Predictors of fair trade apparel consumption, an application of a modified health belief model

1

Predictors of fair trade apparel consumption, an application of a

modified health belief model

Keywords

Ethical consumer behavior, apparel industry, fair trade apparel, modified health belief model [HBM]

Master thesis

MSc in Business Administration – Marketing track University of Amsterdam (UvA)

Under supervision of | dr. S. Wendel

Second supervisor |

By | Kim Koemeester

Student number | 10556559

Date of submission | 29.06.2015

2

Statement of originality

This document is written by Kim Koemeester 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.

3

Table of contents

Abstract

... 5

1. Introduction

... 6

1.1 The apparel industry ... 6

1.2 Ethical consumer behavior in the apparel industry ... 9

1.3 Problem definition ... 10

1.3.1 Research question ... 10

1.3.2 Sub questions ... 11

1.3.3 Delimitations of the study ... 12

1.4 Contribution ... 13

1.4.1 Theoretical contribution ... 13

1.4.2 Managerial contribution ... 13

1.5 Outline of the study ... 14

2. Literature review

... 15

2.1 Ethical consumer behavior ... 15

2.2 Translating ethical intentions into ethical behavior ... 16

2.3 Difficulty of predicting ethical behavior - the intention-behavior gap ... 17

2.4 The Health Belief Model ... 18

2.4.1 Traditional Health Belief Model ... 18

2.4.2 Perceived susceptibility - probability ... 20

2.4.3 Perceived severity ... 20

2.4.4 Perceived benefits ... 21

2.4.5 Perceived barriers ... 21

2.4.6 Interaction ... 23

2.4.7 Self-efficacy ... 24

2.5 Additional constructs in the modified HBM ... 25

2.5.1 Knowledge on the subject ... 25

2.5.2 Norms ... 26

2.5.3 Values ... 26

4

2.6 Conceptual framework ... 28

2.7 The present study ... 29

3. Methodology

... 30

3.1 Sample ... 30

3.2 Research design ... 31

3.3 Procedure ... 31

4. Results

... 39

4.1 Preparation of raw data ... 39

4.1.1 Factor analysis ... 39

4.1.2 Reliability ... 42

4.1.3 Summated scales and dummy variables ... 43

4.1.4 Correlation ... 45

4.1.5 Descriptive statistics ... 45

4.2 Model testing ... 47

4.2.1 Regression assumptions ... 47

4.2.2 Testing the traditional HBM ... 48

4.2.3 Testing the modified HBM ... 49

4.2.4 Testing individual predictor variables ... 50

4.2.5 Testing the interaction effect ... 52

5. Discussion

... 55

... 61

... 67

5

Abstract

The production of apparel is a topic which has received increasing attention in recent years, especially concerning the degrading working conditions in the sweatshops and the opaque supply-chain. Interestingly, while apparel brands, factory owners and local governments are heavily discussed, the role of the consumer is barely put forward in the debate and little change is seen in their behavior. Where ethical consumer behavior might be expected in the form of the consumption of fair-trade apparel, a rise of fast fashion chains appears. In an attempt to explain this behavior, previous literature has studied the individual influences of one or a few variables connected to ethical consumer behavior in the apparel industry. In this research, a theoretical framework is created in which all important variables are incorporated, to eventually be able to predict ethical consumer behavior. This is done by modifying the health belief model, a well-supported theory common in the health literature and applying it to ethical behavior in the apparel industry. A study was conducted via an online survey among 234 Dutch apparel consumers. The study resulted in the modified model being able to predict 38 percent of the ethical consumer behavior in the apparel industry. The main predictors were found to be objective knowledge, feeling of responsibility, micro-societal values and outcome expectancy. Perceived threat, self-efficacy, norms and macro-societal values did not have a significant effect in this study. The main predictors can be used as focus points by managers in the fair-trade apparel business to make their marketing efforts more effective.

6

1. Introduction

1.1 The apparel industry

“Enough is enough. I will not buy any clothes at apparel shops that do business with the controversial Indian spinners anymore.” (Van Es, 2014) This is the statement of a poll on the website of Dutch newspaper De Volkskrant of 28 October 2014. 1900 votes have been counted, 66 percent of the people agree but the other 34 percent disagrees. The amount of people that disagree is shockingly high, especially after reading the corresponding article. How can consumers prefer to buy apparel produced by young girls who are assumingly robbed from their freedom?

In the last decades, fashion has become a daily consumption article while it once was a privilege for the rich elite. Most people only needed a few outfits and a suit for Sunday. Thanks to retailers like Primark, H&M, Topshop and Zara, fashion has been fully democratized (Tilstra, 2013). As a consequence, the average consumer nowadays has 70 garments in his closet. Most of them are barely worn, nearly 70 percent of the Dutch consumers even has between five and twenty garments in the closet that have never been worn (Kien, 2013). The reason for this is the dramatic increase in the number of collections a brand displays per year, a phenomenon called fast fashion (Ghemawat, Nueno & Dailey, 2003). The aforementioned retailers have gained their success by supplying cheap fashionable items, targeting young people who are fashion conscious but with a limited income. To realize this, the retailers embraced a quick response model of production, aimed to dramatically reduce the time spent between the initial design and their eventual arrival in retail outlets (Jones, 2002). The business model is simple: stock inexpensive but fashionable items in small quantities to encourage customers to frequently buy the products and hereby assure that the stock turnaround in stores was ten days to two weeks (Choi, 2012). This has two major advantages; reduced inventory costs and frequently varying items and clothes. The latter is seen as a necessity by many retailers as consumers are demanding ‘new and now’ and favor the fast fashion brands that can deliver this (Lewis & Dart, 2010). Primark is the European textbook example of this preference, a rare and genuine growth story of cheap fashion in an intensely competitive market (Rijlaarsdam, 2015).

On the downside, the apparel industry has seen the largest disasters in its global industry in the past three years (International Labour Rights Forum [ILRF], 2013). Especially in Bangladesh with the fire at Tazreen Fashions in November 2012 killing more than 100 workers and of course the Rana Plaza building collapse in April 2013 where more than 1100 workers died and hundreds were crippled (T.J., 2014). The deaths of these garment workers could have been prevented with legally-required

7 health and safety measures (ILRF, 2013). These measures should be part of the Corporate Social Responsibility (CSR) programs of global Western corporations, which among others promote voluntary and unenforceable safety and labour right commitments and stimulate efforts to improve national labour justice and inspections systems (ILRF, 2013). Companies have used CSR programs to claim responsible treatment of workers in the supply chain, while rejecting legal responsibility for workers’ safety or welfare, shifting it wholly to contractor factories and local governments (ILRF, 2013). However, in emerging economies it is quite common to see individual factory owners pursuing their self-interest, reinforced by the existence of an institutional framework that is permissive and favors enterprises that provide employment, often regardless of actual work and safety conditions (Taplin, 2014). What arises is a picture of forces in the supply-chain of apparel that keep each other in position. Due to the opaqueness of the supply-chain, blame can be shifted from party to party and no responsibility is taken for the wellbeing of the garment workers. Considered together with the aforementioned changing industry structure which has placed an even greater pressure on cost and time, the fast fashion model seems to take its toll. Although the role of the consumer is barely put forward in the general debate, they appear unwilling to pay more for items that are not produced under exploitative conditions, the so-called fair trade apparel (Taplin, 2014). The Western consumer seems to have become accustomed to cheap fashion but as Eyskoot (2014) states in her book Talking Dress; cheap fashion actually does not exist, somebody else just pays the price.

The apparel industry can be seen as one of the world’s most exploitative industries (ILRF Industries, 2015). Worldwide, a total of around 80 billion kilos of textiles is produced annually, which has an impact on both society and the environment (Tilstra, 2013). Examples of the environmental impact include; the fact that after agriculture the apparel industry consumes the most water, that a quarter of the total amount of insecticides and ten percent of all pesticides are used in the cultivation of cotton, and that half of the problems with wastewater are related to the textile industry (Tilstra, 2013). Naturally, these environmental impacts will also have an effect on society as they contribute to the slow depletion of the earth making the earth less livable. Another impact on society is the concerns for the welfare of the workers in the so-called sweatshops of the apparel industry. In the online Oxford Dictionaries a sweatshop is defined as “a factory or workshop, especially in the clothing industry, where manual workers are employed at very low wages for long hours and under poor conditions” (Oxford Dictionaries online, “sweatshop”, 2015). Sweatshop workers are vulnerable men, women and children possibly kept there by illegal tactics and paid little or nothing (International

8 Labour Organization [ILO], 2015). One of the concerns is the possibility of forced labour in the sweatshops. Forced labour takes different forms, including migrants trapped in debt bondage, trafficking and other forms of modern slavery (ILO, 2015). In total almost 21 million people are victims of forced labour of which 19 million victims are exploited by private individuals or enterprises. Another concern which appeals more to the imagination of many consumers is child labour. Child labour can be described as work done by a child that is likely to interfere with his or her right to education, or when it is harmful to their health or physical, mental, spiritual, moral or social development (Stop Child Labour, 2015). According to the Convention on the Rights of the Child (Assembly UN General, 1989), all work done by children under the age of 15 and dangerous work done by children under the age of 18 is illegal. However, according to estimates by the International Labour Organisation (2015), there are 168 million children working worldwide.

Contrasting these views on sweatshops, there are academics that argue in their favor (Powell & Zwolinski, 2010; Kristoff, 2009). Although it is not denied that the working conditions are poor, the emphasize is put on the fact that the exchange between the worker and employer is mutually beneficial as workers are better off in a sweatshop than in alternatives jobs in the domestic economy. It is stated that sweatshops are a method to bring economic development to a country and that they should be seen as a way of alleviating the poor from poverty (Powell & Zwolinski, 2010). Although this economic justification might hold some truth about why sweatshops should not be prohibited, it does not successfully argue why apparel brands and the manufacturers should not improve the working conditions for the workers. The concerns for the human rights of the possibly forced workers and/or children can thus be seen as an ethical issue; “a problem or situation that requires a person or organization to choose between alternatives that must be evaluated as right (ethical) or wrong (unethical)” (BusinessDictionary.com, “ethical issue”, 2015). The people responsible have to weigh the costs of ethics and profit. In this case, the ethical issues also force the consumer to choose for responsible fair trade apparel or the usual apparel of which the background is unknown.

In the media, the tide seems to be turning in the past years. As the CSR programs of corporates have proven to fail in protecting the workers and the abuse of company-controlled social auditing have been exposed and thoroughly debated in the media. This increased focus on businesses behaving in a socially responsible manner has led to growing concerns of the media and governments about apparel sweatshops (Dickson, 2000). Next to the wider media coverage, there are other reasons behind the increased focus on CSR and fair trade values. These reasons include;

9 growing awareness of Third World issues around the world, increasing availability of information about fair trade principles, the increasing amount of fair trade products, and a shift in values towards sustainable development among consumers (Strong, 1997). Companies now have to protect their corporate reputations and brand images when faced with anti-sweatshop, fair trade, and other corporate accountability movements. A step in the right direction is for companies to join the UN Global Compact, a global platform launched in July 2000 to align business strategies and operations with ten universally accepted principles in the areas of human rights, labour, environment and anti-corruption. With nearly 8,000 corporate participants it is now the world’s biggest voluntary corporate sustainability initiative (UN Global Compact Office, 2014). In the last fifteen years, also the mobilizations of thousands of people and hundreds of organizations have contributed to forcing companies to take responsibility (Crabbé, De Koninck & Caudron, 2008). But now that the awareness is there, how do consumers actually react on the knowledge that clothes they purchase are made by low-wage workers under abusive working conditions?

1.2 Ethical consumer behavior in the apparel industry

As the awareness of the environmental and the social impact of their own consumption is growing among consumers, the demand for more ethical products has grown (Doane, 2001). Ethical concerns can include environmental, animal, societal and people or labour issues and can basically be applied to all products and services (e.g. Shaw & Clarke, 1999; Strong, 1997). Doane (2001) adds that the product concerned with an ethical issue is chosen freely by an individual consumer. Consumers can express their ethical concerns by buying products for their positive qualities or by boycotting products for their negative qualities (De Pelsmacker, Driesen & Rayp, 2005). In this study a focus is placed on ethical concerns specifically in the apparel industry; concerns about environmental issues and especially fair labour practices in the production of apparel. Most of these concerns relate to apparel sweatshops, commonly known for having a lack of social responsibility (Dickson, 2000). To avoid buying apparel produced in sweatshops, consumers can now make a difference by choosing for alternative consumption movements like fair trade clothing globally or clothing with the so called “No Sweat” labeling specifically in the United States of America (Straight & Grigsby, 2013; Shaw, Hogg, Wilson, Shiu, Hassan, 2006; Rudell, 2006). Fair trade is a term, mostly used in western countries, that describes international trade with a focus on ensuring a fair price and fair working conditions for producers and suppliers and promoting righteous trading agreements (Shaw et al, 2006). When

10 deliberately choosing for fair trade products, consumers depict ethical behavior in the form of consumption. Some authors see ethical consumption as political activism, in which citizens have the opportunity to bring their political issues to the marketplace and use their consumer choice to attempt to change institutional practices (Straight & Grigsby, 2013; Micheletti & Stolle, 2008). Most often though, fair trade buying is seen as a specific socially conscious consumer behavior (De Pelsmacker et al, 2005). The question remains, as Carrigan and Attala point out (2001), if social consciousness or ethics actually matter in consumer purchase behavior. And if it does, how this ethical consumer purchase behavior can be predicted and influenced?

1.3 Problem definition

In this chapter the main research question will be stated followed by the sub questions which will be answered throughout the study. Also the delimitations of the research will be denoted.

1.3.1 Research question

It is evident that the subject of ethical behavior is quite thoroughly researched. Much of the research though, examines individual influences of one or a few variables. What seems to be missing is an attempt to incorporate these variables into one theoretical framework. The research presented in this report is an attempt to create a more complete framework by modifying a theory common in the health literature and applying it to ethical behavior in the apparel industry. The chosen theory is the health belief model, HBM, originally developed by Rosenstock (1966). The HBM was developed to predict the likelihood of a person to engage in recommended health behavior. Although the HBM was originally developed to explain preventive health behavior, it has also been applied in studies of environmental behavior. Lindsay and Strathman (1997) examined the application of the traditional and a modified HBM to recycling behavior and the results indicated that both versions significantly predicted recycling behavior. The idea to use the HBM as a framework for explaining environmental behaviors originated when predictors of general environmental behavior and health behavior revealed strong similarities (Lindsay & Strathman, 1997). Behavior performed to maintain good health and behavior to conserve nature are both voluntary behaviors to prevent a negative state, like sickness and pollution. Also, both types of behavior are only performed when the negative consequences are severe and likely to happen, and both involve accepting some direct consequences which can be inconvenient, time consuming and even financially demanding in order to achieve long-term positive consequences (Lindsay & Strathman, 1997). These similarities between health and environmental

11 behavior led to the proposal that the HBM, which originally explains health behavior, might also be useful to explain environmental behaviors. When applied to ethical behaviors in the apparel industry, these too are voluntary and are performed to prevent a negative state, in this case the exploitation of workers in sweatshops. Moreover, ethical behavior in the apparel industry is only shown when consumers have knowledge on the consequences and when they are convinced that they are severe. It is probably also for this reason that Dickson (2001) found that only 16 percent of consumers were willing to prioritize their buying decision based on ethics. Finally ethical behavior also involves some immediate consequences as it might require switching from your favorite brand and be more time consuming to find out which brand is fair trade. Moreover, consumers might have to be willing to spend more. As defined in the introduction, ethical behavior is partly environmental and party socially concerned. For all these similarities, it is proposed that the HBM might be useful as a framework for explaining ethical consumer behavior in the area of apparel purchasing. Will the threat of injustice lead to a likeliness to change one´s behavior and counteract?

Yet, despite all the similarities, it has to be noted that there is a difference between health and environmental / ethical behaviors. Where health behavior results in personal benefits, the other two result in more collective benefits. This though, makes it even more interesting to use the HBM because it might provide valuable insights into the psychological processes of consumers, indicating exactly why and when the collective benefit is important to them.

Using the HBM theory provides three important benefits: it is thoroughly tested and supported in previous literature leading to a high reliability (Janz & Becker, 1984), it incorporates variables already identified as important in predicting ethical behavior which makes it very applicable and it gives insight into the psychological processes of an individuals’ decision to perform a certain behavior which might clarify questions in the literature of ethical behavior (Lindsay & Strathman, 1997). Overall the current study aims to complement the literature on ethical behavior by answering the following research question: What are the predictors of fair trade apparel consumption using a modified Health Belief Model?

1.3.2 Sub questions

Several sub questions are formulated to guide the research to eventually be able to answer the main research question. In addition, sub question 5 adds to the theoretical relevance of this study by

12 primarily investigating if the traditional HBM has enough explaining power and thereafter analyzing what the difference in explaining power with the modified HBM is.

- Sub question 1: What is ethical consumer behavior in the apparel industry?

- Sub question 2: How do the constructs perceived -susceptibility, - severity, - benefits, - barriers and self-efficacy of the traditional HBM apply to ethical consumer behavior in the apparel industry?

- Sub question 3: Which specific constructs, concerning ethical consumer behavior in the

apparel industry, discussed in the literature have to be added in order to create a comprehensive model?

- Sub question 4: What are the hypothesized relations based on previous literature?

- Sub question 5: How can the different variables be measured?

- Sub question 6: What correlations and relations can be found when analyzing the results of the research?

- Sub question 7: What is the explaining power of the traditional HBM versus the modified model?

- Sub question 8: What is the significance of the results of this study into the predictors of ethical behavior in the apparel industry?

1.3.3 Delimitations of the study

The core topic of this study is ethical consumer behavior in the apparel industry. This indicates that the literature review is focused on fair trade clothing and does not include other fair trade products like food. Also it is important to mention that this study analyses the ethical behavior of people as a consumer and as a concerned citizen. Measuring actual behavior indicates a very costly and time-consuming experiment, for this reason behavioral science mostly focuses on researching reported behavior or intentions as they have a high probability of predicting actual behavior (Ajzen & Fishbein, 1980).

The literature review of this study explores ethical consumer behavior and why this specific concept is difficult to measure. It explains how behavior is formed introducing the topic of intentions and explains why particularly in the case of ethical behavior, the intention-behavior gap is so evidently present. Also the traditional Health Belief Model will be elaborated on including its original variables; perceived benefits, perceived barriers, perceived severity and perceived susceptibility but also the

13 later added variable self-efficacy. The variable cues to action will be mentioned but not included in the model. For the modified model, several constructs which originate from the literature will be added and elaborated on; knowledge on the subject, norms, values and feeling of responsibility. A construct which could have been added is a test for social-desirability in the answers that respondents will give. This could have been a valuable addition but will be left-out due to the length of the questionnaire.

1.4 Contribution

1.4.1 Theoretical contribution

The purpose of this paper is to explore predictors of ethical behavior, factors that might influence decisions to purchase from socially responsible apparel retailers. The theoretical contribution is a modified HBM specified on the topic of ethical behavior. This larger framework will include various factors important in explaining ethical behavior which have not been incorporated in one model before. This will add to the literature on ethical behavior. In addition, if proven that besides for environmental behavior the usage of a modified HBM is also valid for ethical behavior, this will imply that the model might also be used for other sorts of behavior like e.g. safety or privacy behavior.

1.4.2 Managerial contribution

The intent of the research is to create a modified framework from which managers in the apparel industry and environmentalists might understand what drives the consumer on the topic of fair trade apparel. With this framework ethical consumer behavior can hopefully be predicted, indicating also which constructs are the best to focus on. The planning of any marketing process starts with understanding the marketplace and your target audience (Kotler & Armstrong, 2010). Therefore the insight into what the drivers of the ethical consumer are, will help create more effective marketing campaigns. An example is that a fair trade apparel brand will then know on which attributes of their products to focus on in their advertisement. Or the government or an NGO will then know that e.g. the overall knowledge on the subject is so low that an educative commercial might be most effective to increase the likelihood of behavior change of the Dutch consumer.

14 1.5 Outline of the study

This study has the following structure. In the next chapter, the theoretical framework will be set by a literature review leading to the development several hypotheses. This framework will be examined by a quantitative research for which the research method will be explained in chapter three. The results are presented and evaluated in chapter four. Finally, there is a discussion of the empirical findings including the conclusions, managerial implications, limitations and suggestions for further research.

15

2. Literature review

This chapter starts with the theoretical framework consisting of a review of the literature on the topic of consumer behavior related to social responsibility. To be more concrete, it is a review of purchase behavior in the case of fair trade products. The literature review will consist of four parts; the first describing ethical consumer behavior, the second examining the topic of translating ethical intentions into ethical behavior and the third analyses the difficulty of ethical behavior introducing the intention-behavior gap. After the literature review, the model of this study will be elaborated on. The traditional Health Belief Model and its dimensions will be explained together with the constructs which will be added to the model. This will lead to the development of several hypotheses and the conceptual model will be presented.

2.1 Ethical consumer behavior

Many different types of ethical behavior have been discussed in the literature, examples are water and energy conservation, fair trade consumption, waste reduction, recycling and choosing a sustainable bank. In this research a focus will be placed on ethical behavior in the case of fair trade consumption, specifically in the apparel industry. Although the market for fair trade products is clearly growing, the overall rates of consumer awareness remain low and for a lesser extent the ‘label fatigue’ is growing as more companies begin to make the fair trade claim (Stokes, 2010). Strong (1997) addresses another fundamental problem of translating fair trade, or ethical, intentions into consumer purchase behavior. Accordingly, this problem is caused by the barrier of consumers not recognizing the human element in the ecological marketing agenda (Strong, 1997). The underlying reason for this is the fact that the prevailing emphasis on green consumption of that time mainly stressed the importance of the environmental elements of sustainability and therefore overshadowed the human elements (Strong, 1997). The current fair trade perspective counters this by focusing on the human factors while maintaining concern for the environmental ones. Moreover, the actual knowledge a consumer has of the social issues may play a part in the ethical responsibility and apparel consumption (Shim, 1995). This knowledge includes facts on how exploitative the fashion industry actually is and actual awareness of fair trade brands in the apparel industry. Although this barrier seems to be overcome by the increasing attention to the fair trade perspective, also in the clothing and fashion sector, it is still a problem to influence consumers’ intentions. Like the outcome of the poll

16 pointed out, there is a remarkable group of people that indicate to keep buying non-fair trade clothing, but why?

Another part of the literature on ethical consumer behavior focuses on how to encourage the behavior. Arousing cognitive dissonance seems to be an effective way to change behavior, especially when the subject already supports the desired goal but finds their behavior inconsistent with those beliefs (Dickerson, Thibodeau, Aronson & Miller, 1992). Another option is to use normative social influence, especially when the normative information is about a specific referent group this will guide and spur behaviors (Schultz, Khazian, Zaleski, 2008). Even though these examples show us how to stimulate ethical behavior, the question remains why and when people engage in ethical behavior.

2.2 Translating ethical intentions into ethical behavior

One explanation for non-ethical intentions and behavior can be found by examining the evolutionary basis of human behavior. Kenrick, Griskevicius, Neuberg and Schaller (2010) have a very specific view on what type of factors explain behavior, they draw an important distinction between ultimate and proximate explanations. Behavioral scientists typically tend to focus on proximate explanations of behavior, which are relative immediate triggers for behavior. Ultimate explanations for behavior are part of an evolutionary perspective that asks why humans have evolved to behave in a certain way. Griskevicius, Cantu and van Vugt (2012) consider the evolutionary bases for why people sometimes act in non-sustainable behavior. They propose that this non-sustainable behavior is caused by five evolutionary tendencies; (1) self-interest, tendency to prioritize self-interest over group welfare, (2) relative status, people are more motivated by relative than absolute status, (3) social imitation, unconsciously copying what others are doing, (4) future discounting, people value the present more than the future and (5) impalpable concerns, the tendency people have to disregard problems they cannot see or feel (Griskevicius et al, 2012). Although it is hard to research these ultimate explanations which people conduct unconsciously, it might be important to keep them in mind when evaluating the results of any behavioral research.

A more seizable concept on the topic of behavior remains the proximate explanation. The predominant thinking of many researchers is that personal values influence attitudes which in turn impact behavior (Ajzen & Fishbein, 1980). Although some scholars question the validity of using attitudes as a predictor of behaviors (Hill, 1990), research has shown improved correlations by using certain measurement conditions and by acknowledging differences in attitudes (Dickson, 2000). In

17 particular, behavioral intentions have become an important concept. Several models in the social psychology, including the theory of reasoned action (Fishbein & Ajzen, 1975; Ajzen & Fishbein, 1980), the theory of planned behavior (Ajzen 1985, 1991) and the attitude-behavior theory (Triandis, 1980), agree on the proposal that the most important predictor of a person’s behavior is his or her intention to perform it (Sheeran, 2002). Purchasing being a specific behavior, this indicates that purchase intentions can predict actual purchase behavior. Although these theories have the same assumption; that intentions are the most important predictor of behavior, they do acknowledge that people do not always have sufficient control over the performance of their behavior and suggest that low facilitating conditions can prevent the actual performance of an intended behavior (Ajzen, 1985, 1991; Triandis, 1980). Triandis (1980) also points out that ‘habits’ are quite likely to control certain behaviors instead of conscious intentions. All these remarks lead to the so-called intention-behavior gap which makes it hard to predict actual behaviors.

2.3 Difficulty of predicting ethical behavior - the intention-behavior gap

The aim of this research is to explore predictors of ethical behavior in the apparel industry and this will be done by, inter alia, identifying variables that form intentions as they have proven to be predictors of behavior. Therefore, it is important to acknowledge the intention-behavior gap concerning ethical consumer behavior in the apparel industry and to find the factors forming the intention to explain the ethical consumer behavior.

In the area of ethical behavior, the intention-behavior gap is quite evident. While 46 percent of the European consumers claim to be willing to pay more for ethical products, most of the ethical initiatives often have a market share of less than 1 percent (De Pelsmacker et al, 2005). Theoretically this is mainly due to the fact that attitudes or intentions alone are poor predictors of buyer behavior (Cobb-Walgren, Ruble & Donthu, 1995). Shaw and Clarke (1999) show that this is especially valid in the area of social marketing by focusing on formative factors underlying actual choices instead of only the decision making. Another adversary of the belief that only those intentions or attitudes which are specific to the purchase decision significantly affect the decision, are Dickson and Littrell (1996) who identified other variables with direct effects on the purchase intention. These variables are product-focused attitudes and societal attitudes which have been proven to affect social responsible consumption (Dickson & Littrell, 1996). More practically, the majority of consumers seems to make purchase decisions by evaluating all product attributes. Of course there are exceptions, like in the

18 research of Crane (2001) where consumers refuse to buy products with an unethical background, hereby focusing on only one product attribute. Most consumers however make a buying decision based on the factors; price, quality, convenience and brand familiarity (Carrigan & Attalla, 2001). Also the research of Shaw and Clarke (1999) revealed that even though consumers might place ethical concerns high on their purchase agenda, this is often not at the expense of the more traditional choice factors, like the aforementioned. By taking factors and constructs into account which might cause the intention-behavior gap for ethical consumer behavior and adding them to the HBM model, this will presumably add to the explaining power of the model. Thus a more comprehensive and solid model will be build. Examples of constructs discussed in the literature concerning ethical consumer behavior are knowledge on the subject, consumer commitment, price and willingness to pay, feeling of responsibility, norms, availability of the product and fashion. What is noticeable is that most of these constructs are either value- or expectancy related and together form the consumer intention. Therefore, a model can be formed that hypothesizes behavior based on two variables, (1) the value placed by an individual on a particular goal and (2) the individual’s perception of the likelihood that a certain action will achieve that goal (Janz & Becker, 1984). A well-established model that has these variables as its basic components is the Health Belief Model.

2.4 The Health Belief Model

In this chapter, the traditional Health Belief Model will first be elaborated on, followed by an explanation of each dimension in the context of ethical consumer behavior in het apparel industry.

2.4.1 Traditional Health Belief Model

For this study the Health Belief Model (HBM), originally proposed by Rosenstock (1966), is chosen as a basis. Like mentioned above, the HBM is based on the value-expectancy concept (Stretcher & Rosenstock, 1997). When translating this concept into health-behavior, the desire to avoid illness or to get well is the value, and the belief that a certain action would prevent or improve illness is the expectancy. Correspondingly, Rosenstock’s HBM states that the likelihood of a person to engage in a recommended health behavior is determined by two constructs; a desire to avoid a possibly negative physical condition and by the assessment of the costs and benefits of a certain action (Rosenstock, 1966). In the accompanying model this is depicted as an interaction of two main elements; perceived threat, defined as the assessment of the likelihood and the severity of an event and outcome

19

expectancy, defined as the benefits minus the costs of performing a certain action. So, as Rosenstock (1974) notes, the combination of susceptibility and severity provide the force to act and the perception of benefits minus the barriers form the preferred path of action (Janz & Becker, 1984). However, in their meta-analysis on the HBM, the authors Janz and Becker (1984) find there is a lack of studies in the literature examining the model in this way; as a whole model. Most studies analyze the dimensions separately and do not examine how various combinations of beliefs are more or less likely to change health-related behavior (Stretcher & Rosenstock, 1997). This study supports the idea of interaction of the different dimensions and will therefore not measure the effect of the individual dimensions but only of the two constructs, perceived threat and outcome expectancy, in which the dimensions are merged together.

In the version of Rosenstock and various other early versions of the HBM, it is also thought that the readiness to take action would be triggered by certain cues. Cues to action were found to be necessary to trigger the decision-making process and can be internal (like symptoms) or external (like media communications). However, as these cues can be as small as a sneeze or a barely conscious view of a poster, they are hard to asses (Stretcher & Rosenstock, 1997). Their role has not been systematically studied as only a few studies have attempted to incorporate them in the model (Janz & Becker, 1984). Moreover, the cues for ethical consumer behavior cannot be internal which would make it a different concept. For these reasons the cues to action will in this study also be left out.

In a more recent version of the model, Rosenstock added the concept of self-efficacy by Bandura (1977). Self-efficacy explains the extent to which a person beliefs that he or she can successfully perform the behavior or action to produce the outcome (Rosenstock, 1990). The reason for not incorporating self-efficacy in the earlier model is because this model focused on simple preventive behaviors like accepting a screening test or an immunization, which generally most people could perform (Stretcher & Rosenstock, 1997). However, when trying to adapt lifestyle behaviors the situation is quite different and long-term changes are required. People must feel confident that they can in fact alter their life-long habits in order for them to implement that change (Stretcher & Rosenstock, 1997).

So the recapitulate the traditional HBM; it is generally believed that an individual will take action against a perceived threat only if they regard themselves as susceptible to the condition and if they believe it to have potentially serious consequences (severity). The individuals will also have to belief that the available action will be effective and that the benefits of the action outweigh the barriers

20 (Stretcher & Rosenstock, 1997). Moreover, individuals will have to feel capable of performing the behavior. These three factors together form the intention one has to perform a certain behavior. In the following paragraphs the key variables will be further specified and an explanation will be given of how they will be used in this study.

2.4.2 Perceived susceptibility - probability

The dimension perceived susceptibility originally refers to one’s subjective perception of being at risk for a certain condition (Janz & Becker, 1984). In other words, this dimension covers the individual’s belief of the likeliness of an event to happen. As to be expected, individuals can vary a lot in their feelings of personal vulnerability (Janz & Becker, 1984). For ethical consumer behavior however, this dimension is more difficult to apply as the degrading working conditions will never happen to the consumers themselves. At most, the idea of the sweatshop workers will have an influence on their conscience when buying apparel that is not fair trade. Therefore it is decided to measure the individual’s belief in the likelihood of the situation in the production chain of apparel. To make this variable more comprehensible, it will be called perceived probability instead of susceptibility from now on.

2.4.3 Perceived severity

Perceived severity covers the beliefs concerning the seriousness of the condition (Stretcher & Rosenstock, 1997). So to what extent does it have to be avoided or how serious is it if no action is undertaken. In the case of ethical consumer behavior the severity refers to the perceptions of the situation in the apparel sweatshops and the consequences this has for the people working there.

In this study the dimensions susceptibility and severity will be merged as it is not expected that each will have a relevant effect on itself. To explain in terms of health behavior, everybody beliefs that Malaria is a quit severe disease, but only people that will visit areas where the Malaria mosquito is actually present will get the vaccination. This can be explained by their higher level of susceptibility to the disease. For this reason it does not seem relevant to measure the effects of severity and susceptibility on their own and this study intends to use to HBM as originally proposed. Like mentioned before, the combination of susceptibility and severity is labeled as perceived threat and it is believed that they produce a force leading to a particular behavior to avoid this threat (Stretcher &

21 Rosenstock, 1997). So, the higher the perceived threat is, the higher the likelihood of a behavior change is expected to be. This leads to the following hypothesis:

Hypothesis 1: there is a positive relation between the construct perceived threat and

likelihood of ethical consumer behavior

2.4.4 Perceived benefits

The type of behavior that will be performed depends upon beliefs of the effectiveness of the action that will be taken and this is called the perceived benefits (Stretcher & Rosenstock, 1997). So although an individual does feel the potential threat, he or she is not expected to accept a recommended action unless it is perceived as feasible and effective (Janz & Becker, 1984). Therefore, the benefits are formed by the potential outcome of the individual’s action. Applying this to ethical consumer behavior in the apparel industry, this variable refers to the extent in which an individual actually beliefs that his or her ethical behavior like e.g. boycotting a certain brand will be effective.

2.4.5 Perceived barriers

The potential negative aspects of a particular action may act as a barrier to undertake the recommended behavior (Janz & Becker, 1984). These barriers occur when actions are expected to be expensive, dangerous, unpleasant, inconvenient, time-consuming etc. In the case of ethical consumer behavior there are several barriers that occur in the literature.

The first identified barrier is price as not all consumers are equally willing to pay extra for ethical products (De Pelsmacker et al, 2005). In fact, a high price might create a perception of financial risk that leads to high consumer involvement in the particular purchase decision (Assael, 1987). This is undesirable as this can monopolize the attention of the consumer which will most probably lead to the overshadowing of the ethical intention (Thøgersen, 1999). The general sentiment seems to be that to ensure fair labour conditions and to take the environment into account, extra monitoring and third-party certification is needed in order to be credible (De Pelsmacker et al, 2005). This implies additional costs leading to a higher price for consumers. However, in practice the actual price increase for fair trade apparel differs substantially as it is primarily about creating a living wage for the sweatshop workers and not about additional certification (McMullen, Luginbühl, Nolan, Crabbé & Ajaltouni, 2014). On the one hand, organic cotton can be guaranteed for only a few cents extra per

22 garment. While a complete assurance of fair labour conditions, which can be guaranteed by producing in for example Italy, can lead to prices three times higher than what the retail is used to (Betlem, 2015). But on the other hand, there are also numerous opportunities for creating fair(er) labour conditions in e.g. Asia which will have a minimum price increase. As a recent EU-financed report pointed out in the price breakdown of a simple t-shirt, the wage of the garment worker only represents 0.6 percent of the total price. If the wages would be raised to a fair living wage this will have a minimum impact on the total price of a garment (McMullen et al, 2014). So, for the apparel industry this barrier seems to be a misunderstanding. Moreover, consumers state to not even mind the potential price increase, almost 46 percent of European consumers claim to be willing to pay more for ethical products (MORI, 2000). However, in the end, price does remain one of the most important factors in the buying decision (Shaw & Clarke, 1999).

Another barrier to ethical consumer behavior is the commitment a consumer has to a certain brand. Research in consumer behavior has indicated that consumers become attached to various brands and form ‘relationships’ with them (Fournier, 1998). These relationships vary in strength and as expected, stronger attitudes exhibit greater resistance to negative information. In a research on the consumer response to negative publicity, this consumer commitment has shown to play a critical role in determining the resistance one has to negative information (Ahluwalia, Burnkrant & Unnava, 2000). The moderating role of consumer commitment is also visible in the apparel industry. Although negative information about apparel brands is widely present, consumers keep buying their favorite brand. Although Dickson (2001) identified a potential market for “no sweat labels” in clothing, guaranteeing certain working conditions, she also found that only 16 percent of the consumers would be strongly influenced and willing to prioritize the social labels when making the purchase decision. In addition, research has shown that the fair trade consumers are not often consistent in their ethical behavior (Strong, 1997). This indicates that although most of the time fair trade products are bought, fair trade consumers occasionally buy non-fair trade products maintaining the non-ethical consumerism and hereby giving a mixed message to the apparel brands.

Other studies have shown that barriers to ethical buying behavior can also be caused by the lack of availability of ethical products and lack of credibility of the ethical claim (Carrigan & Attalla 2001; De Pelsmacker et al, 2005). The availability of ethical products is still a challenge as fair trade or people-friendly products are only recently starting to be recognized by the supermarket organizations. This in contrast to environmentally-friendly products which were the prior challenge for

23 activists but which has now widely succeeded (Strong, 1997). Fair trade is however, starting to become a more important niche market. Also in the apparel industry this is visible by the rise of brands like ArmedAngels and Kuyichi leading to a higher availability of the products. The lack of credibility of the ethical claim is partly caused by the aforementioned label fatigue, as more and more companies make claims in the labels, less attention is paid to the actual message and its credibility is lost (Stokes, 2010).

Research among ethical consumer groups also revealed that fashion is very important in the actual clothing choice (Shaw & Tomolillo, 2004). This indicates that although consumers might be willing to purchase fair trade apparel, this will not be at the cost of the style and fashionability of the item (Shaw et al, 2006). This also came forward in the complaints of the ‘unfashionability’ of the many fair trade alternatives available in the research of Shaw and Tomolillo (2004).

All these barriers have to be taken into account when researching ethical behavior in the area of fair trade apparel as they form serious impediments to the intention of performing ethical consumer behavior.

Similarly as done for the construct perceived threat, also the dimensions perceived benefits and barriers will be merged together to form the construct outcome expectancy. The corresponding reason is again that is does not make sense to analyze the dimensions separately but that they have to be considered together to be relevant. Together, as the outcome expectancy, they will form the particular behavior to avoid a perceived threat. Looking at this equation practically, the outcome expectancy will be higher with high benefits and/or low barriers and will then lead to a higher likelihood of a behavior change. This leads to the following hypothesis:

Hypothesis 2: there is a positive relation between the construct outcome expectancy and

likelihood of ethical consumer behavior

2.4.6 Interaction

Another important aspect of the traditional HBM is the interaction between the different beliefs. Going back to Rosenstock’s (1974) statement, in which he describes the combination of susceptibility and severity to provide the force to act and the perception of benefits minus the barriers to form the preferred path of action, this interaction can be noticed. In cases where the force is more urgent, the path of action is expected to be more predictive. This study supports the idea of interaction between

24 the different beliefs as also emphasized by Janz and Becker (1984), with a focus on a moderation between the constructs perceived threat and outcome expectancy. For higher levels of perceived threat, it is expected that the benefits will be better known and some barriers will be taken for granted, resulting in higher levels of outcome expectancy. This in turn leading to a bigger chance of ethical consumer behavior.

Hypothesis 3: there is an interaction between outcome expectancy and perceived threat in their relation to likelihood of ethical consumer behavior

Note, the variables outcome expectancy and perceived threat are also expected to be predictive on themselves like explained above. This moderation is an additional expected relation.

2.4.7 Self-efficacy

Aforementioned, the construct self-efficacy by Bandura (1977) had been added to the model in the more recent versions. As the researched health behaviors became more difficult, the perceptions of one’s own abilities started to play a more important role. Therefore, self-efficacy will particularly be a strong predictor of behavior that requires significant skills to perform (Stretcher & Rosenstock, 1997). Bandura (1977) hypothesized that expectations of personal efficacy determine if a certain behavior will be initiated, how much effort will be spend on it and how long the behavior will be sustained when barriers arise and this hypothesis was supported by his research. In the study at hand, self-efficacy is needed to perform the buying of fair trade apparel. Several factors discussed in the literature on ethical behavior in the apparel industry will not only affect the will of people to perform the behavior but also the feeling of being capable to perform the behavior. This indicates that the higher one’s level of self-efficacy is, the higher the likelihood of initiating the ethical consumer behavior will be and the longer it will last.

Hypothesis 4: there is a positive relation between the construct self-efficacy and likelihood of

ethical consumer behavior

Thus, for an ethical consumer behavior change to succeed, people must feel the threat of their current behavioral pattern and believe that changing it will be favorable by resulting in a valuable outcome at an acceptable cost. However, they must also feel competent to implement the change.

Like elaborated on in chapter 1.3 with the problem definition, the HBM is expected to be useful as a framework for explaining ethical consumer behaviors in the area of apparel purchasing for several

25 reasons. Now, after the literature review, it can also be stated that most constructs which are previously researched in literature on ethical behavior, correspond to constructs of the traditional HBM. This increases the probability of the HBM being able to identify predictors of ethical consumer behavior in the apparel industry. Therefore, the following hypothesis is developed:

Hypothesis 5: the traditional HBM can predict the likelihood of ethical consumer behavior in the apparel industry

2.5 Additional constructs in the modified HBM

In addition to the traditional HBM including self-efficacy, it is expected that the inclusion of four other variables may prove useful in understanding ethical consumer behavior. Based on previous literature it is hypothesized that the knowledge on the subject, normative behavior of people in the close surroundings, one’s values and the feeling of personal responsibility should contribute to the explanatory power of the HBM.

2.5.1 Knowledge on the subject

In order to perform ethical consumer behavior in the apparel industry, one must have the adequate knowledge on this subject of how to perform it. This includes basic knowledge on which brands are fair trade and which are not and where they can be bought. Unfortunately, several studies in the literature report a lack of available information on ethical products (Carrigan & Attalla 2001; De Pelsmacker et al, 2005). On a contrary, in the last twenty years several websites of organizations like the Fair Wear Foundation (1999) and Rank-a-Brand (2010) have emerged that guide consumers in their consumption choices. In 2013 also a new app called Talking Dress was launched to help specifically women find sustainable brands. In this study the knowledge of the respondents concerning fair trade apparel will be tested. It is expected that lower levels of knowledge on the subject will lead to a lower likelihood of ethical consumer behavior as people do not know where to go to buy fair trade brands or even which brands to buy.

Hypothesis 6: there is a positive relation between the construct knowledge on the subject and likelihood of ethical consumer behavior

26

2.5.2 Norms

Where some individuals are quite confident in their assessment of the costs and benefits of ethical consumer behavior, others are expected to be prone to informational - or normative social influence (Lindsay & Strathman, 1997). Normative social influence is defined as “an influence to conform with the positive expectations of another, a group or one’s self” (Deutsch & Gerard, 1955, p.629). Informational social influence is defined as “an influence to accept information obtained from another as evidence about reality” (Deutsch & Gerard, 1955, p.629). It is common that these social influences come together and it is expected that they influence one’s intentions and thus one’s behavior. Ajzen and Fishbein’s (1980) theory of reasoned action conceptualizes norms in a similar way; as a coercive social influence with which individuals feel motivated to comply. Furthermore there is little discussion in the literature about how or why the normative beliefs of people exert influence on their behavior (Rimal & Real, 2005). From the previous conceptualizations of norms though, an expectation of the impact of norms can be formed. The expected impact is if more people in one’s community perform ethical consumption, the higher the likelihood of one’s own ethical behavior will be as they will feel motivated to conform. Thus, the perceived level of ethical consumption of community members has a positive relation to one’s own intention of ethical consumption.

Hypothesis 7: there is a positive relation between the construct norms and likelihood of

ethical consumer behavior

In addition to the perceived proportion of ethical consumers in the community, the conceptualization of norms used in this study implies that an implicit rule exists among community members that performing ethical consumption is the right thing to do.

2.5.3 Values

The values people have, have appeared to be of significant impact on their ethical consumption behavior (De Pelsmacker et al, 2005). Values can be defined as abstract principles that reflect one’s self-concept (Dickson, 2000) and are often part of one’s personality. They serve as standards that guide our behavior by complying to firm personal beliefs that a given behavior or outcome is desirable or good (De Pelsmacker et al, 2005). Values can be divided in two groups, instrumental and terminal values. Instrumental values relate to desired modes of conduct, so how people want to reach a desired goal (Rokeach, 1973). Terminal values reflect the desired goals themselves and can be

27 separated into two basic dimensions; macro-societal (socially centered) and micro-societal (self-centered) goals (Dickson, 2000). For this study the terminal values are the ones of interest as they are expected to influence ethical behavior. This is supported by Dickson (2000), she stresses the importance of personal values along with beliefs, knowledge and the right attitudes relating to intentions to buy apparel from socially responsible businesses and also solely focusses on terminal values. Other studies also segment consumers in terms of their values or ethical concern, Fritzsche (1995) had this approach and concluded that the values of people behaving ethically were significantly different from the values of people behaving unethically. From this theory it is expected that people with a higher score on macro-societal values will be more likely to perform ethical consumer behavior, this leads to the following hypothesis:

Hypothesis 8: there is a positive relation between the construct macro-societal values and

likelihood of ethical consumer behavior

2.5.4 Feeling of responsibility

The last construct that will be added to the model is feeling of responsibility. Strong (1997) stresses the difference between the issues of whose responsibility the problems and their solution are, for social and purely ecological problems. For the ecological problems, the general view is “we’ve all got to do our bit” or “at least I’m doing something”. While for the social problems the same consumer may think that it is “the Indian Government’s responsibility to protect children” or consumers hold the belief that the workers now at least earn something which is better than nothing (Strong, 1997). This phenomenon can be seen as the shifting of responsibility which is likely to reduce the feeling of one’s own responsibility for the situation.

In addition, the moral expectations that consumers have of business managers play a role. As stated in the research of Marcus and Fremeth (2009), people expect managers to use resources wisely and responsibly and to act in a sustainable way. These sustainable expectations are not just environmental but social and economic as well. Here, the social dimension of sustainability covers the impact an organization has on society, including its employees, customers, community, business partners but also its supply chain (Marcus & Fremeth, 2009). Elaborating on this insight, consumers expect their favorite apparel brands to be sustainable even if not specifically mentioned in the apparel label. This will, perhaps unconsciously, lead to lower levels of feeling of responsibility. Moreover, people can feel responsible for solving the situation in the apparel sweatshops by their own actions or

28 they can feel that it is the responsibility of the brand, the factory owners or the governments. Thus, for lower levels of personal feeling of responsibility for the situation, the likelihood of an ethical intention is expected to be lower as well:

Hypothesis 9: there is a positive relation between the feeling of personal responsibility and

likelihood of ethical consumer behavior

As can be noticed, all of the mentioned constructs have been discussed and supported in the previous literature on ethical behavior in the apparel industry. Adding these constructs to the traditional HBM seems necessary to make the model as complete as possible. In this way they will presumably add to the explaining power of the new modified model which leads to the following hypothesis:

Hypothesis 10: the modified HBM can predict the likelihood of ethical consumer behavior in the apparel industry and, can explain extra variance in this behavior compared to the traditional HBM

2.6 Conceptual framework

As a convenient way of summarizing the new model with all its variables, the conceptual framework can be found in figure 1. In this framework it is clearly shown that there are 12 variables in total, one dependent variable; likelihood of ethical consumer behavior, 7 independent or predictor variables and 4 extra variables needed to compose 2 of the independent variables. The conceptual framework includes the traditional HBM in the upper part and the added constructs in the lower part, together forming the modified HBM. It is important to note that the 7 predictor variables in the framework together form one’s intention to perform a certain behavior, therefore leading to a likelihood of ethical consumer behavior. All independent variables show their corresponding hypothesis and whether the relation to the dependent variable is positive or negative. Hypothesis 3 covers the interaction between the construct outcome expectancy and perceived threat. Hypothesis 5 and 10 refer to the models, correspondingly the traditional HBM and the modified HBM, being able to predict the likelihood of ethical consumer behavior in the apparel industry. An overview of the hypotheses can also be found in appendix 1.

29 Figure 1. Conceptual framework of the modified HBM; predictors of ethical consumer behavior

2.7 The present study

To test the formulated hypotheses and answer the research question, a survey will be performed in the form of an online questionnaire. In the questionnaire, respondents will be asked questions regarding their knowledge on the subject of ethical consumer behavior in the apparel industry, the frequency of currently performing this ethical consumer behavior, their opinion and attitude towards several aspects of fair trade apparel and some personal questions concerning self-efficacy, norms, values and demographic and socioeconomic factors (gender, age, educational level, income and place of residence).

30

3. Methodology

In this chapter a description of the sample will be given, followed by a description of the research design and the measurement procedures.

3.1 Sample

The population of his research consists of all Dutch apparel consumers. The purchasing power of the Dutch consumer has been declining since 2010 and they have been forced to be reticent (Erich, 2013). Nevertheless, the Dutch fashion sector is expected to have realized a turnover of 8.8 billion in clothing and 1.6 billion in shoes in 2014 (Erich, 2013). The spending of Dutch consumers in the apparel industry has the following structure; the biggest segment by far is the ´clothing for women and girls 12+´, consisting of 43 percent. Next is the ‘clothing for men and boys, 12+’ with 26 percent. On ‘underwear’ more is spent than on ‘children clothing (2-11)’ with correspondingly 9 and 7 percent (Rabobank Cijfers & Trends, 2014/2015). The smallest segments in the expenditure overview are baby clothing, accessories and swimwear. Of the total apparel turnover of 8.8 billion, the share of online apparel shopping accounts for 14 percent, which indicates a spending of more than one billion. The aim is to create a sample which resembles this population as accurately as possible. This implies that the sample has to include both men and women of all ages, children and both on- and offline shoppers. However, to increase reliability and accuracy of the data, children were not be included in the sample. The age of the respondents will ideally be 16 and older.

With a population this big it is impossible to identify a sampling frame, therefore non-probability sampling was chosen as the method to select respondents. Although the population is roughly set out in the previous paragraph, it is not precise enough to apply quota sampling. Convenience sampling was therefore be applied by e-mailing the online survey to friends and family. To decrease biased selection and to create some sort of random selection, also volunteer sampling in the form of self-selection on different websites and online forums was used. The goal for data collection was to achieve a sample as large as possible to increase the representativeness of the sample and to eventually be able to generalize conclusions. According to Harris’s (1985) formula for producing the minimum number of participants, the absolute minimum of 10 participants per predictor variable is appropriate for regression equations. However, with approximately 30 participants per variable, a researcher will have better power to detect a small effect size (VanVoorhis & Morgan, 2007). Having eight variables, the aim of this study was to gather between 30 and 240 participants.