good deeds, can license an ethical transgression. They examined when and why an actor’s prior good deeds make observers more willing to excuse his or her morally dubious behaviour. They demonstrated that previous good deeds can license and justify suspicious behaviour of a person. Although this study is based on persons, this theory might also be applicable to companies. A company with a good CSR reputation might already have a ‘buffer’, which could make CSI less harmful. A company with a bad reputation does not have such a buffer and CSI might therefore be more harmful to this company than it is for a company with a prior good CSR reputation. However, Einwiller et al. (2006) found that prior
16 positive associations can buffer the negative effect of moderately negative associations, but that this buffer effect disappears when the associations are extremely negative.
In addition, Wagner et al. (2009) examined the effect of inconsistent CSR information on companies with both prior positive abstract and concrete CSR statements. They found that prior concrete CSR statements, followed by conflicting CSI, are more harmful to a company than abstract CSR statements followed by conflicting CSI. The company is in that case seen as more ‘hypocrite’. However, it has been found that the inconsistent information in both scenarios increases perceptions of hypocrisy. For example, H&M has recently been known for organic cotton, which is good for the environment. When a consumer finds that they are actually harming the environment in some other way, the consumer will perceive them as inconsistent and hypocrite (Wagner, 2009). Thus, it has been found that CSR inconsistency has a negative impact on consumers’ attitudes toward firms, both directly and indirectly through negatively affected CSR beliefs. This way, CSR statements can actually be counterproductive if a company fails to comply with its own CSR statements.
However, Wagner et al. (2009) did not pay attention to the ‘relatedness’ of the CSI with the firm’s CSR statements; they only studied ‘related’ CSI, i.e. CSI in the same domain as the abstract or concrete CSR statement. These related transgressions might be considered worse than potential unrelated transgressions. Inconsistent CSR information can come in many different forms. For example, a company could have a positive CSR statement regarding a specific social issue, such as labour conditions, but consumers may then be confronted with negative CSR information about the company's impact on the environment. This study will contribute to the study of Wagner et al. (2009), by studying the effect of the relatedness of the CSI with the prior domain of the firm’s prior CSR domain. Additionally, Wagner et al (2009) looked at specific CSR ‘statements’. But how will the consumer respond if the company has an established reputation in the field of CSR, because it is known for multiple previous CSR activities, and subsequently shows CSI? Will this reputation be detrimental to the company when they show CSI, because the company will be perceived as hypocrite, like Wagner et al. (2009) found for inconsistent CSR statements? Or will this positive CSR reputation be beneficial and function as a buffer, like Effron & Monin (2010) found for the good reputation of people who make ethical transgressions?
17 Two contradictory hypotheses can thus be expected from these prior studies. Firstly, as the study of Effron & Monin (2010) showed, prior good deeds can licence ethical transgressions. Based on this licencing theory, it can be expected that a firm’s prior good CSR reputation can work as a buffer and licence and justify CSI, which will lead to less negative consumer attitudes than when the company would have a prior negative CSR reputation or no CSR reputation at all. It could thus be expected that consumers will respond in a more negative way on corporate ethical transgressions when the company already has a prior bad CSR reputation, because there is no such buffer available (Effron & Monin, 2010). Besides, CSR activities of a company with a bad CSR reputation can already lead to negative consumer attitudes because of profound suspicion regarding the sincerity of the company’s motives (Yoon et al., 2006). When a company with a bad CSR reputation makes a transgression in one of its CSR statements, this is expected to conduct even more negative consumer attitudes. This leads to the following hypothesis:
Hypothesis 1A: Corporate irresponsible behaviour will conduct less negative consumer attitudes if the company has a prior good CSR reputation than if the company has a prior bad CSR reputation.
However, a good CSR reputation might also be detrimental to a company when it shows CSI, because the company could be perceived as hypocrite, like Wagner et al. (2009) found for inconsistent CSR statements. Additionally, the expectancy-violation perspective states that consumers expect more of a firm with a good reputation, which subsequently will be seen in a more critical way in the case of a violation of expectations (e.g. Coombs & Holladay, 2001; in: Helm & Tolsdorf, 2013). Therefore, it can be expected that companies with a bad CSR reputation do not have much to lose in case of violation of expectations. A CSR transgression by a company with a bad CSR reputation could therefore be seen as less harmful because there are lower expectations. This leads to the following hypothesis:
18 Hypothesis 1B: Corporate irresponsible behaviour will conduct more negative consumer attitudes if the company has a prior good CSR reputation than if the company has a prior bad CSR reputation.
This research will make a contribution to these prior studies by both examining both the prior valence of the CSR reputation of a company in case of CSI and the influence of the relatedness of the CSI with the domain of the firm’s prior CSR reputation. E.g.: if the company has a prior good CSR reputation, is it more harmful if the domain of the CSI is related with the domain of the firm’s prior CSR reputation, or if the domain of the CSI is unrelated to the domain of the prior CSR reputation? And is this different from a company with a bad CSR reputation in the same CSR domain? The concept of the 'relatedness' of the CSI will be discussed in more detail in the next section.
2.4 The relatedness of CSI
As mentioned earlier, apart from to the valence of the CSR reputation of a company, the relatedness of the domain of CSI with the domain of the firm’s prior CSR reputation could have an impact on consumers’ response to CSI (Forehand & Grier, 2003; Yoon et al., 2006). Aforementioned studies mostly studied the effect of corporate ethical transgressions in a ‘related’ context (e.g. Wagner et al., 2009; Sweetin et al, 2013; Kim, 2014). But how does the consumer respond if a company shows corporate irresponsible behaviour in a totally different domain than their prior CSR domain? For example, Gap Inc. has been known for their eco-friendly policies, but was also confronted with negative articles about its refusal to sign the legally binding international accord to protect Bangladeshi garment workers.
Nowadays companies often engage in multiple CSR activities and mostly put emphasis on one CSR topic in particular. The broad range of different CSR domains and subsequent potential consumer value, suggests considerable fragmentation within the marketing field (Peloza and Jingzhi, 2011). A transgression in the same domain as the firm’s prior positive CSR reputation could be considered worse than if the transgression takes place in a totally different domain (Effron & Monin, 2010). In
19 other words, the type of corporate ethical transgression, in terms of whether it is related or unrelated to the firm’s prior CSR reputation, may moderate the relationship between unethical corporate behaviour and attitudinal consumer response.
Yoon et al. (2006) have already done research on the relatedness of CSR activities in the context of companies with bad reputations. They conclude that the relatedness of CSR activities with the domain of the firm’s negative reputation has a significant impact on consumer responses. CSR activities that are related with the company’s business could increase the salience of firm-serving benefits (Forehand & Grier, 2003; Yoon et al., 2006). For example, in case of an oil company which claims to be eco-friendly or a tobacco company which supports a cancer fund, consumers are presumably aware that the company has a negative image and hopes to improve it through the CSR activity. Such combination arouses profound suspicion regarding the sincerity of the company’s motives, which could eventually backfire on the company, and this may lead to an even more negative image than without this CSR activity (Yoon et al. 2006).
However, Yoon et al (2006) only looked at the impact of the relatedness of CSR activities in combination with companies with prior bad reputations; they did not study the impact of the relatedness of CSI activities. Additionally, they only studied companies with overall bad reputations, like tobacco companies and oil companies. They did not specify this to a specifically bad CSR reputation. Finally, they do not compare their findings to companies with good CSR reputations. Therefore, there are still a lot of uncertainties about the influence of the relatedness of a CSI. However, there are some interesting theories that could already give an indication about the potential influence of the relatedness of CSI, which are described in the following section.
2.4.1 Hypocrisy
Transgressions that are in the same domain as the firm’s prior moral behaviour could create the appearance of making a contradictory claim about its values. In other words, “saying one thing, but doing another” (Barden, Rucker & Petty, 2005). According to Festinger (1957) the inconsistency between two cognitions (e.g. a CSR statement and a display of counter-attitudinal behaviour) arouses
20 an aversive psychological state, which is called cognitive dissonance, which leads to perceptions of hypocrisy. People dislike hypocrites because of a general distaste for inconsistent behaviour.
When a firm claims to be something that it is not, the firm can thus be seen as hypocrite (Wagner et al., 2009). Wagner et al. (2009) introduced the concept of perceived hypocrisy as consequence of inconsistent CSR information. They define corporate hypocrisy as “a firm’s behaviour with regard to CSR when the behaviour is contrary to their stated standards of social responsibility” (Wagner et al., 2009, p. 77). They found that corporate hypocrisy has a negative impact on consumers’ attitudes toward firms, both directly and indirectly, through negatively affected CSR beliefs.
According to Effron & Monin (2010), perceived hypocrisy elicits negative attributions about the prior moral related behaviour, which as a result could seem less sincere or more selfish. They found that the effect of the good deeds is depending on the transgression type and the relatedness of the transgression domain versus the domain of the good deeds; perceptions of hypocrisy were increased when the transgression was in the same domain as the prior good deeds, but the perceptions of hypocrisy were less when the transgression was unrelated to the good deeds. Based on these findings of Effron & monin (2010) it could thus be expected that CSI that is related with the company’s prior positive CSR reputation increases more negative consumer responses than CSI that is unrelated with the company’s prior CSR reputation.
When a firm shows CSI, a prior good CSR reputation in a different CSR domain might reduce observers’ condemnation by seeming to balance out the bad behaviour, whereas prior good deeds in the same domain might make the firm appear hypocritical (Effron & Monin, 2010). For example, when the consumer knows a firm for their organic cotton (i.e. environmental issue), but then finds they didn’t treat their employees well (i.e. social issue), the consumer might accept somehow that they cannot excel in all domains simultaneously. In such a situation, a consumer might be less disappointed about the firm compared to a situation in which the CSR reputation conflicts with the related CSR behaviours. Effron & Monin (2010) found that blatant transgressions are especially licenced by good deeds that are in a different domain. This leads to the following hypotheses:
21 Hypothesis 2: When the company has a good CSR reputation, related corporate irresponsible behaviour will conduct more negative consumer attitudes than unrelated corporate irresponsible behaviour.
Hypothesis 3: When the company has a prior good CSR reputation, the impact of the relatedness of the domain of the CSI with the domain of the CSR reputation on consumers’ attitude toward the firm is mediated by hypocrisy.
However, Effron & Monin (2010) only studied the relatedness of moral transgressions in the context of people. Consumers might be more suspicious towards companies than towards people. Additionally, they only looked at ethical transgressions in combination with prior good reputations, with no attention to prior bad reputations. But what if the reputation is already bad? This situation will be shortly discussed in the next section. This study will make a contribution to these findings by examining the relatedness of ethical transgressions in the context of companies and in the context of both good and bad prior CSR reputations.
2.4.2 Expectancy
According to Yoon et al. (2006) CSR activities of a company with a bad CSR reputation can already lead to negative consumer responses because of profound suspicion regarding the sincerity of the company’s motives (Yoon et al., 2006). So, what if a company with a bad reputation shows corporate irresponsible behaviour? Will this be even more harmful to the company?
Just as people have expectations of others in interpersonal relationships, consumers have certain expectations of organizations in company-consumer relationships. And just as transgressions violate a partner’s expectancies in personal relationships, organizations’ transgressions can violate consumers’ expectancies in a company-public relationship (Kim, 2014). According to the expectancy violation theory, when information about a target violates stereotype-based expectations, evaluations are getting more extreme in the direction of that violation (Biernat et al., 1999). The expectancy-violation perspective predicts that customers expect more of a firm with a good reputation, which
22 subsequently has more to lose than a company with a bad reputation in the case of violation of expectations (e.g. Coombs & Holladay, 2001; in: Helm & Tolsdorf, 2013). This way, it can be expected that companies with a bad CSR reputation do not have much to lose in case of violation of expectations. A CSR transgression by a company with a bad CSR reputation could therefore be seen as less harmful because there are lower expectations.
Additionally, Kim (2014) found that stakeholders with prior negative expectancies tend to reveal less negative valence regarding an ethical transgression of an organization, whereby the company eventually perceives less reputational damage then when stakeholders have prior positive expectancies. When a company has a prior bad CSR reputation, consumers do probably have negative expectancies about the firm in the field of CSR. This way, it can be expected that companies with a bad CSR reputation in a specific CSR domain, do not have much to lose in case of a transgression that is related with its prior negative CSR reputation. CSI could thus be highly expected from a company that already had a bad CSR reputation, especially when the CSI occurs in the same domain as the firm’s prior bad CSR reputation and is thus related. Expectancy-violation is not likely to occur in this situation. Because of these low expectations, a related CSR transgression by a company with a bad CSR reputation could be seen as less harmful; it could already be expected. It could be said that there is a lower level of expectancy, because the consumer is already known with the bad reputation in that specific domain but might hold a higher level of expectancy in relation to other domains. Unrelated CSI could in this situation be less expected. Therefore, when a company has a prior bad CSR reputation and subsequently shows CSI, a possible mediator for the impact of the relatedness of CSI on consumers’ attitude toward this firm could be consumers’ expectancy of the CSI. This leads to the following hypotheses:
Hypothesis 4: When the company has a prior bad CSR reputation, related corporate irresponsible behaviour will conduct less negative consumer attitudes than unrelated corporate irresponsible behaviour.
23 Hypothesis 5: When the company has a prior bad CSR reputation, the impact of the relatedness of the domain of the CSI with the domain of the CSR reputation on consumers’ attitude toward the firm is mediated by expectancy.
2.5 Consumers’ attitude to CSR and CSI
The effect of CSIs will be studied by focussing consumers’ attitudes toward a firm. Past studies have suggested that there is a relationship between an individual's attitude and behaviour (Calder and Ross, 1973; Fishbein and Ajzen, 1975). This is based on the theory of reasoned action (Ajzen and Fishbein, 1980) which suggests that attitude toward an act is a determinant of behavioural intention. This means that consumers’ attitudes toward companies likely play a key role in their willingness to buy products of these companies. When a consumer holds a general attitude toward a company, that attitude probably will have a direct effect on the person’s company-specific quality perceptions (MacKenzie and Lutz 1989). Several empirical investigations provide support for the positive relationship between attitude and patronage (e.g., Eastlick and Liu 1997; Korgaonkar et al.1985). Thus a more favourable attitude toward a company is likely to result in a behavioural intention (e.g. purchase behaviour). However, a more unfavourable attitude toward a company, e.g. caused by corporate irresponsible behaviour, is therefore likely to result in negative purchase intentions or even boycotting intentions.
Many studies demonstrate how information about CSR affects consumers’ attitudes toward firms (e.g. Brown & Dacin 1997; Wagner et al. 2009). Research indicates that consumer attitudes toward companies active in CSR are largely positive (Webb and Mohr 1998) and attitudes toward companies that show inconsistent CSR behaviour are largely negative (Wagner et al., 2009). In addition, willingness to purchase a company's product is also positively influenced by the company's CSR activities (Smith and Alcorn 1991).
Because attitudes toward the firm have been demonstrated to affect consumers’ actual purchase patterns in the marketplace (Lichtenstein, Drumwright, and Braig 2004; Pan and Zinkhan 2006), this study adds behavioural relevance by studying how consumers’ attitudes with regard to CSI are influenced by the valence of the company’s CSR reputation and the relatedness of the CSI.
24
2.6 Conceptual model
In the previous sections of the literature review the variables studied in this research and their relationships have been explained. From this theoretical framework a conceptual model can be derived. This conceptual model is shown in figure 2.
Figure 2. Conceptual model
3. Method
This study consists of an online quasi-experiment, using Qualtrics, to investigate the influence of CSI on consumers’ attitude toward a firm, by investigating (a) the influence of the valence of a company’s CSR reputation (positive or negative) and (b) the congruency between the domain of the CSR reputation and the domain of the company’s CSI. The questionnaire was distributed via Facebook and participants were randomly assigned to a condition. Participants rated their attitudes toward a fashion company after reading general information about this fictitious organization, its CSR reputation and, subsequently, specific CSI by this company.
Because of this deductive quasi-experimental approach, the independent variables are manipulated and controlled to easily determine the cause and effect relationship. However, no direct causal relationships can be determined due to extraneous and confounding variables that exist in a
The relatedness of the domain of the CSI with prior
CSR domain of the firm (related/unrelated)
The valence of the CSR reputation of the firm
(good/bad/ no CSR)
Consumers’ attitude toward the firm CSI
Hypocrisy
25 social environment. However, utilizing this quasi-experimental design minimizes the threats to external validity, as natural environments do not suffer problems of artificiality (Shadish et al., 2002).
In this section, the design, the pre-tests, the stimuli, the dependent variable and other variables will be described. Subsequently, the participants and the procedure will be described.
3.1 The research design
This study used a 3x3 between-subjects design, as shown in figure 3. The first independent variable is the valence of the firm’s reputation, which is either positive, negative or there is no information about the CSR reputation. This last condition functions as a control condition. The second independent variable is the domain of the CSI, which is either related or unrelated with the domain of the firm’s CSR reputation. The domain of the first CSI is ‘working conditions’ which is related to the firm’s CSR reputation. The domain of the second CSI is ‘environment’, which is unrelated to the firm’s CSR reputation. The domain of the last CSI is ‘animal treatment’, which is again unrelated to the firm’s CSR reputation. There are two unrelated conditions to rule out the effect of a specific CSR domain, even though the domains are equal to each other in terms of ethicality, responsibility, harmfulness, credibility and reliability (see 3.2).
Figure 3: Graphical representation quasi-experimental research design
CSI: working conditions - related Condition 1 Good CSR reputation CSI: environmental crisis - unrelated Condition 2 CSI: animal crisis – unrelated Condition 3 CSI: working conditions - related Condition 4 The company Bad CSR reputation CSI: environmental crisis - unrelated Condition 5 CSI: animal crisis – unrelated Condition 6 CSI: working conditions - neutral Condition 7 No CSR / Neutral CSI: environmental crisis - neutral Condition 8 CSI: animal crisis - neutral Condition 9
(Prior domain CSR reputation: working conditions)
26
3.2 Pre-tests
Before the start of the experiment, texts were manipulated and measured. Therefore, several pre-tests were done in order to test the instruments, to ensure they meet their purpose. By private messages on Facebook, the participants were asked to join one of the three pre-tests. During the first pre-test, the valence of the company’s reputation was measured, which should be positive, negative or neutral (no information about the CSR reputation). The texts for the positive and negative reputation story were fairly similar to each other; the texts only differed by several words that are positive in one version and negative in the other version. E.g.: in the positive version: “Fashion Loft is known as the industry leader when it comes to socially responsible business practices.” and in the negative version: “Fashion Loft is known as the industry laggard when it comes to socially responsible business practices.” The full texts of the different CSR reputation scenario’s can be found in Appendix I.
Before starting the experiment, the different reputation scenarios were tested, to make sure that the right effect is created and to prevent side effects. The positive and negative CSR reputation were tested on a 7-point Likert scale, operationalised as ‘perceived social responsibility’ (Wagner et al., 2009), using three item questions (α = 0,98), to see if they were actually perceived as positive and negative in terms of social responsibility. In this pre-test, 24 respondents participated. The results are shown in table 2.
There was a significant difference between the positive and the negative CSR reputation story based on social responsibility (F(22)=0,47, p<0,01). As shown in table 2, the positive CSR story was seen as relatively social responsible and the negative CSR story was seen as relatively social irresponsible. This means that the manipulation worked out as it was supposed to.
Table 2: Means and standard deviations of the CSR reputations on ethicality (7-point Likert scale) Positive CSR reputation (N=12) Negative CSR reputation (N=12) 1. Socially responsible (SD) 6,00 (1,13) 1,75 (0,75) 2. Concerned to improve
well-being of society (SD) 5,92 (0,90) 1,50 (1,00)
3. High ethical standards (SD) 5,67 (0,89) 1,33 (0,65)
Overall mean (SD) 5,86 (0,87) 1,53 (0,59)
In the second pre-test, the congruency between the CSR reputation and the CSR domain of the company’s irresponsible behaviour was manipulated based on real life news articles about fashion companies; the irresponsible behaviour was either related to the firm’s prior CSR domain, in this case in the field of the firm’s working conditions, or unrelated. The full texts of the five different CSIs that were used in this pre-test can be found in Appendix II. There was one related transgression, three unrelated transgressions and one semi-related transgression. These were all transgressions in different CSR domains; working conditions (related; similar to Wagner et al. 2009), environment (unrelated), animal treatment (unrelated), community crisis (unrelated) and health care (semi-related; adapted from on Wagner et al. 2009). The different transgression scenarios were pre-tested to check if they were equal to each other in terms of perceived ethicality (multidimensional ethics scale (MES) by Reidenbach and Robin, 1990), perceived harmfulness (based on Effron & Monin, 2010), perceived responsibility of the firm (based on Effron & Monin, 2010), realism and credibility (based on Skarmeas & Leonidou, 2013). The participants of this pre-test just read one of these transgression scenarios, and received no information about the CSR reputation of the firm. Table 3 shows how the five different CSI scenarios score on perceived ethicality (three items, α= 0,91), perceived harmfulness (one item), perceived responsibility (one item), realism (one item) and credibility (one item). In this pre-test, 34 respondents participated. The full questionnaire of this pre-test can be found in Appendix III.
28 Table 3: Ethicality, Perceived harmfulness, Realism & Credibility of the different transgressions (7-point Likert scale).
Ethicality (SD) Harmfulness (SD) Responsibil- ity (SD) Realism (SD) Credibility (SD) CSI 1: Working conditions (N= 7) 2,00 (1,02) 5,29 (1,70) 5,00 (1,41) 5,14 (1,57) 5,29 (1,60) CSI 2: Environment (N= 8) 1,92 (0,75) 5,88 (0,35) 5,50 (0,76) 5,50 (0,67) 6,00 (0,76) CSI 3: Animal crisis
(N= 7) 2,48 (0,86) 4,43 (1,51) 4,43 (1,27) 4,29 (1,70) 5,14 (0,69) CSI 4: Community
crisis (N= 6) 2,56 (0,75) 4,33 (1,86) 4,83 (1,47) 3,33 (1,51) 3,17 (1,47) CSI 5: Health care
(N= 6) 2,56 (0,81) 3,50 (1,38) 4,67 (1,21) 3,83 (1,72) 4,00 (1,90) Total (SD)
(N= 34) 2,27 (0,84) 4,76 (1,52) 4,91 (1,22) 4,85 (1,58) 4,47 (1,62)
Using a One-way Anova, it was shown that there were no significant differences between the five different transgressions based on ethicality (F(4)=0,98; p=0,44) and on responsibility (F(4)=0,80; p=0,53). However, there were significant differences between the five different transgressions based on harmfulness (F(4)=3,19; p=0,03), credibility (F(4)=2,87; p=0,04) and realism (F(4)=4,30; p<0,01).
For this study, the CSI’s should be equal to each other and a CSI should ideally score low on ethicality (i.e. ‘morally wrong’, ‘unacceptable’) and high on perceived harmfulness, responsibility realism and credibility. Table 3 shows that the CSI in the field of community crisis and the CSI in the field of health care score relatively low on perceived harmfulness, realism and credibility on a 7-point Likert scale (i.e. <4, below midpoint). Therefore it was decided to not use these two CSIs in the actual experiment.
The CSI in the field of animal crisis scored relatively high on ethicality, and low on perceived harmfulness compared to the two remaining CSIs. Therefore, it was decided to adjust this transgression by letting this CSI seem a bit worse. This was done by changing the last sentence of this scenario. The last sentence of this scenario in the first pre-test was: “A coalition of animal rights
29 groups released a gut-wrenching exposé into the angora industry in China, which is responsible for 90 percent of the world’s supply of angora wool.” This sentence was changed into: “A coalition of animal rights organizations revealed heart breaking practices, such as live plucking of angora rabbits. In European countries, the live plucking of angora rabbits prohibited.” Additionally, it was decided to translate the scenarios to Dutch, because the experiment would take place in the Netherlands, and because less educated people had to be reached in order to achieve a representative population. The translated Dutch scenarios can be found in appendix IV.
A third pre-test, in which 18 respondents participated, was done to test the adjustments. As shown in table 4, the three transgressions scored relatively low on ethicality, and relatively high on harmfulness, perceived responsibility, realism and credibility. Additionally, the texts seem to be very similar in terms of perceived ethicality, perceived harmfulness, perceived responsibility, realism and credibility.
Table 4: Ethicality, Perceived harmfulness, Realism & Credibility of the different transgressions (7-point Likert scale).
Ethicality (SD) Harmfulness (SD) Responsibility (SD) Realism (SD) Credibility (SD) CSI 1: Working conditions (N= 7) 2,05 (0,45) 5,00 (2,00) 5,71 (0,76) 4,86 (1,86) 5,86 (0,69) CSI 2: Environment (N= 5) 1,80 (0,77) 5,80 (1,10) 5,40 (1,67) 5,00 (1,87) 5,20 (2,05) CSI 3: Animal crisis (N= 6) 1,56 (0,75) 5,17 (2,14) 5,00 (2,45) 4,83 (1,94) 4,50 (2,07) Total (SD) (N=18) 1,81 (0,15) 5,28 (1,78) 5,39 (1,65) 4,89 (1,78) 5,22 (1,66)
30
3.3 Independent variables and stimuli
The independent variables were the valence of a company’s prior CSR reputation (positive, negative, neutral) and the congruency between the domain of the CSR reputation and the domain of the company’s CSI (related or unrelated).
The actual experiment consisted of a short scenario about a fictitious fashion company. The scenario consisted of three parts. The first part of the scenario provided some neutral and objective information about the fashion company. Every participant read this section. The second part of the scenario provided the information about the CSR reputation of this company. In this study, the prior CSR domain of this company was about working conditions. There were three versions of this part of the scenario: a positive story about the CSR reputation with regards to the working conditions, a negative story about the CSR reputation with regards to the working conditions and finally, there was a version where the information about the CSR reputation is omitted, which is the control condition.
The third part of the scenario provided information about corporate irresponsible behaviour that this company has shown; a transgression in a specific CSR domain. There were three versions of this part of the story: a transgression which was related to the prior CSR domain of the company, which is in this case working conditions and two transgressions that were unrelated to the prior CSR domain of the company. The first unrelated transgression was about an environmental crisis and the second unrelated transgression was about an animal scandal. These transgressions were carefully determined during the pre-tests. Because the experiment has taken place in the Netherlands, and also less educated people had to be reached in order to achieve a representative population, it was decided to have the scenarios read in Dutch by the participants. The Dutch version of the full scenarios can be found in Appendix IV. The English version of the full scenarios can be found in Appendix V.
31
3.4 Dependent variable
The dependent variable in this experiment is attitude. The participant could indicate their attitude toward the firm in the following way (Wagner et al., 2009; Nan & Heo, 2007):
In general, my feelings toward Fashion Loft are … 1. Very unfavourable (=1) / Very favourable (7) 2. Very bad (=1) / Very good (=7)
3. Very unpleasant (=1) / Very pleasant (=7)
These measures are in line with previous studies in consumers’ attitude toward a firm, such as Wagner et al. (2009) and Nan & Heo (2007).
3.5 Mediators
Two potential mediators were measured: hypocrisy and expectancy. The measures of hypocrisy are based on the studies of Wagner et al. (2009) and Effron & Monin (2010), who also measured hypocrisy caused by moral transgressions. The measures of expectancy are based on the study of Biernat et al. (1999).
3.6 Controls
A manipulation check on the valence of the CSR reputation of the firm, right after the reputation scenario, checked if the CSR reputation manipulation worked out in the right way. This manipulation check was adapted from the study of Wagner et al. (2009).
There was no specific manipulation check for the manipulation of the relatedness of the CSI’s with the domain of the firm’s prior CSR reputation since no valid way has been found to test this (similar to Effron & Monin, 2010). However, there was a control for the domain of the CSI, to check whether the participants have read the information about the CSI correct (see Appendix XVIII). Finally, credibility and reliability were tested to check if the scenarios were perceived as credible and realistic.
32
3.7 Other variables
In the last part of the questionnaire, the demographics were measured: gender, age, education and nationality. And since this study is part of a larger research project (a.o. by Ms. Vock), two extra variables were measured during this study: negative emotions and boycott intentions (based on the study of Grappi et al, 2013). However these variables will not be used during this specific study.
3.8 The participants
For this study the population consists of mainly Dutch consumers, with the exception of one German, one Belgian and one Polish participant, since the language in the experiment was Dutch. We strived for equality of men and women. For this experiment, 403 participants opened the experiment online in their browser of whom 135 did not start the experiment and 19 participants dropped out during the experiment. Eventually, 249 participants completed the full experiment, of which 51% were female and 49% were male. Similar to the study of Wagner et al (2009), mainly students participated in this experiment. This is reflected in the age and education of the participants. The participants varied in age from younger than 15 to 65 or older. However, 59% of the participants were between 15 and 24. There was a variety in education, from primary school as highest education to doctoral degrees. However, 69% of the participants had finished their HBO, bachelor’s degree or a higher education. More details about the demographics of the participants can be found in Appendix VII. The participants were randomly assigned to one of the nine conditions. Figure 4 shows the number of participants in each condition.
33 Figure 4: Distribution of the number of participants on the different conditions.
CSI: working conditions - related Condition 1: N=26 Good CSR reputation CSI: environmental crisis - unrelated Condition 2: N=29 CSI: animal crisis – unrelated Condition 3: N=26 CSI: working conditions - related Condition 4: N=31 The company Bad CSR reputation CSI: environmental crisis - unrelated Condition 5: N=28 CSI: animal crisis – unrelated Condition 6: N=29 CSI: working conditions - neutral Condition 7: N=28 No CSR / Neutral CSI: environmental crisis - neutral Condition 8: N=26 CSI: animal crisis - neutral Condition 9: N=26
3.9 The procedure
The sampling technique that was used during this study was a ‘snowballing’ technique; participants received an invitation on Facebook with a link to the online experiment. These Facebook friends were asked to take part in the research. Subsequently, it was asked to share the message on Facebook in order to increase the range of Facebook members. The full questionnaire in Dutch can be found in appendix VIII and this questionnaire can also be found in English in appendix IX.
4. Results
In this section, the results of the study will be described. Visual inspection of the p-plots and histograms showes that all dependent variables were fairly normally distributed. See appendix X for a visual presentation of the distribution of the dependent variables by histograms.
4.1 Descriptives
Table 5 shows the means, standard deviations, correlations and Cronbach’s Alpha’s of all the continuous variables on a 7-point Likert scale. This table shows some interesting correlations. For example, it shows that there is a significant correlation between gender and the attitude toward the firm (r = 0,21; p<0,01). Apparently, men and women have different attitudes toward firms showing CSI (see 4.7). There is also a significant correlation shown between gender and perceived hypocrisy (r
(Prior domain CSR reputation: working conditions)
34 = -0,25; p<0,01). Additionally, there is a positive correlation found between the age of the participants and perceived hypocrisy (r = 0,16; p<0,01). Apparently, the older the participant is, the more they perceive firms showing CSI as hypocrite. The table also shows a negative correlation between education and credibility (r = -0,13; p<0,05). Apparently, the higher the education, the lower participants rate the credibility of the scenario. The higher educated participants might thus be more critical regarding the credibility of the scenario. Next, a high positive correlation between credibility and realism is shown (r = 0,62; p<0,01). The table also shows a negative correlation between attitude and expectancy (r = -0,16; p<0,05). Apparently, the less the participant expected the CSI, the higher is the attitude toward the firm. A negative correlation is found between hypocrisy and attitude (r = -0,14; p<0,05). This means that the more the firm is perceived as hypocrite, the more negative the attitude toward the firm is. The final interesting correlation is found between hypocrisy and expectancy (r = -0,34; p<0,01); the more the CSI behaviour was expected, the less it was seen as hypocrite.
Table 5: Means, Standard Deviations and Correlations (7-point Likert scale) M SD # of items 1 2 3 4 5 6 7 8 1. Gender 1,49 0,50 1 - 2. Age 2,75 1,16 1 0,02 - 3. Education 4,11 1,41 1 0,09 0,02 - 4. Attitude 2,86 1,24 3 0,21** -0,09 -0,02 (0,89) 5. Credibility 4,99 1,27 1 -0,09 0,05 -0,13* -0,13* - 6. Realism 5,10 1,24 1 -0,12 0,02 -0,08 -0,11 0,62** - 7. Expectancy 4,31 1,27 3 0,03 0,03 0,03 -0,16* 0,10 0,13* (0,94) 8. Hypocrisy 4,69 1,41 2 -0,25** 0,16** -0,12 -0,14* -0,01 0,02 -0,34** (0,80)
** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).
35
4.2 Controls
Credibility and realism
The first control of the experiment is the control of the perceived realism and credibility. As shown in table 5, the credibility and realism of the scenarios is relatively high (>4, above midpoint). Specific information about the credibility and realism of the different scenarios can be found in Appendix XVII.
Reputation manipulation check
To check if the manipulation of the CSR reputations of the firm worked out, the participants who read the good or bad CSR reputation scenario were asked to rate the firm on ethicality right after they read about the CSR reputation of the firm. The participants that read the scenario in which no information about the CSR reputation was given were not asked about the ethicality of the firm. Using an independent T-Test it is shown that there is a significant difference between the good and the bad reputation based on ethicality (F(167) = 0,11; p<0,01). Table 6 shows that the good CSR reputation is indeed seen as more ethical than the bad reputation. This means that the manipulation worked out in the right way.
Table 6: Means and Standard Deviations of the good and bad CSR reputation scenario’s based on ethicality on a 7-point Likert scale
M SD
Good CSR reputation scenario (N=81) 5,23 1,25
Bad CSR reputation scenario (N=88) 2,27 1,07
The SPSS output of the reputation manipulation check can be found in Appendix XII
4.4 Hypothesis 1
The first two hypotheses proposed contradictory effects of the valence of the CSR reputation on consumers’ attitude with regard to corporate irresponsible behaviour; it was hypothesized that corporate irresponsible behaviour would conduct an either less (H1A) ore more (H1B) negative
