CERTIFICATION LABELS SIGNALLING ETHICAL FASHION:
DIFFERENCES IN EFFECTIVENESS
ON CONSUMERS’ WILLINGNESS TO PAY
AND ORGANIZATIONAL EVALUATION
Thesis MSc Business Administration
Entrepreneurship and Management in the Creative Industries
Lesley Moelker (10448802) August 31st 2015, Amsterdam
Statement of originality
This document is written by Student Lesley Moelker who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents
Abstract Page 4
1. Introduction Page 5
2. Literature review Page 9
2.1. Signalling ethical fashion using certification labels Page 10 2.2. Organizational benefits of signalling ethical fashion Page 12 2.3. The signalling party: First Party and Third Party Certification Labels Page 15 2.4. Factors influencing the effectiveness of certification labels Page 16
2.4.1. Certification label reputation Page 17
2.4.2. Certification label credibility Page 18
3. Data and Method Page 23
3.1. Research design Page 23
3.2. Data collection Page 24
3.3. Sample Page 27
3.4. Variables Page 28
4. Results Page 31
4.1. Descriptive statistics Page 31
4.2. Correlations Page 33
4.3. Hypothesis testing Page 36
4.3.1. Certification labels’ effectiveness on consumers’ willingness to pay Page 36 4.3.2. Certification labels’ effectiveness on organizational evaluation Page 37 4.3.3. Differences in certification label reputation and credibility Page 38 4.3.4. Moderating effects – dependent variable: consumers’ willingness to pay Page 39 4.3.5. Moderating effects – dependent variable: organizational evaluation Page 40
4.3.6. Additional insight Page 42
5. Discussion Page 43
5.1. Hypotheses Page 43
5.1.1. Consumers’ willingness to pay Page 43
5.1.2. Consumers’ organizational evaluation Page 46
5.2. Managerial implications Page 48
5.3. Limitations and future research Page 49
6. Conclusion Page 51
7. References Page 52
8. Appendices Page 58
1 – Mock-up advertisements Veylinx Page 58
2 – Survey questions Veylinx (Dutch) Page 59
3 – Data analysis graphs Page 60
Abstract
Preceding research on the use of certification labels has predominantly been conducted in the empirical setting of the food industry. This study contributes to the literature by extending this knowledge to the dynamic empirical setting of the fashion industry. Additionally, this study adds to the literature by focusing on signaling ethical production methods opposed to the often-discussed increasing pressure on the environment.
This study investigates whether fashion organizations benefit from licensing a third party certification label, as a tool to communicate corporate social responsibility (CSR), compared to communicating through a self-created certification label, referred to as a first party certification label. Two possible organizational benefits retrieved from signalling CSR, an increase in consumers’ willingness to pay and organizational evaluation, are studied. Differences in effectiveness of diverging certification labels were tested through auctioning an ethically produced T-shirt via online auction platform Veylinx.
Results show that certification labels signaling CSR result in higher consumer willingness to pay, irrespective of the type of certification label used. Certification labels do however not affect consumers’ level of organizational evaluation. This study has highlighted the ineffectiveness of current certification systems, as no significant differences were found between first and third party certification label effectiveness. Therefore, it is concluded that fashion organizations do not benefit significantly more from licensing a third party certification label, compared to signaling CSR with a first party certification label.
Keywords – Corporate social responsibility, ethical fashion, willingness to pay, organizational evaluation, first party certification labels, third party certification labels, reputation, and credibility.
1. Introduction
The apparel industry has gone through a profound modification compared to just a few decades ago, when fashion designers used to present only two collections a year; spring/summer and autumn/winter. Todays’ fashion industry is generally defined by fast-fashion – a system that combines at least two components; short production cycles and distribution lead times, enabling brands to rotate their collections every two to three weeks, at low prices (BSR, 2012). This system enables fashion organizations to closely match fashionable supply with uncertain demand (Cachon & Swinney, 2011). Fortunately, we are witnessing a backlash against fast fashion, as the generation that grew up with fast fashion is becoming older and richer (Sull and Turconi, 2008). After years of fashion organizations striving for the lowest possible production prices, an increasing number of organizations and consumers are rethinking this widely accepted strategy in the fashion industry, realising that somewhere else in the world people are paying the real price. The uses of child labour, dangerous and unhealthy working conditions, and exploitation and harassment of workers in sweatshops, both sexual and/or physical, are no exceptions (Rudell, 2006). We are witnessing a trend towards consumer activism with respect to the “social behaviours” of organizations (Auger, Burke, Devinney & Louviere, 2003; McGoldrick & Freestone, 2008). Consumer purchasing behaviour is changing, as consumers increasingly express their social concerns through consciously choosing to buy from organizations that produce ethical fashion (De Pelsmacker, Driesen & Rayp, 2005). The term ‘ethical fashion’ refers to “fashionable clothes that incorporate fair trade principles with sweatshop-free labour conditions while not harming the environment or workers by using biodegradable and organic cotton” (Joergens, 2006).
In response to changing consumer demands and increasing awareness, which goes beyond merely looking at the environmental footprint (Auger et al., 2003), fashion organizations are increasingly investing in Corporate Social Responsibility (CSR) programs (Nan & Heo, 2007). The literature reveals at least two possible organizational benefits that can be retrieved from signalling CSR programs: an increase in consumers’ willingness to pay for products that are produced in conformity with a certain set of minimum standards (Basu & Hicks, 2008), as well as an increase in
organizational reputation (Fombrun & Shanley, 1990; Aguinis & Glavas, 2012), which translates into, among other things, an increase in consumers’ organizational evaluation (Sen & Bhattacharya, 2001; Nan et al., 2007; Brown & Dacin, 1997; Creyer, 1997).
In an attempt to increase awareness about CSR and educate consumers about the ethical dimensions of the products that they purchase (Auger et al., 2003), a wide variety of certification systems have been developed as a tool for signalling ethical fashion and helping consumers to make well-informed purchasing decisions. Burke, Eckert and Davis (2014) highlight the need to educate consumers by their research showing that nearly 60 percent of consumers respond negative, indifferent or ambivalent towards ethical purchasing. Those who were ambivalent, were often confused about what actually makes a product ethical and were additionally sceptical about how ethical, the often more expensive products, really are.
Prevailing studies now show a backlash against the high variety of certification systems. Research conducted by Jahn, Schramm and Spiller (2005) for example found up to fourty percent of fraud in organic food labelling in southern European countries. They claim that in situations of information asymmetry, organization directly involved in the production or sale of the product, as well as independent, external certification systems, may show opportunistic behaviour through green washing – meaning that they signal false information about their product quality – thereby misleading consumers. The large variety of certification labels currently available in the market has increased scepticism around certification labels, making it increasingly difficult for consumers to distinguish organizations actually producing ethical fashion from organizations only claiming to do so (Kirmani & Rao, 2000). The severity of this effect has been the motivation for this research, questioning whether fashion organizations benefit from licensing an expensive third party certification label as a tool to communicate CSR, compared to communicating through a self-created certification label, henceforward referred to as a first party certification label.
The following research questions will be tried to answer. Are certification labels effective tools for influencing consumers’ willingness to pay, as well as their organizational evaluation for producing ethically responsible? Does signalling ethical fashion through a third party certification label result in higher benefits compared to communicating ethical production methods themselves? Do
third party certification labels benefit higher perceived label reputation and credibility? And finally, does this influence consumers’ willingness to pay, as well as their organizational evaluation for producing ethically responsible?
A substantial amount of research has been devoted to consumers’ willingness to pay for ethically produced products, especially in the empirical setting of the food industry (Auger et al., 2003; Bougherara & Combris, 2009; Loureiro et al., 2002). Results have shown that consumers are willing to pay a price premium when they think that a food signals taste, or provides them with safety and health benefits (Bougherara et al., 2009). Studies on consumers’ willingness to pay for other ethically produced products have however shown more contradicting outcomes (Sen et al., 2001). More importantly, there remains inconsistency about whether or not certification labels have an actual positive affect on consumers’ willingness to pay for ethical products (Loureiro et al., 2000; Rudell, 2006; De Pelsmacker et al., 2005). Results on consumers’ willingness to pay can often be considered questionable, as many studies have gathered data by using questionnaires. Additionally, research has proven that people tend to overstate their hypothetical willingness to pay, relative to when actual money is being used, especially in case of important product features, like ethical features, to which socially correct answers exist (List & Gallet, 2001; Ellis, McCracken & Skuza, 2012; Auger et al., 2003). Therefore, the experiment in this study will by conducted through Veylinx, an online auction platform.
The effectiveness of signalling CSR on organizational reputation has been a consistent finding in the literature (Aguinis et al., 2012). However, little research has been conducted on how to best communicate CSR (Du, Bhattacharya & Sen, 2010) and the effectiveness of informing consumers through signalling certification labels. Due to time restraints of this research, organizational evaluation for producing ethical fashion is measured instead of organizational reputation, as reputation is built overtime (Herbig and Milewicz, 1995).
This paper is structured along the following lines. First an extensive literature review will be presented in the theoretical framework, subsequently leading to the hypothesis. Next, the research methodology and data collection method of the experimental study will be described. Afterwards, the
findings of the experiment will be presented and thoughtfully discussed. Finally, practical implications of the results and future research suggestions will be presented, followed by an overall conclusion.
2. Literature Review
A growing consumer segment is increasingly concerned about whether of not apparel is made in an ethical way (Uusitalo et al., 2004). Consumers have the power to increase organizational concern for ethical production processes by changing their purchasing behaviour through consciously choosing to buy ethically produced products over conventionally produced products. This is referred to as ‘ethical consumerism’. Ethical consumerism is not only concerned with the environmental impact of production processes, but also includes “the ethical and moral aspects present in production and delivery of goods, for example, the use of child labour, suppressing or preventing labour unions, and testing on animals” (Uusitalo et al., 2004).
Fashion organizations can choose to respond to ethical consumerism by investing in Corporate Social Responsibility (CSR) programs. By communicating their CSR acts, organizations can signal their ethical and sustainable business practices to consumers (Porter & Kramer, 2006; Nan et al., 2007; Crane, 2001). According to The World Business Counsil for Sustainable Development (WBCSD, 1999, p. 3) “Corporate Social Responsibility is the continuing commitment by business to contribute to economic development while improving the quality of life of the workforce and their families as well as of the community and society at large.” (WBCSD, 1999, p.3). According to Perry and Towers (2009) any CSR definition should stress that “action must go beyond immediate profit-maximization interest of the organization and beyond merely obeying the law or other regulations”. In the fashion industry, apparel that is produced in an ethical and sustainable way is referred to as ‘ethical fashion’; fashionable clothes produced from organic cotton and in accordance with fair trade and labour standards, thereby taking into account the environment and worker conditions (Joergens, 2006).
The first section of this literature review will elaborate on how certification labels are used to signal ethical fashion. After, two possible organizational benefits of signalling ethical fashion will be discussed, followed by different types of certification labels. Finally, two possible factors influencing the effectiveness of certification labels, label reputation and credibility will be evaluated.
2.1 Signalling ethical fashion using certification labels
Consumer judgement of ethical organizational conduct is a multifaceted process and especially complex in the fashion industry. The ability of consumers to evaluate product quality and characteristics varies across products. The literature makes a distinction between four different products. First of all, information economics claim that some products are search goods; as for those products consumers can easily assess product features and characteristics before purchase. Second, products for which quality characteristics are difficult to be evaluated prior to purchase are called experience goods (Nelson, 1970). Credence goods are products for which its quality can neither be assessed before nor after purchase (Darby and Karni, 1973). External organizations and public authorities do have the ability to observe these products’ attributes, however at high costs. Therefore, consumers have to rely on their judgement of the product, transforming the credence attribute into a ‘quasi-search attribute’ (Jahn et al., 2005). Finally, and of relevance to this study, are the unobservable process-oriented product attributes of products, which are unable to be analysed at the end-product level, neither by consumers nor by external organizations. These process-oriented attributes are referred to in the literature as Potemkin attributes and include, among others, organic production and fair trade production processes (Jahn et al., 2005; Moussa & Touzani, 2008). Potemkin attributes form a particular problem in the fashion industry, due to complex and obscure global subcontracting relationships (Perry et al., 2009).
Organizations producing Potemkin goods, such as apparel, usually possess more information about the production process than consumers do (Moussa et al., 2008). In the literature, this unfair information distribution - in which consumers are less informed and face higher uncertainty about an organization’s commitment to delivering desirable product features than the organization’s managers - is referred to as information asymmetry (Grossman & Stiglitz, 1980; Moussa et al., 2008). Constraint by time, effort and information available, consumers seek to maximize the available information through social construction processes that enables them to evaluate and form an opinion about the social properties of an organization (Deephouse et al., 2005; Bitektine, 2011). Additionally, a general perception in the literature is that the more uncertainty stakeholders face, the more likely it is that their attitudes will be influenced by external information cues in order to maximize the available product
information, reduce the uncertainty that they face in evaluating relevant product attributes (Weigelt & Camerer, 1988; Rindova et al., 2005; Fombrun & van Riel, 1997; Fischer & Reuber, 2007; Erdem & Swait, 2004) and to be better able to evaluate and form an opinion about the social properties of an organization (Deephouse & Carter, 2005; Bitektine, 2011). Therefore, scholars working from the economics perspective propose that firms can reduce quality information gaps through signalling (Rindova et al., 2005).
Signals are marketer-controlled, easy-to-acquire information cues, extrinsic to the product itself (Bloom & Reve, 1990) and can be communicated through “price, brand name, advertising spending, warranties, money-back guarantees, retailer reputation, and quality and certification labels” (Moussa et al., 2008; Fombrun et al., 1990; Shapiro, 1983). They are used as information cues by consumers to make relational assumptions about product quality and value (Bloom et al., 1990; Moussa et al., 2008) and an organization’s intentional future behaviour (Rao, 1994; Herbig & Milewicz, 1995).
In order to support consumers in making conscious product choices, the use of certification labels to decrease information asymmetry is a widely supported method (Caswell & Mojduszka, 1996; Basu et al., 2008; Caswell & Padberg, 1992; Moussa et al., 2008). Certification labels can be used as a tool to address the issues of scepticism around ethical behaviour (Boulstridge & Carrigan, 2000) and may act as a filter through which consumers can interpret the standards along which an organization’s products are produced, thereby assessing specific product qualities (Graffin & Ward, 2010; Caswell et al., 1992; Loureiro, 2000). Ethical production standards may include minimum wages, reasonable working hours and conditions, safe working environment, no exposure to toxic chemicals, no use of child labour and a guaranteed price floor for poor producers (Basu et al., 2008). A study by Marymount (1999), interviewing ethically conscious consumers, found that 56 percent of the respondents preferred the use of certification labels to a list of stores and companies (33 percent), as a tool to avoid buying clothes that were produced in sweatshops. In response to this demand, both fashion and production organizations themselves, as well as independent third parties, have designed a wide variety of quality and certifications labels, as the answer to the lack of information transparency regarding ethical behaviour in production processes (Basu et al., 2008).
2.2 Organizational benefits of signalling ethical fashion
As already stated before, fashion organizations are increasingly involved in CSR acts. Former research mainly focussed on the financial payoff gained from CSR, as it was believed that an organization’s economic and social objectives could not combined (Boulstridge et al., 2000). However, as it is becoming increasingly apparent that these concepts are inseparable (Porter & Kramer, 2002; Werther & Chandler, 2005), more recently studies are also investigating the effects of CSR on the reactions of consumers and other stakeholders (Sen et al., 2001; Brown et al., 1997). This section will discuss these two possible organizational benefits of signalling CSR programs – an increase in consumers’ willingness to pay and organizational evaluation - in more detail.
The effect of signalling CSR on consumers’ willingness to pay – The majority of preceding research
has measured the relationship between CSR programs and financial performance by comparing organizations on the Fortune Corporate Reputation Index (Stanwick et al., 1998). Empirical evidence for the financial effects of signalling CSR behaviour appears to be equivocal (Sen et al., 2001). Some studies have concluded that the benefits do not outweigh the financial costs (Stanwick & Stanwick, 1998). However, the majority of research shows an, at least weak, positive link (Boulstridge et al., 2000; Sen et al., 2001; Stanwick et al., 1998). The method used to study this relationship is however questioned, as the Fortune Corporate Reputation Index “is heavily weighted based on the financial position of the firm” (Stanwick et al., 1998, p. 196). This study will measure the effectiveness of signalling CSR, more specifically ethical fashion, on financial performance by investigating consumers’ willingness to pay for ethically produced apparel.
Studies on consumers’ willingness to pay for eco-labelled products have often been conducted in the organic food industry. It is often found that consumers buy organically labelled food, as they assume these products to provide them with health and safety benefits. A study by Loureiro et al. (2002) found that eco-labeled apples had a significant positive affect on the probability of consumers willing to pay a price premium, as these apples were believed to be of higher quality. The actual increase in willingness to pay was however rather small, with an increase of five percent. Research by Bougherara and Combris (2009) on the other hand, found that the willingness to pay for eco-labeled
orange juice did not significantly differ from non-labeled orange juice. Additionally they found that consumers do not buy eco-labelled food products for higher perceived taste or food safety attributes, but other motives.
A handful of research has specifically been conducted on consumers’ willingness to pay for ethically produced apparel (Ellis et al., 2012). Auger et al. (2003) found that consumers are willing to pay significantly more for ethically produced products (soap and athletic shoes) than was previously believed. This is supported by Shen, Wang, Lo, and Shum (2012) who found that nowadays, consumers are actually more willing to pay a price premium for socially responsible fashion compared to environmentally responsible fashion. A survey conducted under UK consumers by McGoldrick et al. (2008) showed an average increase in consumers’ willingness to pay for ethical fashion of 11,2%. Also, research by Casadesus‐Masanell, Crooke, Reinhardt and Vasishth (2009) showed a substantial increase in willingness to pay of $6.58 for T-shirts produced from organic cotton, compared to conventional produced shirts. However, the outcomes of both studies should be interpreted with caution, as both studies have been based on the unconstraint, stated willingness to pay, which has proven to be a typical overestimate of the actual willingness to pay (Millock & Hansen, 2002; Ellis et al., 2012; McGoldrick et al., 2008). More recent studies have been estimating willingness to pay through product auctions. Studies by Hustvedt and Bernard (2008), investigating willingness to pay for organic cotton socks when auctioned, and Ellis et al. (2012), auctioning T-shirts produced from organic cotton, also found a significant price premium of respectively $1.86 and an increase in willingness to pay of twenty-five percent. Elliott and Freeman (2001) found that, depending on the product price, consumers were willing to pay up to twenty-eight percent more for $10 products and up to fifteen percent more for $100 products when produced ethically responsible. Finally, Hustvedt and Bernard (2010) found that consumers were willing to pay significantly more for apparel when provided with information about labour-related attributes. Therefore:
Hypothesis 1: Signalling ethical fashion using a certification label will increase consumers’
The effect of signalling CSR on organizational evaluation for producing ethically responsible - Apart
from the possible increase in consumers’ willingness to pay as a result of signalling ethical fashion, more recently studies show a second possible organizational benefit of signalling ethical organizational behaviour, based on the inclusion of the reactions of specific stakeholder groups such as employees and consumers (Sen et al., 2001). Research has shown that CSR programs have the ability to positively influence organizational reputation (Fombrun et al., 1990; Stanwick et al., 1998), which translates into, among other things, favourable organizational evaluations by consumers (Nan et al., 2007; Brown et al., 1997; Creyer, 1997; Sen et al., 2001). Literature therefore states that CSR programs can be used by organizations to differentiate itself and its products from competitors (Brown et al., 1997; Sen et al., 2001).
Research by Brown et al. (1997), specifically studying the effect of an organization’s CSR acts on consumers’ organizational evaluations, found that CSR behaviour showed to have a positive effect on consumers’ organizational evaluation for producing ethically responsible and additionally results in being a favourable choice for future purchases. Furthermore, they concluded that communicating CSR behaviour is effective because “unless consumers know about it, it may do little good or have little effect on product responses” (Brown et al., 1997, p.81).
Qualitative research by Boulstridge et al. (2000) has shown that even though many organizations are expecting positive outcomes of promoting their corporate reputation, the actual consumer knowledge on corporate responsible behaviour is often small. They therefore conclude that the positive returns of CSR behaviour might not be as effective as anticipated, due to a general lack of knowledge on ethical production methods by the average consumer. This study however proposes that signalling ethical fashion by using certification labels can be used as a tool to educate and inform consumers. Therefore:
Hypothesis 2: Signalling ethical fashion using a certification label will positively increase consumers’
2.3 The signalling party: First Party and Third Party Certification Labels
In essence, two types of certification labels can be distinguished based on the party issuing the label. Certification labels created and communicated by the organization that is directly involved in the production or sale of the product (Moussa et al., 2008) are from now on referred to in this paper as first party certification labels. Certification labels created by an independent third party such as an external, independent organization or public authority (Starobin & Weinthal, 2010), are henceforward referred to as third party certification labels.
Both, first party certification labels and third party certification labels, can be classified as sale-independent, default-independent signals, meaning that they involve upfront money loss; irrespective of whether the products carrying the label will be sold or not (Kirmani et al., 2000). The associated costs for signalling ethical fashion with a first party certification label are relatively low, as a first party certification label does not involve a licensing agreement with an external, independent issuing party. However, when signalling with a first party certification label, there will be no independent party monitoring the organization’s claims. The lack of control by an external party does provide the opportunity for greenwashing (Cason & Ganghadaray, 2002), meaning that the fashion organization may not actually fulfil its claims to be producing according to certain ethical standards. Third party certification labels, on the other hand, do require fashion organizations to apply for a licence (Graffin et al., 2010), which will only be issued once the fashion organization conforms to the standards set by the third party (Basu et al., 2008). Third party certifiers often guarantee inspections on Potemkin attributes, covering the whole production process and supply chain (Jahn et al., 2005). They increase transparency and provide organizational evaluations and endorsements to inform consumers (Graffin et al., 2010; Quagrainie et al., 2003; Rao, 1994). Licensing a third party certification label is often seen as a strategic choice used for differentiation. It is a resource that is costly and sometimes unavailable to obtain by competitors (Rindova et al., 2005), in that case offering a competitive advantage (Uusitalo et al., 2004).
The literature often refers to third party certification labels as superior in evaluating organizations compared to first party certification labels. Rao (1994) found that third party endorsement in the automobile industry increased an organization’s performance in terms of better
access to resources and extended lifecycle chances. Research by Graham and Bansal (2007) affirmed this, by finding a significant positive effect of third party endorsement on an airline company’s reputation and consumers’ willing to pay a premium of $36 for a plane ticket. Third party certification labels allow consumers to reveal a higher willingness to pay for those products that they believe to have been produced in conformity with a certain set of minimum social standards (Loureiro, 2002; Basu et al., 2008). According to Walster, Aronson and Abrahams (1966) any signal “will be more effective and will be seen as more credible when it is arguing for a position opposed to his own best interest”. This is confirmed by research of Larceneux (2001), who found that a certification label must come from an external objective party, foreign to the manufacturer, competent and not at all interested in the sale of the product bearing the certification label. This leads to the following hypotheses:
Hypothesis 3: A third party certification label has a greater positive effect on willingness to pay than a
first party certification label.
Hypothesis 4: A third party certification label has a greater positive effect on consumers’
organizational evaluation for producing in ethically responsible than a first party certification label.
2.4 Factors influencing the effectiveness of certification labels
Even though third party certification labels are often identified as superior when compared to first party certification labels, not all third party certification labels are equally effective and reliable. Merely providing consumers with a third party certification labels might not always succeed in reducing information asymmetry (Moussa et al., 2008) “as they struggle to process the new information and integrate it with their prior knowledge and experiences” (Rudell, 2006, p. 292). Consumers might question the labels’ message or intension, or face uncertainty about the standards against which the quality of a product is assessed (Graffin et al., 2010). Just like first party certification labels, third party certification label systems can be sensitive to opportunistic behaviour and greenwashing, when control is not regular, when the organization to be inspected can choose its
own auditor or when inspection procedures and occupational qualifications have not yet been sufficiently well defined by the third party (Jahn et al., 2005).
A certification label’s effectiveness in changing consumers’ believes attitudes or behaviour may, among other things, be dependent on the credibility of the label issuer and the amount and quality of information available (De Pelsmacker et al., 2005). The following paragraphs will therefore examine first and third party certification label reputation and credibility in more detail.
2.4.1. Certification label Reputation
According to the Macmillan English Dictionary (2007), reputation is “the opinion that people have about how good or how bad someone or something is”. The definition in academic literature is however more equivocal, as reputation is a multidimensional construct (Bitektine, 2011), extensively studied by different academic and practitioner literatures (Deephouse, 2000; Fombrun, 1997; Rindova et al., 2005; Fombrun et al. 1997). As this paper is exploring the reputation of a product, the certification label, reputation will be discussed as described by signalling theorists in the economics perspective.
Signalling theorists claim that reputations are information signals (Fombrun et al., 1997; Herbig et al. 1995), emphasizing the perceived quality dimension (Shapiro, 1983; Milgrom & Roberts, 1986; Weigelt & Camerer, 1988; Rindova et al., 2005) of product- or organizational features that are hidden from consumers’ view (Herbig et al., 1995; Rindova et al., 2005). Reputation of product quality is the consumers’ judgement of the “overall excellence, esteem or superiority of a product relative to alternative products” (Moussa et al., 2008, p. 529) and is built by fulfilling its communicated signals, generally assessed through past performance (Sobel, 1985; Weigeld et al., 1988; Herbig et al., 1995; Graffin et al., 2010). Consumers must be able to interpret the quality of earlier performance as signals to indicate future quality (Herbig et al., 1995; Shapiro, 1983). Reputation therefore can also be described as the “estimation of consistency over time, based on an organizations’ ability and willingness to perform an activity repeatedly in similar fashion” (Herbig et al., 1995). According to Jahn et al. (2005), a certification labels’ reputation can only be built if the
certifier is able to revealing critical aspects and opportunistic behaviour. The inability of consistency will eventually lead to a decrease in reputation.
In the case of first party certification labels, issued by the organization that is directly involved in the production or sale of the product, the reputation of the certification label is converted from the organization that has created and is signalling the label. In the case of an unknown fashion organization, the perceived reputation of the label can thus be expected to be low. Perry et al. (2009) rightly questioned whether small and medium-sized enterprises’ reputation is of sufficient magnitude to enable credible self-certification. Although reputation can be build over time by delivering upon their promise, reputation adjustment can only be achieved with a time lag (Shapiro, 1983).
A third party certification label, on the other hand, can be seen as a rented reputation from an independent, external certifying body. Fashion organizations can license a third party certification label once they confirm to the standards set by the third party and are subsequently allowed to signal their ethical organizational behaviour by using the certification label. Organizations licensing a third party certification label can benefit from the so called “halo effect”, taking advantage from the instant recognition of the third party’s established reputation. Consistency is key in building a favourable reputation, as “one bad egg may as well spoil the entire basket” (Herbig et al., 1995). Therefore, third party issuers place high demands on quality control, as all certified organizations are an extension of the third party certifiers’ reputation. It is therefore hypothesized that:
Hypothesis 5: Third party certification label reputation will be higher than first party certification label
reputation.
2.4.2 Certification label Credibility
Apart from a certification label’s reputation, consumers also consider the believability of the certification label when trying to distinguish the low from the high quality labels (Kirmani et al., 2000; Erdem et al., 2004). In the literature, believability of signalled information is referred to as credibility (Herbig et al., 1995; Starobin et al., 2010; Erdem et al., 2004; Hovland & Weiss, 1951; Fogg, Marshall, Laraki, Osipovich, Varma, Fang, … and Treinen, 2001). Herbig et al. (1995) define
credibility as consumers’ ability to rely on what a signalling party claims it will do, at a specific moment in time. Therefore, for a certification label to be credible, it must be “verified in a consistent manner” (Starobin et al., 2010), delivering upon the promised quality (Herbig et al., 1995).
The establishment of credibility is a dynamic process (Herbig et al., 1995). Just like reputation, when talking about credibility, one is talking about the perceived credibility by consumers (Fogg et al., 2001; Herbig et al., 1995). A credible signal has the ability to influence consumers’ acceptance of an information signal (Eisend, 2006; Yoon, Kim & Kim, 1998), and can thereby decrease information asymmetry. The higher the credibility of a signal, the more persuasive it will be (Herbig et al., 1995) in changing consumer’s perceptions, attitudes, behavioural persistence and purchasing intentions (Dentoni, Tonsor, Calantone & Peterson, 2014; Ohanian, 1990; Pornpitakpan, 2004; Yoon et al., 1998). Additionally, when a signalling party’s communicated message is perceived as credible, it will subsequently increase an organization’s reputation (Herbig et al., 1995; Moussa et al., 2008). Perceived credibility increases when the signalling party’s actions agree with its communicated intentions; when it consistently provides accurate and reliable information (Sobel, 1985; Herbig et al., 1995). Credibility decreases if actions and intentions are inconsistent (Herbig et al., 1995).
The concept of credibility has received much attention from scholars and practitioners in the marketing field (Fogg et al., 2001). Scholars have agreed that credibility perceptions are formed by evaluating multiple dimensions simultaneously (Fogg et al., 2001), leading to many operationalizations of source credibility in the literature (Ohanian, 1990). Studied components of credibility that have been identified by factor-analytic studies (Pornpitakpan, 2004) are, among others, physical attractiveness (Ohanian, 1990; Pornpitakpan, 2004), objectivity (Starobin et al., 2010; Homer & Kahle, 1990), status (Pornpitakpan, 2004; Erdem et al., 2004), sincerity and accuracy. However, the majority of scholars agree on identifying two main components of credibility: expertise and trustworthiness (Erdem et al., 2004; Starobin et al., 2010; Homer et al., 1990; Ohanian, 1990; Fogg et al., 2001; Eisend, 2006; Pornpitakpan, 2004; McGinnies and Ward, 1980).
Extensive research has been conducted on whether high- or low credibility signals are more effective, or whether they will have no different effects in changing attitudes and behavioural
compliance of consumers (Pornpitakpan, 2004). Low credibility sources can be used by organizations that have existed long enough in the market so that consumers have experienced the product first hand and show a positive attitude towards the brand. However, it is generally well supported in the literature that high credibility signals have a stronger positive effect on causing positive attitude change and behavioural intentions compared to low credibility signals (McGinnies et al., 1980; Yoon et al., 1998; Sternthal, Dholakia & Leavitt, 1978), especially when consumers have no prior experience with the product (Ohanian, 1990; Pornpitakpan, 2004).
McGinnies et al. (1980) found that the most favourable condition for credibility to be effective is when consumers consider the signal to be both expert and trustworthy. Although researchers have not yet fully agreed on the relative importance of trustworthiness and expertise (Erdem et al., 2004), expertise appears to be the most persuasive component of credibility (Homer et al., 1990; Rindova et al., 2005). This dimension is also indicated in the literature as authoritiveness, competence, expertness, experience or qualification (Pornpitakpan, 2004). It reflects the extent of which the signalling party is believed to be capable of making valid assessments (Fogg et al., 2001; Ohanian, 1990; Homer et al., 1990) by virtue of having relevant knowledge, experience, skills and competences (Moussa et al., 2008; Homer et al., 1990). Thus, for a signalling party to be credible, the issuer must be an expert in evaluating the attributes of a product (Lanceneux, 2001). Previous studies have often measured expertise among the dimensions of trained-untrained, informed-uninformed and educated-uneducated (Ohanian, 1990).
Trustworthiness is consumers’ believe that the source provides honest information, without motivation for manipulation or deception (Eisend, 2006). It reflects consumers’ believe that the signalling party intends to tell the truth (Eisend, 2006; Ohanian, 1990; Homer et al., 1990) and captures the “perceived goodness or morality of the source” (Fogg et al., 2001). In order to be trustworthy, the signalling party is said to be independent of the awarding organizations and not have any interests in the sale of the product-bearing label (Larceneux, 2001). This is supported by research by Dentoni et al. (2014) who found that consumers did not choose to buy beef endorsed by the government, as the government was perceived as having too high vested interests.
Research by Larcenaux (2001) concludes that signal credibility is likely to depend on the awarders’ credibility. Subsequently, first party and third party certification label credibility need a different theoretical approach. First party certification label credibility must be regarded as corporate credibility (Pornpitakpan, 2004). When talking about corporate credibility, the organization that produces the product is seen as the communication source. Corporate credibility is based on the organizations’ reputation, the sum of its past behaviour (Erdem et al., 2004; Herbig et al., 1995) and its reputation for keeping its intentions (Herbig et al., 1995). It is its perceived level of expertise to fulfil its claims, and whether or not the organization can be trusted to tell the truth or not, that make up its perceived credibility (Newell & Goldsmith, 1997). First party certification labels are often perceived as low credibility signals, due to credibility problems for private information sources (Shapiro, 1982). Quality signals communicated by first party certification labels are often expected to be self-interested and therefore lacking credibility (Erdem et al., 2004). They are thought to be untrustworthy, as their positive claims are not inspected and controlled by independent third parties (Larceneux, 2001; Erdem et al., 2004). However, research by Sternthal et al. (1978) did find that only when consumers were already favourably predisposed towards an organization, a moderately credible source could induce more agreement and support than a high credible source, as it gave consumers the feeling that they could make their own choice.
Third party certification label credibility must be regarded as source credibility (Pornpitakpan, 2004). Source credibility refers to the extent to which the source endorsing the organization is perceived as having relevant expertise and can be trusted to give an objective opinion on the subject (Ohanian 1990). Source credibility has the ability to favourably enhance the believability of an organization’s message (Ohanian, 1990). The often-included attractiveness component of source credibility, used to measure perceived attractiveness of a human endorser, will not be taken into account (Yoon et al., 1998; Ohanian, 1990), as the attractiveness of a certification label expressed by its design characteristics will not be investigated in this study. Third party certification labels are often perceived as more credible than first party certification labels. De Pelsmacker et al. (2005) found that for a fair-trade coffee label to be perceived as credible, it needs to be “efficiently monitored and subjected to third-party certification”. By virtue of their specialization in collecting and disseminating
information, they are likely to be viewed as having superior access to information and/or expertise in evaluation organizations (Homer et al., 1990). Results by Graham et al. (2007) support this as they found that the endorsement of airline companies by the Federal Aviation Administration, assessing airline safety, had the largest influence on perceived organizational reputation, as it provided consumers with independent confirmation of an airlines’ expertise and worthiness. Therefore:
Hypothesis 6: Third party certification label credibility will be higher than first party certification label
credibility.
Additionally, this paper hypothesizes that third party certification label reputation and credibility will have a positive influence on the effectiveness of certification labels, thereby further increasing consumers’ willingness to pay and organizational evaluation for producing ethical fashion. This leads to the final hypotheses:
Hypothesis 7a: Third party certification label reputation will increase the effectiveness of a third party
certification label on willingness to pay, compared to a first party certification label.
Hypothesis 7b: Third party certification label credibility will increase the effectiveness of a third party
certification label on willingness to pay, compared to a first party certification label.
Hypothesis 8a: Third party certification label reputation will increase the effectiveness of a third party
certification label on organizational evaluation for producing ethical fashion, compared to a first party certification label.
Hypothesis 8b: Third party certification label credibility will increase the effectiveness of a third party
certification label on organizational evaluation for producing ethical fashion, compared to a first party certification label.
3. Data and Method
This section will elaborate on the empirical setting of the research and data collection. First, the research design will be presented. After, the data collection method and sample will be discussed. Finally, all variables measuring different constructs in this study will be operationalized.
3.1 Research design
This research’s empirical setting is the fashion industry and aims to explain the relationship between the use of different certification labels signalling ethical fashion and two possible organizational benefits of CSR: an increase in consumers’ willingness to pay and organizational evaluation for producing ethically responsible.
The majority of preceding studies on consumers’ willingness to pay were conducted by using surveys. Elliot et al. (2004) state that surveys invariably find that the majority of people would choose a garment produced under better conditions over conventional production methods, even at small additional costs. On the other hand, evidence exists that stated willingness to pay, expressed by a simple, unconstraint attempted to quantify the monetary value of these consumer preferences, are typically overestimates of the actual willingness to pay (Millock et al., 2002; Ellis et al., 2012; McGoldrick et al., 2008). Surveys conducted on issues that provoke socially desirable answers result in more positive attitudes measured than actual behaviour, as respondents are more likely to be willing to make a good impression on the researcher (De Pelsmacker et al, 2005; King & Bruner, 2000). As a result, respondents additionally tend to overstate their hypothetical willingness to pay relative to when actual money is being used (List et al., 2001; Ellis et al., 2012; Auger et al., 2003).
Therefore, over the past 35 years, an increasing amount of studies have been conducted by using experimental auctions (Bougherara et al., 2009). Finally, the majority of preceding studies covering this topic used surveys that were conducted under a sample of merely students. This makes the level of representativeness of the population doubtful (Ellis et al., 2012).
Choosing an experiment as research strategy enables manipulation of the independent variable while holding others constant. This makes enables researchers to observe the effects of changes in the
independent variable. In order to be able to analyse differences in effectiveness of diverging certification labels, this research will collect its data though an online experimental auction.
3.2 Data collection
Veylinx - In order to test if and how the independent variable in this study – diverging
certification labels – effect the two dependent variables (consumers’ willingness to pay for ethically produced apparel and organizational evaluation), this study’s experiment is conducted in collaboration with online auction platform Veylinx. Veylinx is build according to one of the most popular auction methods named the Vickrey second-price auction (Vickrey, 1961). In a Vickrey auction, the final winner will be the participant that placed the highest bid, but will only be paying the second-highest bid price (Bougherara et al., 2009). This means that the winning participant will always be paying less than his or her indicated willingness to pay. As Veylinx hides biddings from other participants, the experiment controls for collusion or affiliation (Harrison, Harstad & Rutström, 2004) and competition (Bougherara et al., 2009).
Veylinx works as follows: consumers can sign-up at wwww.veylinx.com. Once they are a member, they will be sent a notification by email every time that they are identified as a target consumer for a specific product that is being auctioned. By clicking the link provided to them in the email, they will enter the auction website and are able to place a bid on the product that is presented to them, either on a picture or in a movie clip. After having placed their bid, representing their willingness to pay for that specific product, participants are asked to answer a maximum of five survey questions related to the auction.
Veylinx is a very convenient data collection platform for researchers. The ability to add survey questions after the auction is of great value to researchers, as it provides them with more consumer insights. Additionally, Veylinx gives researchers access to their sizable database, providing extensive data on participants’ consumer behaviour. It collects all data retrieved from preceding auctions, providing researchers with additional information like age, gender, education, as well as for example a participant’s interest in organic products. In return for using Veylinx, students and researchers must recruit an additional 100 members before running their experiment.
Stanley and Stella – Ethically produced T-shirts from the brand Stanley and Stella were auctioned
through Veylinx. Basic, white T-shirts were chosen over coloured shirts in order to be able to reach as many auction participants as possible. Stanley and Stella is a brand pioneering in the imprint market and only sell their products business-to-business. Therefore it is most probable that the brand was unknown to auction participants and previous experience with the brand was unlikely. By choosing this brand, the chances of the results being influenced by the brand name were minimalized. Additionally, this brand was chosen because of its highly sustainable production process: from the selection of fibres, manufacturing, and distribution to packaging. Finally, they are one of the few fashion brands that is licenced more than one certification label: they signal CSR with both the Fair Wear Foundation label and Global Organic Textile Standard (GOTS) label.
Treatments – The Fair Wear Foundation is an independent, non-profit organization that works together
with fashion organizations and factories to improve labour conditions for garment workers. The Fair Wear Foundation has created the ‘Code of Labour Practices’ as a standard to which all members have to comply (http://www.fairwear.org/22/about/), before being allowed to signal their membership. In order to check the commitment of organizations and factories to the standards set by the Fair Wear Foundation, they have created a verification system active on three levels: factory level, complaints procedure and company level. Factory-level verifications included both announced factory visits as well as offsite interviews with workers prior to the factory visit. The complaints procedure is enabled by a local complains handlers active in every country where the Fair Wear Foundation is operating. Finally, the Fair Wear Foundation usually visits companies after one year of membership.
According to the website of GOTS, the GOTS standard “stipulates requirements throughout the supply chain for both ecology and labour conditions in textile and apparel manufacturing using organically produced raw materials” (http://www.global-standard.org/about-us/visionmission.html). Selected professionals and independent certification bodies execute verification of licensee’s commitment to the environmental and social standards set by GOTS through onsite inspections.
The Dutch organization Milieu Centraal, who provides independent, reliable and practical information about environmental conscious products, and aims to help consumers making sustainable
choices, has developed a certification label guide evaluating diverse certifications used in the fashion industry. They evaluate certification labels on four pointers: environment, animal welfare, fair trade and control. Possible scores rank from 0 – 5, with 0 being not applicable, and 5 being very good (http://keurmerken.milieucentraal.nl/keurmerken/inhoud/kleding/kleding). According to Milieu Centraal, the Fair Wear Foundation label scores no points on environment and animal welfare, but does score 3 points (moderate) on fair trade, as well as 3 points on control. The control system is evaluated as moderate, as the Fair Wear Foundation trains external inspectors, which leads to a so-called two party certification (http://keurmerken.milieucentraal.nl/keurmerken/inhoud/kleding/ kleding/fair-wear-foundation). Milieu Centraal evaluates the GOTS label as a much better label, as it scores 3 points on environment (moderate), 4 points on fair trade (good) and even 5 points (very good) on control.
In order to evaluate if and how different types and qualities certification labels, both first and third party, differ in their effectiveness to influence consumers’ willingness to pay, as well as their organizational evaluation of Stanley and Stella, diverging certification labels will be tested. The experiment is divided into four treatment groups, enabling controlled tweaking of the advertisements presented to participants on Veylinx. The experimental treatments only differ in the certification label that was signalled. The picture of the shirt, the advertisement texts, colours used and brand name were kept unchanged and identical (see appendix 1).
A baseline was created to be able to compare the effects of different certification labels. This was done in treatment 1, by not using any certification labels. To test the effect of using a first party certification, which, as already explained in chapter 2, means that the certification label is created and communicated by the organization directly involved in the production and/or sale of the product, a first party certification label was designed for the brand Stanley and Stella (see appendix 1, treatment 2). The final treatments include the two certification labels that Stanley and Stella has been issued; the Fair Wear Foundation label (treatment 3) and GOTS label (treatment 4).
Table 1 – Experimental treatments. See appendix 1 for mock-up advertisements
3.3 Sample
In collaboration with Veylinx, the data collected for this experiment could be retrieved from a large sample. Participants in this study were, as already explained, a member of Veylinx. In total, 765 members opened the auction by clicking on the link in the email sent to them. The male/female ratio of the sample is represented by respectively 57,8% and 42,2%. In comparison to the male/female ratio in the Dutch population (49,5% men, 50,5% women), the relatively high percentage of male participants might be explained by the product that was being auctioned: a basic white T-shirt might be more popular amongst men than women. Noticeable is the high percentage of participants in the age range of 20-40 years and 40-60 years old, respectively representing 48,1% and 38,6% of the sample. An explanation for this occurrence could be the obligatory recruitment of 100 new members to Veylinx before running an experiment, as many students use Veylinx as an instrument to collect their research data. This assumption is supported when looking at the relatively high percentage of students participating in the experiment (18% compared to 7% of the total Dutch population being a student). Looking at the highest level of education completed by the auction participants also confirms this, as 39,4% of the sample has completed a WO education, compared to 21% of the total Dutch population. Table 1 shows an overview of key demographics of the sample compared to the Dutch population.
When looking at the four different treatment groups, it can be concluded that there are no huge differences between the division of men and women, average age and level of education. It is therefore appropriate to compare the four conditions (table 3).
Table 3 – Comparison of treatments
3.4 Variables
Dependent variables – This research investigates two dependent variables; willingness to pay and
organizational evaluation for producing ethically responsible.
First, the dependent variable willingness to pay was measured by the biddings placed by auction participants, measured in eurocents. Bidding amounts of 0,00 euros were accepted. The variable was labelled ‘bid_amount’ for statistical analysis.
Second, the dependent variable organizational evaluation for producing ethically responsible was measured by one of the survey questions asked through Veylinx after participants had indicated their willingness to pay for the T-shirt by means of their bidding. Participants were asked to score their agreement with the statement “The brand Stanley and Stella produces ethically responsible apparel” on a five-point Likert Scale ranging from 1 = “completely disagree” to 5 = “completely agree”. The variable was labelled ‘organizational evaluation’.
Independent variable – In order to test for differences among diverging certification labels, a
comparative model was used (Bougherara et al., 2009), as already explained in section 3.2. Through tweaking the auction advertisement designs by only changing the certification labels signalling ethical produced production methods, differences in certification label effectiveness can be measured on the two dependent variables (again see appendix 1 for the mock-up advertisements).
Moderators – In this study, two moderating variables are tested: certification reputation and
certification label credibility.
In order to test for the moderating effect of perceived certification label reputation, respondents were asked whether or not they agreed with the high quality of the certification label that was presented to them (Fombrun, Gardberg & Sever, 2013). Again, this was measured on a five-point Likert Scale ranging from 1 = “completely disagree” to 5 = “completely agree”.
The moderating effect of perceived certification label credibility was measured on a two-item construct. Participants were asked whether or not they agreed with the high trustworthiness of the certification label presented to them, as well as its high expertise, again on a five-point Likert Scale ranging from 1 = “completely disagree” to 5 = “completely agree”. The scale was adapted from Dentoni, et al. (2014) by changing its measurement from a seven-point Likert Scale to an easier to interpret five-point scale. The survey questions asked to participants through Veylinx can be found in appendix 2. Prior to statistical analysis, the two items measuring credibility had to be combined into one variable. Tests for internal consistency (α ≥ .8) were run for both trustworthiness and expertise. Factor analysis showed that both items loaded on the same component, which explained 91,75% of the variance. The items also showed to be reliable (Chronbach’s α = .91). Therefore, trustworthiness and expertise could be combined into a new variable labelled ‘credibility’.
Control variables – This study controlled for the possible effects of four variables: gender, age,
education and frequency of buying organic products.
Gender. In order to test for differences among gender, the experiment was conducted under
both men and women. Findings in regards to the effect of gender differences on willingness to pay show lots of consistency. Most studies identity female consumers as being more environmentally concerned than male consumers (Berkowitz & Lutterman, 1968; Hustvedt et al., 2008) and are therefore more likely to buy ethical products (Gilg et al., 2005; McGoldrick et al., 2008). Laroche et al. (2001) also found that 57 percent of the female participants in their study showed a higher willingness to pay for ethical products, compared to only 40 percent for the male participants. Research by McGoldrick et al. (2008) also found a slightly higher willingness to pay by female
consumers. Their results however showed not to be significant. Gender is dummy coded into 0 = male, 1 = female.
Age. Studies by Auger et al. (2003), Gilg, Barr and Ford (2005) and McGoldrick et al. (2008)
found that age is a significant influencer on the likelihood of consumers to consider and buy ethical products. Younger consumers are expected to be more willing to choose ethically produced products over conventional produced equivalents. The actual willingness to pay and price premium are however considered to not be influence by age (Laroche, Bergeron & Barbaro-Forleo, 2001; McGoldrich et al., 2008). A variable measuring age was computed by using year of birth of the participants, extracted from Veylinx database.
Education. Previous studies have found a positive correlation between education and organic
purchases. Consumers that have enjoyed higher levels of education are more likely to choose organic over conventional products (Thompson, 1998; Gilg et al., 2005). Highest level of education completed was found in Veylinx’ database of 578 participants. Education was measured as categorical variable, with 1 = primary education, 2 = secondary education, 3 = MBO/Intermediate vocational education, 4 = HBO/University of applied sciences and 5 = WO/University.
Frequency of buying organic products. The literature states that consumers who buy organic
food are very much likely to do so regularly and show a higher willingness to pay (Loureiro et al., 2001; Hustvedt et al., 2010). More specifically, Ellis et al. (2012) found that consumers with a higher frequency of organic food purchases were also willing to pay a price premium for organic cotton. Data on this variable was retrieved from the Veylinx database, collected amongst 577 participants by asking the questions “how often do you buy organic products?”, measured on a five-point Likert Scale ranging from 1 = “never” to 5 = “daily”.
Measurement – IBM SPSS Statistics 21 was used to run all statistical analyses.
4. Results
In this chapter the results of statistical data analysis will be presented. First, data collected on the dependent variables will be described using descriptive statistics. Afterwards, relevant correlations between variables will be examined. Finally, results per hypothesis will be presented.
4.1 Descriptive statistics
Willingness to pay – By exploring the dataset for the first dependent variable ‘bid_amount’ (N = 765),
measuring consumers’ willingness to pay, a mean willingness to pay of €2,82 with a standard deviation of 3,63 was found, together with a minimum and maximum bidding amount of respectively €0,00 and €25,00.
Looking at the four different treatment groups shows that 268 participants were not signalled a certification label. Participants in treatment group 1 showed a mean willingness to pay of €2,08 (SD = 3,13). The second treatment group, consisting out of 153 participants was shown the first party certification label. These participants placed an average bidding amount of €3,47 (SD = 4,04). The third party Fair Wear Foundation label was shown to the third treatment group, consisting out of 166 participants. Their average bidding was €3,26 (SD = 4,06). Finally, the last treatment group was shown the third party GOTS label. This group contained 178 participants, with a mean willingness to pay of €2,96 (SD = 3,36) (see table 4). Levene’s Test of Equality of Error Variances was used to evaluate whether or not the four groups were suitable to be compared. It was found that the assumption of homogeneity of variance had not been met with alpha α = .01 (Levene’s F (3,761) = 12.15, p < .01) (field, 2013).
When looking at the distribution of the bidding amounts in eurocents, a distribution far from normal was presented with a positive skewness of 1.71 (SE = 0.09), indicating clustered scores on the right hand side of the histogram (Field, 2013). Additionally, a positive kurtosis of 3.88 (SE = 0.18) showed a leptokurtic distribution, meaning that many scores are located in the tails of the distribution. This distribution can easily be explained by the 330 participants (43.1%) showing a willingness to pay of €0,00. Figure 1 in appendix 3 shows the histogram and boxplot of the variable bid_amount.
Willingness to pay excluding bidding amounts of zero eurocents – In an attempt to normalise the
distribution, the natural logarithm of the variable bid_amount was computed, as it reduces the skewness (Murphy, 1985). This resulted in a new variable labelled LNBid_amount. The new variable LNBib_amount however still showed not to be normally distributed. Fifty-three participants were deleted listwise from the dataset, as they had not filled in the questionnaire that was presented to them after the bidding. SPSS reckoned another 322 cases as missing. This resulted in a dataset suitable to be used for the statistical analysis with N = 390. Figure 2 in appendix 3 shows the histogram and boxplot displaying this distribution, with a skewness of -2.38 (SE = 0.12) and kurtosis of 10.67 (SE = 0.25).
A final attempt to improve the distribution of bid_amount, and limit the number of cases being left out, involved removing all cases bidding less than 1 euro. This resulted in a dataset of N = 442 with a much-improved skewness of -0.84 (SE = 0.12) and kurtosis of 1.76 (SE = 0.23), which can be considered normally distributed (George and Mallery, 2010). Still, assumptions of normal distribution are violated when looking at Kolomogorov-Smirnov (442) = 0.17, p < .001 and Shapiro-Wilk (442) = 0.94, p < .001. However, based on visual inspection of the histogram, as shown in figure 3 in appendix 3, the distribution of LNBid_amount can be assumed to be normally distributed. Additionally, in line with the central limit theorem, LNBid_amount can be assumed to be normally distributed, as N > 30 (Field, 2013). Again, Levene’s Test of Equality of Error Variances was run and found that the assumption for homogeneity of variance had been met with alpha α = .01 (Levene’s
F (3,438) = 2.70, p = .05) (field, 2013), meaning that the four different treatments groups are suitable
to be compared for LNBid_amount.
It should be kept in mind that the results of LNBid_amount, representing willingness to pay excluding bidding amounts of zero eurocents, are the natural logarithm of bidding amount, not the actual bidding amount in eurocents. This variable will be use in all statistical analyses, as it only includes participant that were actually interested in the product by showing a willingness to pay.
Consumers’ organizational evaluation for producing ethically responsible – The second depend
variable, labelled ‘organizational_evaluation’, showed another 60 missing values, resulting in N = 652 used for statistical analysis. The mean organizational evaluation score is 2.92 (SD = 0.73). When
