The relationship between legitimacy and movie downloads and the moderating role of online trust

(1)

Study: MSc Business Administration – Marketing Track

Institution: Amsterdam Business School / University of Amsterdam Thesis Supervisor: dr. J.Y. Guyt

Student: Marinda van Eersel Student number: 10868909 Date: 29-01-2016, final draft

MSc Business Administration

Master Thesis

The relationship between legitimacy and movie

downloads and the moderating role of online

(2)

Acknowledgements

This Master thesis is the final individual assignment for the master Business Administration – Marketing track - at the University of Amsterdam. It explores the relationships between legitimacy, online trust in the supplier and the number of movies that people download.

I would like to use this opportunity to thank the people who have been involved in this research. First of all, I would like to thank my supervisor, Jonne Guyt, for his helpful

feedback and continuous help throughout the process. Furthermore, I would like to thank my family and especially my boyfriend Ernst van der Wiel for their understanding for not seeing me as much and for their support. Last but not least, I would like to thank all the respondents that contribute to my research by filling out the survey.

Marinda van Eersel Werkendam, January 2016

Statement of originality

This document is written by student Marinda van Eersel 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

(3)

Abstract

In this thesis the relationship between legitimacy and digital piracy (i.e. movie downloads) and the direct and moderating effect of online trust in the supplier are examined. This has been tested with two complementary, quantitative studies. In the first study, attitude toward unethical behaviour is used as a proxy for legitimacy, which is expected to have a negative relationship with the number of movies that people download illegally. Neither attitude toward unethical behaviour nor the interaction effect of online trust had a significant relationship with the number of movie downloads. However, online trust in the supplier appeared to have a direct, positive effect on the number of movie downloads. So when online trust in the supplier increases, the number of movie downloads rises as well. In the second study, rule of law was used as a proxy for legitimacy and the relationship between rule of law in different countries and the number of movie downloads was examined. The hypothesized relationship was negative, so that a country with a high rule of law (strict rules and

regulations) would have lower download statistics than a country with a lower rule of law, but the effect was not found to be significant. The moderating effect of online trust in the

supplier, which was expected to decrease the strength of this relationship, was found to be significant. Moreover, this study also provided support for the direct positive effect of online trust on the number of movie downloads. These results suggest that online trust in the supplier interact with the rule of law in a country, despite the lack of a main effect of legitimacy. Overall we can conclude that legitimacy does not have a direct effect on the number of movie downloads, but online trust in the supplier does. Online trust in the supplier as a moderator is partially proved to interact with legitimacy. The final section of this study deals with the implications and avenues for future research.

(4)

1. Introduction

1.1 Background

The availability of fast Internet connections around the world is growing, combined with sophisticated and accessible mobile devices. This provides significant benefits like new economic business models and the ubiquitous access to information. However, this

technology on the Internet also creates opportunities for unethical consumer practices that can have a major impact on companies’ profits. A growing number of marketing studies focus on consumer practices that are considered unethical, such as digital piracy (Yoon, 2010; Tan, 2002; Liao, Lin & Liu, 2009). Digital piracy is taking an increasing part of revenues of several industries like software, music, and motion picture (movie). Especially the motion-picture industry is affected since this industry already faces difficulties with profitability, like low return rate and high volatility (Walls, 2008). Peer-to-peer networks on the Internet, fast Internet connections and the perceived benefit of saving time and money have made digital piracy tempting to engage in. The laws and regulations regarding digital piracy are often unclear and governments are struggling to control the unlimited share of files, because it seems to be an impossible operation. However, the fact remains that digital piracy is illegal (in most countries) and a serious concern that impacts both society and the economy (Walls, 2008).

Given the nature of digital piracy people need a particular feeling to be comfortable with it: trust. Trust becomes important in situations that are perceived risky or vulnerable, such as transacting online (Schlosser, White and Lloyd, 2006). Digital piracy can be seen as a very risky online ‘transaction’, since it is also illegal to engage in.

(7)

Many studies examined factors that influence digital piracy, but little research is conducted on how online trust and legitimacy affect the relationship between attitude and behaviour of digital piracy. Online trust might be an important moderator of the relationship between legitimacy and the behaviour of illegally downloading a movie. People may have a positive attitude toward unethical behaviour and are therefore likely to download movies from a torrent website for example, but when the trust in the supplier (e.g. The Pirate Bay) is low, this may decrease the willingness to download. Moreover, differences in legitimacy also may influence the attitude-behaviour relation. For example, people who live in a country where the prices for a movie ticket are fairly high might have the intention to download the movie, but not engage in the actual behaviour because of strict laws and regulations regarding digital piracy.

Even though literature has been paying attention to the role of trust in purchase intentions online, little research has been conducted on the role of trust in an unethical environment and how this influences the behavioural intention of consumers and their actual behaviour. Besides that, trust can be perceived differently depending on the country because of different laws and regulations, which, to the best of my knowledge, has not been researched yet. In this study, we consider online trust as a moderator between legitimacy and the illegal download of movies.

This study focuses on the effect of online trust and legitimacy and their influence on the attitude and behaviour to (illegal) download movies. When rules and regulations are strict, the illegal download of a movie might have serious consequences. Therefore it is relative more risky to do than in a country where governments are not so strict regarding digital piracy. It is interesting to examine whether online trust can affect the direct effects between legitimacy

(8)

and the number of movie downloads. Whereas other studies mainly focus on examining the behavioural intention, this study also added a second study wherein real data concerning movie downloads are analysed to consolidate the research.

1.2 Research question and sub questions The research question is established as follows:

What is the effect of online trust in the supplier on the relationship between legitimacy and the number of movie downloads?

The sub questions that should contribute to answer the research question are: - What is legitimacy?

- How does attitude toward unethical behaviour and rule of law influence the intention to download movies?

- Why is trust in the supplier important for online purchases? - What is the effect of trust on attitude and behavioural intention?

1.3 Scientific relevance

This research makes an important scientific contribution. It focuses on examining the effect of legitimacy and online trust in the supplier, since the influence of these factors is still largely unclear and unknown. This study extends literature about important aspects for consumers to engage in illegal download behaviour. Several factors influencing unethical decision-making has been examined, but research regarding the effect of online trust and legitimacy is lacking.

(9)

1.4 Managerial relevance

The study is relevant for practitioners and governments as well. Governments and

practitioners did not find a successful manner to decrease digital movie piracy (Al-Rafee and Cronan, 2006) so far. Furthermore, a high percentage (35,2%) of the content that has been downloaded consists of movies (Yale University, 2011). Therefore there is a growing need for research on how trust and legitimacy can be regulated to decrease digital movie piracy.

1.5 Structure

The structure of this research is as follows. The first chapter served as an introduction to the field of research and to denominate why this study is needed. The second chapter discusses the theoretical background that forms the framework of this study. It provides a critical review on the core concepts of the study and introduces the proposed hypotheses of this study. Chapter 3 is a detailed description of the adopted method in this study. It explains how data is collected and which methods and models are used to do this. Moreover, the measures that have been used to examine each construct are discussed. In chapter 4 the empirical findings of the study are presented. The collected data have been analysed using the methods described in chapter 3. Thereafter, the proposed hypotheses are tested. Chapter 5 discusses the findings presented in chapter 4 and aims to link these results back to the theoretical

framework. Subsequently, an answer to the research question is provided. The main purpose of this chapter is to critically reflect on the findings and to elaborate on how this study

contributes to the literature and provides new insights for practitioners. In het final chapter the limitations of this study are discussed and suggestions for future research are given.

(10)

2. Theoretical framework

This chapter discusses the most relevant findings from the current literature about the key concepts of this study and introduces the hypotheses that have been tested. Initially, the key concepts of digital piracy, legitimacy and online trust are described. Subsequently, this chapter outlines how online trust in the supplier influences the relationship between legitimacy and the behaviour of illegally downloading a movie. The chapter ends with a research model that graphically illustrates the stated hypotheses.

2.1 Digital piracy

Digital piracy is defined as “the illegal copying/downloading of copyrighted software and media files” (Cronan and Al-Rafee, 2008, p. 528). According to Cronan and Al-Rafee (2008) digital piracy is a growing societal problem that contains serious costs and consequences, especially considering the fact that this behaviour is illegal in most countries. Earlier studies regarding digital piracy mainly focused on the software industry, while the consequences of digital piracy are also affecting the entertainment industry, with serious implications for marketers in this industry. The entertainment (and in particular the motion- picture) industry is highly dynamic because of new, faster and easier technologies, the on-going releases of new movies and rules and regulations that are constantly changing. The worldwide motion-picture industry faced a self-estimated revenue loss that exceeded $18 billion in 2005 (MPAA, 2005). Even though companies and governments are constantly trying to decrease digital piracy, both prevention and deterrence do not seem to work (Al-Rafee and Cronan, 2006). Multiple scholars describe the illegal download of movies as a significant threat to the growth of the motion-picture industry (Yoon, 2010; Liao, Lin & Liu, 2009).

It is important for both management and practice to understand why people would engage in illegal activities such as digital piracy. Digital piracy is considered unethical behaviour, but

(11)

many studies have claimed that individuals do not see piracy as an unethical issue or a crime (Al-Rafee and Cronan, 2006; Reid et. al, 1992). While the illegal downloading of movies increased in recent years, many consumers do not feel that downloading causes harm to companies or industries (Levin, Conway Dato-on and Rhee, 2004). According to them, people do not feel that one single download could harm the artists or producers and they feel like the business deserves it since the financial risk of buying a CD with only a few good songs is high. Their research shows that people that are downloading have lesser ethical concern and a greater willingness to engage in other unethical behaviour (Levin, Conway Dato-on and Rhee, 2004). Even though Levin, Conway Dato-on and Rhee focused on the music industry, the underlying reasons they mention are highly transferrable to the movie industry. Despite the fact that downloading is illegal behaviour, the threshold to do so is not very high. Not only can you obtain a movie for free which saves you money, but the great speed of Internet also created timesaving opportunities and made it a lot more attractive for people to download (Levin, Conway Dato-on and Rhee, 2004). Moreover, rules, regulations and therefore consequences of digital piracy are not always clear.

2.2 Legitimacy

Legitimacy can be defined as “a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions” (Suchman, 1995, p. 574). Noteworthy about this definition is that legitimacy is a perception or assumption, meaning that it is subjective. Second, it is based on the actions or behaviour of an entity. This means that it represents the shared norms, values, beliefs and definitions of some social group (Suchman, 1995). In his study, Suchman lists three dimensions of (organizational) legitimacy: pragmatic, moral and cognitive

(12)

the ‘rightness’ of a particular action. It focuses on the question: ‘Is it the right thing to do?’ (Thomas and Lamm, 2012). Even though the articles of Suchman (1995) and Thomas and Lamm (2012) are concentrated on organizational legitimacy, they provide an overview of how and when people perceive specific actions or behaviour as legitimate. As legitimacy is an abstract concept that knows not only one operationalization, for practical reasons our two different studies use slightly different versions that both theoretically correlate highly with legitimacy: ‘attitude toward unethical behaviour’ and ‘rule of law’. Both can be categorized under the umbrella of legitimacy. The attitude toward unethical behaviour of a particular group of people with shared norms, beliefs, values and definitions can also be explained as the legitimacy of that particular group. Rule of law is primarily based on a set of beliefs of what is acceptable in a specific country and therefore this term is considered as a proxy for legitimacy in this study.

2.2.1 Attitude toward unethical behaviour

A number of studies in the area of marketing focus on consumer practices that are considered to be unethical (Levin, Conway Dato-on and Rhee, 2004; Yoon, 2010; Liao et. al, 2010). The main reason for this is that unethical consumer practices can have a major effect on company profits. According to Muncy and Vitell (1992) consumer ethics can be defined as “the moral principles and standards that guide behaviour of individuals or groups as they obtain, use and dispose of goods and services” (p. 298). In other words, consumer ethics are the guidelines of what can be seen as proper behaviour and what not. According to Tan (2002) researchers are increasingly interested in how and why consumers make decisions to engage in unethical behaviour. This is mainly because decisions regarding unethical behaviour are based on a specific group of stimuli that affect consumers’ ethical decision process (Tan, 2002).

(13)

Many studies examined factors that lead to the intention to engage in digital piracy, which is considered to be unethical behaviour (Cronan and Al-Rafee, 2008; Yoon, 2010; Liao, Lin and Liu, 2009). Cronan and Al-Rafee (2008) suggested that attitude, perceived behavioural control, moral obligation and past piracy behaviour have a significant influence on the intention to commit digital piracy, whereas ‘past piracy behaviour’ was the strongest

indicator. A study conducted by Yoon (2010) also examined factors influencing the intention to commit digital piracy. The results of his study indicated that attitude, subjective norms, moral obligation and perceived behavioural control were of significant influence on the behavioural intention, whereas moral obligation had the largest effect. Liao et al (2009) indicate that attitude, perceived behavioural control affects the intention to use pirated software. Tan (2002) examined consumer’s ethical attitude as one the key determinants influencing software piracy. In all these studies attitude towards the behaviour has a significant effect on the behavioural intention.

As said before, attitude toward digital piracy is different than other unethical behaviours because digital piracy is not perceived to be unethical (Al-Rafee and Cronan, 2006; Reid et. al, 1992). However, digital piracy is illegal and therefore involves risk. Perceived risk is one of the key variables that determine ethical decision-making so people should take this into account when they make their decision regarding unethical behaviour (Tan, 2002). The fact that downloading movies is illegal makes it interesting to study, since people are likely to form their attitudes differently and other factors might influence the relation between attitude and behaviour. This study elaborates on the study of Levin, Conway Dato-on and Rhee (2004), which state that downloaders have lesser ethical concern than non-downloaders. We examine the direct effect between people’s attitude toward unethical behaviour in general and the number of movies they download with a work hypothesis under the main effect of

(14)

legitimacy. It is expected that there is a negative relationship between attitude toward unethical behaviour and the number of movie downloads, meaning that people with a high ethical concern (i.e. a low score) download fewer movies than people with a favourable attitude toward unethical behaviour (i.e. a high score).

2.2.2 Rule of law

Rule of law can be defined as “the degree to which the behaviour of individuals and organizations (including government authorities) complies with formal legal rules” (Licht, Goldschmidt and Swartz, 2003, p. 664). Countries that are characterized by a strong rule of law usually have a strong and independent court system and solid (political) institutions (Oxley and Yeung, 2001). A strong rule of law has several features, which will be discussed hereafter. A strong rule of law leads to greater stability and transparency concerning the boundaries of acceptable behaviour. This means that people are better aware of what is appropriate behaviour or not and under which circumstances law protects them or not. Moreover, a strong rule of law enables companies to build up a credible and trustworthy reputation, since breaking the law may lead to high sanctions. Finally, a strong rule of law increases the level of trust people have in markets and contracts by influencing their overall attitudes (Oxley and Yeung, 2001). This trust is mainly important in e-commerce and will be further discussed in section 2.3. In countries with a weak rule of law, this lack of trust might lead to a greater need for other safeguard for people to protect their privacy (Steenkamp and Geyskens, 2006). Different countries have different laws, rules and policies and consequently, they will differ on how strong their rule of law is. Since digital piracy is illegal, anti-piracy legislation is one of the elements that are affected by this rule of law. Countries with a strong rule of law are likely to be stricter about piracy behaviour as well. As discussed before, there is a growing body of research regarding the economic effects of digital piracy. However, little

(15)

research has been conducted about how anti-piracy legislation affect people’s behaviour of committing digital piracy (Orme, 2014).

A strict rule of law might safeguard consumers’ interests. A provider in a country with very strict data regulations may not be allowed to give up information on the people behind ip-addresses, which may give consumers a ‘safe’ feeling. This then, will lead to a higher number of movie downloads. However, following the theory of Oxley and Yeung (2001) we expect that a strict rule of law has a direct negative effect on downloading movies. When you would like to see a movie but you do not want to pay for a movie ticket, you might have a favourable attitude towards illegally downloading a movie. However, if you know that legislation in the country you live in is very strict and you face serious consequences when getting caught, it might stop you from actually doing it.This is likely to weigh more heavily than the feeling of being protected by your service provider. The following hypotheses regarding legitimacy have been formulated:

H1: There is a negative relationship between legitimacy and the number of movie downloads, meaning that when the rules and regulations are stricter, the number of movie downloads are lower.

H1a: Ethical attitude has a negative effect on the number of movie downloads, meaning that people with a high ethical concern download fewer movies than people with low ethical concern.

(16)

2.3 Online trust

Trust is a term that is mainly studied in relationship marketing research. It consists of several important aspects; it is a psychological state (Rousseau, Sitkin, Burt and Camerer, 1998), has to do with the willingness to be vulnerable and to rely on the intention and behaviour of others (Jevons and Gabbott, 2000; Rousseau et. al, 1998; Morgan and Hunt, 1994). According to Becerra and Korgaonkar (2011) trust can be defined as “the willingness of a party i.e. trustor, to be vulnerable to the actions of another party i.e. trustee, based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to control or monitor that other party” (p. 938). This research is focused on online trust, which is different from offline trust in the following ways. With online trust, the object that needs to be trusted can either be the Website, the Internet or the technology. In contrast, with offline trust the trust needs to be formed around other persons or a particular store. The way of building trust can be somewhat similar and has to do with interaction and familiarity. When a consumer becomes familiar with a website, has positive impressions and accepts vulnerability, trust can be built (Bart, Shankar, Sultan and Urban, 2005).

Online trust is important when people engage in perceived risky, uncertain and/or vulnerable situations, such as transacting online (Schlosser et. al, 2006; Gefen and Heart, 2006). Online trust can help consumers to decrease uncertainty and risk perceptions and allows them to start building on a trust-based relationship with Web vendors. It is online trust that ensures that consumers are willing to depend on a website, share personal information, make an online transaction or acting on a Web vendor’s advice (McKnight, Choudhury and Kacmar, 2002). Becerra and Korgaonkar (2011) distinguished between three different kinds of online trust: brand trust beliefs, product trust beliefs and vendor trust beliefs. Trust in the brand is likely to influence consumers’ decision-making process since a brand contains valuable information.

(17)

This is especially important online, where consumers cannot feel or inspect a product. Trust in a product may differ across product type, newness, personal experiences, (lack of)

information and the inability to inspect the product (Becerra and Korgaonkar, 2011). Last but not least is trust in the vendor. This is likely to decrease consumers’ perceived online

uncertainties. Since Internet is ubiquitous, there is physical distance between buyer and seller and relationships are difficult to predict and interpret, several marketing studies find that trust in the online vendor is especially important for consumers to deal with the perceived risk in an online transaction (Mukherjee & Nath, 2007; McKnight et. al, 2002). Therefore, the focus in this research is on online trust in the vendor. Both competence (Schlosser et al., 2006) and benevolence and integrity of e-tailers (McKnight et al., 2002) influence the consumers’ intention to share personal information or to make an online transaction.

Linking online trust to this specific research, downloading a movie involves several potential risks. Firstly, by downloading a movie users can inadvertently download a virus onto their computer. Secondly, the quality of the downloaded movie is not guaranteed. Thirdly, by downloading a movie, people share personal information (ip-address), that could link the illegal activity to them, with which they effectively place themselves in a potentially

troublesome situation. In such uncertain situations, where consumers have to make a decision whether they will download or not, trust can be an important aspect (Schlosser et al., 2006). It can become a crucial element in consumers’ decision-making since it reduces the perceptions of uncertainty, risk and/or vulnerability (Kim and Benbasat, 2006 in Becerra and Korgaonkar, 2011). Therefore, we expect a main effect of online trust on downloading movies, as trust will influence consumers’ online intentions (Gefen and Heart, 2006). Moreover, online trust in the supplier is expected to be an important moderating factor on the relation between legitimacy

(18)

and the actual download of a movie. The following hypotheses are related to these expectations:

H2: There is a direct, positive effect between online trust in the supplier and the number of movies, meaning that when the online trust is high, the number of movie downloads also will be higher.

H3: Online trust in the supplier moderates the negative relation between legitimacy and the number of movie downloads, so that that the negative relationship is stronger (weaker) when the online trust in the supplier is lower (higher).

2.4 Conceptual framework

This conceptual framework visualizes the four work hypotheses that have been tested during two studies. It presents the relationships between the dependent variable (movie downloads) and the independent variables (attitude toward unethical behaviour and rule of law

respectively). Furthermore, the direct and interacting effect of online trust in the supplier is incorporated in the model.

Figure 1. The effect of online trust in the supplier on the relationship between legitimacy and movie downloads

(19)

3. Method

This chapter presents the method that is adopted to collect the data in this study. First, the data collection procedure is explained. Second, a detailed description of the sample is given and finally, the measures that have been used to test the variables are described.

3.1 Data collection procedure

A distinguishing feature of this research is that two complementary studies are taken to test the theory and to answer the proposed research question and sub questions. In the first study, the relationship between attitude toward unethical behaviour and the number of movie downloads, the direct effect of online trust and the moderating role of online trust in the supplier are examined. However, as this study only investigates behavioural intentions and does not include real behaviour, a complementary second study was executed. The second study is added to extend the results and to investigate how the rule of law per country is influencing movie downloads, using a dataset that comprises of actual download behaviour. Although it was impossible in these studies to use exactly the same constructs, highly

theoretically correlated constructs were used to mimic the setting in both studies to the largest extent possible. In specific, attitude toward unethical behaviour and rule of law are highly correlated, since they both represent the legitimacy of a particular group of people. In the first study, a cross-sectional quantitative study by means of a survey is conducted. The use of a survey as the data collection instrument is considered to be suitable because this research is explanatory. Moreover, a survey is a well-known and frequently used method to obtain data about respondents’ opinion and behaviour, and in this specific instance provides individual level data on the constructs discussed in the theory section. The survey is designed in

Qualtrics and spread both online and offline to reach a wide set of potential respondents. Prior to the collection of the data, a pre-test was executed among a representative group of

(20)

respondents (n = 20) and small adjustments in the survey have been made accordingly. Despite the suitability of a survey, it does not come without drawbacks. A usual, and in our setting potentially problematic problem is the social desirability bias and the potential lack of assurances for anonymity. A short introduction to the survey has been given to inform respondents about the confidentiality. Moreover, the pre-test allowed us to verify that respondents interpreted the questions in the intended way. In the second study, quantitative database analysis is used to complement study 1. In this study, we use secondary data that has been collected in 2012-2013. A big advantage of the use of this dataset is that it provides data about movie downloads for a period of approximately six months. This provides us

unobtrusive, longitudinal data that is a very useful addition to the first study. Section 3.3 provides a detailed description of which variables have been added to complete the dataset.

3.2 Sample

In order to collect the appropriate data for the questionnaire, non-probability sampling has been used. Prerequisite for the sampling frame was that the respondent was between 15-34 years, since this is the group that downloads the most according to Poort and Leenheer (2014). The authors’ network was used in order to reach a large group of people. Moreover, since it was expected that many respondents would fall within the categories 20-24 and 25-29 years, offline surveys have been spread at a high school in Sleeuwijk (the Netherlands) to reach younger respondents. This not only raised the number of respondents, but the debriefing and feedback also provided better insights in the opinion of the sample. Moreover, high school students are expected to download movies often. A total number of 359 surveys were collected within the distribution period of three weeks. Of these 359, 34 were incomplete, resulting in 325 unique and valuable respondents. However, we are interested in finding a relationship between legitimacy and the number of movie that respondents download. So it is

(21)

interesting to test those respondents that have both the technical knowledge and the right conditions (e.g. proper internet connections, a computer/laptop) to download a movie. Because of this we decided to exclude respondents that indicate they have downloaded no movies since January 2015. Moreover, people that have downloaded more than 100 movies are not representative for the sample size, so we also decided to make a cut-off point at 100 downloads. After excluding these respondents, 225 unique respondents are left. For a complete description of the sample, see Appendix III.

3.3 Measure development Study 1

In the first study, the following links are examined:

- the direct effect of ‘attitude toward unethical behaviour’ on ‘movie downloads’ (H1a) - the direct effect of online trust on ‘movie downloads’ (H2)

- the moderating role of online trust on this relationship (H3)

Figure 2. The moderating effect of online trust on the relationship between attitude toward unethical behaviour and movie downloads

(22)

As already explained in section 3.1, this study is performed by means of a cross-sectional survey. To maximize the reliability measures in this study are based upon measures used in previous studies. These are discussed below.

Attitude toward unethical behaviour

To investigate consumers’ attitude toward unethical behaviour, the 5-point Likert scale of Muncy and Vitell (1992) is used. Muncy and Vitell (1992) developed a 20-item 5-point Likert scale to measure respondents’ attitudes toward unethical behaviour. This list includes several ‘unethical’ activities such as ‘getting too much change and don’t say anything about it’ and ask respondents to indicate whether they ‘strongly believe it is wrong (anchored as 1) or ‘strongly believe it is not wrong’ (anchored as 5). A lower score indicates a greater ethical concern. To measure the construct attitude toward unethical behaviour a final set of three of the original items were used, this is discussed in section 4.1.2. A discussion of the pre-test and what has been changed can be found in Appendix I.

Online trust

The items that are used to measure online trust in the supplier came from Becerra and Korgaonkar (2011), which based it on the original scale of McKnight et al. (2002). The original scale of McKnight et al. (2002) has a Cronbach’s alpha of 0,96 and the reliability of the study Becerra and Korgaonkar (2011) are all between 0,89 and 0.95, which is considered very good. The items that were selected measure respondent’s disposition to trust (e.g. I generally trust others), vendor trust (e.g. I would rely on vendor…) and channel trust (e.g. Online vendors act in consumers’ best interest). The items were tested on a 5-point Likert scale to ensure that the survey contains consistent scales. After the pre-test, three items are

(23)

deleted in order to improve the reliability of the scale and a total of eleven items have been used. In the final setting, three items have been used; this is discussed in section 4.1.2.

Movie downloads

To demonstrate how legitimacy and online trust can affect the number of movie downloads, respondents had to indicate how many movies they have been downloading since January 2015. This item is formulated as an open question in order to prevent social desirability bias (i.e. when respondents are faced with categorical options, these may lead them to adjust their reported behaviour downwards/upwards to fall within a more socially acceptable bracket). As explained in the description of the sample, respondents that have downloaded more than zero and less than 100 are retained for data analysis. For the full questionnaire, see Appendix II.

Study 2

Study 1 examines the intention of movie downloads, but not the real behaviour. Therefore, study 2 is added, which includes both database analysis and theoretical research. After extensive deliberation and inspection of possible analogous option, rule of law was used in study 2. Rule of law, which is strongly correlated with attitude toward unethical behaviour, is theoretically investigated in different countries and compared with statistical data of movie downloads in those specific countries. In short, this study aims to answer the question: does rule of law affect actual downloading behaviour? Moreover, this study aims to investigate the role of online trust in this situation, both the direct and the interaction effect. Hence, Study 2 tests H1b, H2 and H3 as the empirical analogue of the hypotheses in study 1. In the remainder of this section the sample frame and operationalization of all variables in study 2 will be discussed.

(24)

Figure 3. The effects of online trust on the relationship between rule of law and movie downloads

Rule of law

To investigate the rule of law and how this affects the number of movie downloads a subset of countries was selected. Following the statistical data that is represented about the actual number of movie downloads, the following countries are selected: Australia, Belgium,

Canada, China, Germany, Denmark, Spain, Finland, France, the United Kingdom, India, Italy, Japan, the Netherlands, Norway, New-Zealand, Russia, Sweden, Ukraine and the United States, mainly based on the research of Steenkamp and Geyskens (2006).Similar to

Steenkamp & Geyskens (2006) the score for the rule of law per country is extracted from the World Databank. The World Databank provides statistical worksheets with objective and up-to-date scores for the rule of law per country, so these ratings are used. The score for each country lies between -2,5 and 2,5 and indicates how weak (negative) or how strong (positive) the governance performance in a specific country is. The data used is from 2013, which corresponds well to the data on movie downloads.

(25)

Movie downloads

In order to analyse the effect on movie downloads we use a dataset that compromises of all torrent downloads from week 42, 2012 to week 31, 2013. This dataset has weekly information per country on the number of movie downloads.

Online trust

The weekly nature of the movie download data allows for our analysis to be conducted on a weekly level. Hence, in this study we aim to find a way to measure online trust on a weekly level. In a particular week, online trust can be very high, but a disruptive event in a specific country is expected to decrease the online trust in the supplier and therefore affects the

number of movie downloads. A set of twenty countries was selected, for which data on online trust was collected. These countries are associated with digital piracy. Moreover, disruptive events that may have influenced the online trust have been found in the timeframe for which statistical data was available (week 42, 2012 to week 31, 2013). These disruptive events vary from changes in anti-piracy laws to shutdown of major piracy websites. However, they provide a clear insight in what effects a change in online trust provokes. Appendix IV provides an overview of the weeks per country and the specific week in which a disruptive event took place. Since it is difficult to evaluate how many weeks a disruptive event has an effect, step dummy variables are used, implying that there is a long-lasting effect.

Control variables

The population per country and the week number are included as control variables. In this way, it is possible to check whether the time of the year or the number of people living in a country also influence the number of movie downloads. Data about the population per country are extracted from Worldometers (2014). Moreover, data regarding people’s

(26)

ownership of a computer for each country has been added as a control variable. This item is could be correlated with rule of law, but moreover having a computer or laptop is an important condition to be able to download a movie. Data for this item are extracted from Nationmaster (2005).

(27)

4. Results

In this chapter the key results of both study 1 and 2 are discussed. First, the variables and measurements are presented and second, the proposed hypotheses are tested.

4.1 Results study 1

4.1.1 Variables and measurements Reliability of scales

First, the reliability for the construct items was measured. The original scales appeared to be reliable (all Cronbach’s Alpha > 0,7), as displayed in Table 2.

N = 225 Cronbach’s Alpha Number of items M SD

Ethical attitude .760 12 35,94 6,401

Online trust .727 11 31,50 5,715

Table 1. Reliability of scales

The dependent variable in study 1, the number of movie downloads, was measured using an open question (i.e. What is the approximate number of movie that you download last year, from January 2015?).

Correlations Ethical

attitude

Online trust Interaction Gender Age Education Ethical attitude 1 Online trust ,104 1 Interaction -,215** -,065 1 Gender -,188** ,019 ,151* 1 Age -,068 -,035 ,032 ,001 1 Education -,068 -,014 ,063 ,018 ,850** 1 ** Correlation is significant at the 0.01 level (2-tailed)

* Correlation is significant at the 0.05 level (2-tailed)

Table 2. Correlation table study 1

As we can see from the correlation table, the independent variables age and education are highly correlated (r = .850) on a significant level (p < .01). As a result, we have to be cautious for multicollinearity. Multicollinearity between independent variables makes it impossible to

(28)

obtain unique estimates of the regression coefficients (Field, 2013). Even though these numbers are not met, age and education are not perfect in line with the other variables. Moreover, Field (2013) indicates that removal or addition of one predictor variable can result in enormous changes to the model. Therefore we remove the control variable age, since this improves the collinearity statistics and allows us to obtain a better model. Moreover, to reduce multicollinearity in the model, all variables have been centered.

4.1.2 Data preparation

In order to make sure the proper items are used for the hypothesis testing, we ran factor analyses on the constructs ethical attitude and online trust. As a first stage selection

mechanism, we retained twelve items for ethical attitude and eleven items to measure online trust, but given the large number of questions we can be more selective and specific without risking reliability. In order to reduce the dimensionality, we ran two factor analyses. Ethical attitude appeared to have four different factors, from which the first factor has the strongest link with online behaviour. The following items belong to this factor: using computer software or games that you did not buy, downloading a movie or music instead of buying it and taping a movie off the television. These three items are used to measure the construct ethical attitude. The reliability of these three items appeared to be .705, which is still considered to be appropriate. A new, centred variable has been computed for the multiple regression analysis.

(29)

Table 3. Overview of response regarding online trust

The original construct of online trust contains eleven items. From table 3 it becomes very clear that respondents distinguish between ‘disposition to trust’ and ‘online trust’, even though this was not explicitly communicated in the survey. People generally trust others (64.3%) and have a high tendency to trust (45.5%). However, when it comes to online trust and online vendors are included, people are not so trusting anymore. Especially when

incorporating an illegal website like The Pirate Bay, online trust in the supplier decreases. The main observation from this table is that people are not comfortable providing personal

information to such a website. The factor analysis suggested three different factors for the construct online trust. We selected those three that were strongly linked to the Pirate Bay, as we are interested in testing online trust in the supplier (i.e. The Pirate Bay is one of the suppliers for illegal movie downloads). Those items are represented as bold in table 8. The reliability turned out to be .701, which is still considered to be appropriate. Using a subset of three items for online trust also slightly increased the variance that is explained by the model. The centred sum of these three items is computed as a new variable.

Item N Totally not

agree

Not agree I do not have an opinion

Agree Totally Agree I generally trust others 322 7 2.2% 61 18.8% 18 5.5% 209 64.3% 27 8.3% I generally have faith in

humanity

322 16 4.9% 71 21.8% 44 13.5% 174 53.5% 17 5.2% I tend to trust something even

when not knowing it

322 17 5.2% 143 44.0% 56 17.2% 102 31.4% 4 1.2% My tendency to trust is high 321 13 4.0% 89 27.4% 56 17.2% 148 45.5% 15 4.6% It is easy for me to trust others 322 11 3.4% 96 29.5% 35 10.8% 155 47.7% 25 7.7% I am comfortable relying on

Internet vendors

322 19 5.8% 132 40.6% 68 20.9% 99 30.5% 4 1.2% Online vendors act in

consumers’ best interest

322 26 8.0% 120 36.9% 89 27.4% 79 24.3% 8 2.5% I would rely on www.thepiratebay.org 322 35 10.8% 126 38.8% 84 25.8% 62 19.1% 15 4.6% I would trust www.thepiratebay.org 322 47 14.5% 144 44.3% 61 18.8% 62 19.1% 8 2.5% I am comfortable providing personal information to www.thepiratebay.org 322 113 34.8% 147 45.2% 37 11.4% 22 6.8% 3 0.9%

Internet’s legal and other

(30)

N = 225 Cronbach’s Alpha Number of items

Ethical attitude .705 3

Online trust .701 3

Table 4. Reliability of reduced items

4.1.3 Hypothesis testing

Figure 4. The effect of online trust on the relationship between attitude toward unethical behaviour and movie downloads

Hypothesis 1a, 2 and 3 are examined in the first study. Since all variables are scale variables, multiple regression analysis has been used to test the hypotheses. The regression equation for the first study is as follows:

! = !! +!!_!!_!+ !_!!_!+ !_!!" + !_!!_!+ !_!!_!+ !

The number of movie downloads is the outcome variable (Y), ! is the constant and ! is the residual or error, the part that is not explained by the model. In this research, the effects of five independent and control variables are examined. Firstly, we examine the direct effect between attitude toward unethical behaviour (!_!) and the number of movie downloads. Second, the direct effect between online trust and the number of movies downloads is examined. Third, we consider the interaction effect of online trust and ethical attitude (!!).

(31)

Hypothesis 1a

Hypothesis 1a suggests there is a negative relationship between attitude toward unethical behaviour and the number of movie downloads. In other words, it is expected that people with a great ethical concern download fewer movies than people with a low ethical concern. A multiple regression analysis is performed to test whether the independent variables have a predictive effect on the dependent variable.

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate

1 ,232a _,054 _,032 _17,645

a. Predictors: (Constant), Ethical attitude, Online trust, Interaction effect, Gender, Education

Table 5. Model summary study 1

Table 5 shows whether the model is significantly good at predicting the outcome (Field, 2013). This table shows that the model is significant (F(5) = 2.471, p = .033 < .05). The model summary indicates that both the predictor and control variables together explain 5.4% of the variance in the outcome variable movie downloads. This means that there still is a large part of the variance in the outcome variable that cannot be explained by the regression model, a discussion of potential drivers of this follows at the end of the discussion of the results of study 2. However, the focus is on the part that we can explain, therefore we take a look at the individual coefficients in the regression results (table 6).

Coefficients Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 8,607 4,914 1,751 ,081 Ethical attitude ,706 ,517 ,092 1,366 ,173 Online trust 1,054 ,502 ,140 2,099 ,037 Interaction effect ,064 ,197 ,022 ,324 ,746 Gender -1,960 2,457 -,054 -,798 ,426 Education -1,998 ,944 -,140 -2,117 ,035

a. Dependent Variable: Movie downloads Table 6. Regression results study 1

(32)

Since the unstandardized coefficients are dependent on the units of measurement of the variables (Field, 2013), we will look at the standardized beta coefficients to interpret the effects. First of all, ethical attitude has a standardized coefficient of .092, which indicates that when the ethical concern increases with one standard deviation, the number of movie

downloads would also increase with .092 standard deviations. However, this effect is not significant (p = .173 > .05). In other words, having a favorable or unfavorable attitude regarding unethical behavior does not seem to have a significant effect on the number of movies that people download. Therefore, H1a is not supported in this study.

Hypothesis 2

Hypothesis 2 suggests a direct, positive effect between online trust and the number of movie downloads. As we can see from table 6, online trust has a positive, significant (standardized β = .140, p = .037 < .05) effect on the number of movie downloads. This result seems to

indicate that higher levels of trust in the online environment encourage more movie downloads. Therefore, H2 is supported.

Hypothesis 3

Hypothesis 3 suggests that the negative relationship between ethical attitude and movie downloads is stronger for high online trust in the supplier than for low online trust in the supplier. In order to check this, we return to table 6. Here we can see that the beta coefficient for the interaction effect (!_!) is positive (.022), but not significant (p = .746 > .05). Therefore we can conclude that the interaction effect does not have a significant effect on the

relationship between ethical attitude and the number of movie downloads. In other words, attitude toward unethical behaviour and online trust in the supplier are not co-dependent. H3 is not supported.

(33)

The model also shows the control variables gender and education. As we can see, both coefficients of the control variables are negative, but only education appears to be significant (p = .035 < .05). This means that when people have a higher education, they will download fewer movies. An explanation for this effect might be that people with a higher education have less time to download or more money to watch movies in an alternative or legal way (e.g. going to a cinema), or perhaps are more susceptible to potential negative externalities of engaging in downloading movies.

In sum, we can conclude that the results of study 1 are not completely satisfying for several reasons. Statistical evidence neither is found for a direct relationship between legitimacy and the number of movie downloads, nor for the moderating effect of online trust. However, online trust did appear to be have a direct, positive effect and the control variable education also is significant. Besides this lack of statistical evidence for some hypotheses, it is

disappointing that we can only explain 5.4% of the variance with these independent variables. This may be due to the fact that in study 1 we examine intentional behaviour and not real behaviour. Moreover, it is also possible that during the pre-test, some items were deleted that were essential to examine the construct of online trust. Unfortunately, results have been fairly inconclusive in study 1. The results of the second study are presented in the next section.

(34)

4.2 Results study 2

4.2.1 Variables and measurements Rule of law

Study 2 provides insight in the relationship between rule of law in a specific country and the number of movie downloads within a time period of approximately six months (week 42, 2012 to week 31, 2013). Moreover, the direct and moderating effect of online trust in the supplier are investigated. As mentioned in the method section, twenty countries have been compared. Rule of law is the independent, predictor variable in this study (X). Below a graph with respectively the rule of law per country and the number of movie downloads is

presented.

Figure 5. Histogram of rule of law and movie downloads

RU CN IN UA IT ES BE FR _JP US DE GB CA AU NL NZ DK FI SE NO 0 500 1000 1500 2000 2500 3000 -0,78 -0,46 -0,1 -0,08 0.36 1 1.40 1.40 1.41 1.54 1.62 1.67 1.74 1.75 1.81 1.86 1.87 1.93 1.95 1.97 M ovi e d ow n load s (x1000) Rule of law

(35)

Correlations Rule of law Movie downloads Online trust Interaction effect Population Week Pc ownership Rule of law Movie downloads Online trust Interaction effect Population Week Pc ownership 1 -,074* 1 ,201** _,169** ₁ ,030 ,289** ,831** 1 -,604** _,355** _-,104** _-,031 ₁ ,000 ,296** ,539** ,478** ,000 1 ,885** ,004 ,173** ,039 -,580** ,000 1

* Correlation is significant at the 0.05 level (2-tailed) ** Correlation is significant at the 0.01 level (2-tailed) Table 7. Correlation matrix study 2

Pc ownership has a very high correlation with rule of law. However, collinearity statistics do not indicate serious problems. Therefore we decided to incorporate the variable in the model. Just like in study 1, all variables have been centered to reduce multicollinearity.

Movie downloads

Each country has unique data regarding the movie downloads in a specific week. This variable has been called MD (movie downloads) and is the dependent variable (Y) in the regression equation.

Online trust

Study 2 examines the relationship between rule of law and movie downloads and whether a change in online trust moderates this relationship. Since both the dependent, independent and moderating variable are scale variables, a multiple regression is performed to investigate the hypotheses. The regression equation for the second study is similar to the regression equation of study 1, however the meaning of the variables is slightly different:

(36)

As said before, the number of movie downloads is the outcome variable (Y). X_! represents the rule of law per country, X_! the direct effect of online trust, XM is the moderating effect of online trust in the supplier and X!, X! and X! are the control variables (week, population and

pc ownership respectively).!

4.2.2 Hypothesis testing Hypothesis 1b

Hypothesis 1b suggests that a strict rule of law results in a lower frequency of movie

downloads. Since both the independent variable (rule of law) and the dependent variable (the number of movie downloads) are scale variables and we want to examine the linear

relationship between X and Y, we run a multiple regression analysis on the data.

Model Summary

Model R R Square Adjusted R Square Std. Error of the Estimate

1 ,596a ,355 ,350 794,07730

a. Predictors: (Constant), Rule of law, Online trust, Interaction effect, Week, Population and PC ownership

Table 8. Model summary study 2

The ANOVA table indicates that the model is significant with F (6) = 71.247, p = 0.000. In the table above a summary of the model is presented. The value of R2 is .355, which means 35.5% of the variation can be explained by the model. In other words, the independent

variables ‘rule of law’, ‘online trust’, the interaction between rule of law and online trust and the control variables ‘week’ and ‘population’ together explain 35.5% of the variation in the number of movie downloads. This also means that 64.5% (100 – 35.5%) cannot be explained by the model and that other variables might have an influence too. However, the focus is on the part we can explain and therefore we take a look at the individual coefficients.

(37)

Coefficients Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) -1304,035 136,028 -9,587 ,000 Rule of law -30,072 74,063 -,025 -,406 ,685 Online trust -737,113 112,876 -,374 -6,530 ,000 Interaction effect 632,245 71,868 ,480 8,797 ,000 Population 1,451 ,097 ,556 15,031 ,000 Week 21,721 2,824 ,266 7,692 ,000 PC ownership 1,658 ,252 ,396 6,569 ,000

a. Dependent Variable: Movie downloads

b. Movie downloads (x1000), Population (x1000000) Table 9. Regression results study 2

The table above indicates the slope of the effects of the different variables and their

significance. As we can see, online trust as a predictor has a negative effect on the outcome and all other variables have a positive effect on the outcome. As stated before, because of the differences in measurements we look at the standardized beta coefficients to interpret the effects. Rule of law, which is suggested to have a direct, negative of effect on movie

downloads (H1b), appears to have a negative effect (standardized β = -.025, p = .685), but is not significant. So even though the coefficient indeed is negative, as expected, we cannot interpret the result. Therefore we can conclude that H1b is not supported and that legitimacy does not have an influence on the number of movie downloads.

Hypothesis 2

Hypothesis 2 suggests that online trust in the supplier has a positive, direct effect on the number of movie downloads. The standard coefficient for this direct effect is negative (-.374) and significant (p = 0.000). This seems to be in contrast with the proposed hypothesis,

however the negative coefficient should be interpreted with caution. When something

disruptive happens in a particular country and the online trust becomes lower, this is indicated by an increase in codes from 0 to 1. As the changes in online trust (from 0 to 1) generally

(38)

reflect a decrease, and not an increase, in online trust, the variable is essentially reverse-coded. As a result, a negative coefficient is in line with the findings in study 1 and reflects that an increase in online trust leads to a higher number of movie downloads. Therefore, study 2 also provides support for hypothesis 2.

Hypothesis 3

Hypothesis 3 suggests that online trust in the supplier moderates the relation between rule of law and movie downloads. More specifically, the (standardized) coefficient for the interaction term is .480 and significant (p = .000), implying that online trust more negatively influences the number of downloads for countries with a low rule of law. Thus, the effect of online trust and rule of law on the number of movie downloads are co-dependent. Therefore, H3 is supported.

Study 2 incorporated three control variables: population, week and pc ownership. Unsurprisingly, ‘population’ has the strongest effect (standardized β = .556) which is significant (p < .05). This means that if the population in a country increases with one

standard deviation, the number of movie downloads increases by .556 standard deviations. In other words, when the number of people living in a country rise, the number of movie

downloads also rise significantly. The control variable ‘week’ also turns out to have a positive and significant effect (standardized β = .266, p = .000). This means that the later in the year, the more movies are downloaded. This is likely to be related to the release date of new movies. The third control variable, pc ownership, also has a positive, significant effect

(standardized β = .396, p = .000). This is an expected result, since pc ownership is a condition in order to download a movie.

(39)

H1 There is a negative relationship between legitimacy and the number of movie downloads, meaning that when the rules and regulations are stricter, the number of movie downloads are lower.

Not supported

H2 There is a direct, positive effect between online trust in the supplier and the number of movies, meaning that when the online trust is high, the number of movie downloads also will be higher.

Supported

H3 Online trust in the supplier moderates the negative relation between legitimacy and the number of movie downloads, so that that the negative relationship is stronger (weaker) when the online trust in the supplier is lower (higher).

Partially supported

(40)

5. Discussion

This chapter elaborates on the findings of the study. First, a general discussion on the findings is given. Second, a thoughtful interpretation of the results is given based on the literature that formed the basis of this research. Finally, the practical implications are discussed.

5.1 General discussion

The core of the two studies was to examine the relationship between legitimacy and the number of movie downloads and how this relationship is affected by online trust in the supplier. Collectively, the studies provide valuable insights on these relationships. The following research question was phrased: What is the effect of online trust in the supplier on the relationship between legitimacy and the number of movie downloads? In order to provide a solid answer to this research question, the interpretation of the statistical outcomes of the proposed hypotheses are discussed first.

Figure 7. Final results study 1 and 2

The main direct effect examined whether legitimacy has a direct, negative effect on the

number of movies people download (H1). This has been examined with two work hypotheses. The first work hypothesis (H1a) suggested there would be a negative relationship between ethical attitude and the number of movie downloads. However, the multiple regression

(41)

analysis did not provide statistical evidence to support this hypothesis. This means that attitude toward unethical behaviour does not have a significant effect on the number of illegally downloaded movies. This is in contrast with the study of Levin, Conway Dato-on and Rhee (2004). Whereas they have showed that downloaders have lesser ethical concern, this study did not find a direct effect between people’s ethical attitude and the number of movie downloads. In the studies of Liao et. al (2009), Tan (2002), Cronan and Al-Rafee (2008) and Yoon (2010), attitude towards a specific behaviour has a significant effect on the actual behaviour, which is also not in line with this study. As explained by Al-Rafee and Cronan (2006) and Reid et. al (1992) people do not perceive digital piracy to be illegal. Moreover, Yoon (2010) explains that potential outcomes also play a role in people’s intention to pirate. Perceived benefits (positive outcomes) are weight against perceived risks (negative outcomes). Therefore a possible explanation for the lack of statistical evidence can be allocated to the perceptions people have regarding digital piracy. People might have an unfavourable attitude toward unethical behaviour, but if they do not perceive downloading a movie as illegal or when they feel the benefits are greater than the risks, they still might download a relative high number of movies. The second work hypothesis (H1b) suggested that a strict (positive) rule of law results in a lower number of movie downloads. In other words, when you live in a country that has very strict regulations regarding digital piracy, it was expected that the number of movie downloads is lower than in a country where the rules are not so strict. A multiple regression analysis has been performed in order to test this hypothesis, but no statistical evidence for this negative relation was found either. This result is in contrast with the expectation based on the theory of Oxley and Yeung (2001). They suggested that in a country with a strict rule of law there is more transparency and stability regarding illegal behaviour. This means that people living in countries with a strict rule of law should be aware of what is acceptable behaviour and what is not. Lack of significant support

(42)

for a negative relationship may be due to an explanation already discussed in the theoretical framework. It could imply that a strict rule of law ‘safeguards’ consumers’ interests. When a country has strict data regulations, it could be forbidden for providers to share personal information and ip-addresses with third parties. Therefore, consumers can feel more ‘safe’ and download a movie anyway. Moreover, it also could be that people living in a country with a strict rule of law are not better aware of what is acceptable behaviour and what is protected by law than people living in a country with a less strict rule of law. At least, it is not empirically tested whether it really works this way. Are people living in countries with a strict rule of law truly better aware under what circumstances law protects them or not? Further research is required to assess this assumption. A last alternative explanation may be allocated to the selection of countries. As can be seen in Appendix IV, twenty countries have been selected that can be linked to digital piracy. However, most of these countries have a positive score for rule of law, so the small dispersion of these scores may have limited the results. The third direct effect (H2) has been examined in both studies. Both studies found empirical evidence for the direct and positive relationship between online trust in the supplier and the number of movie downloads. So when trust in the online environment grows, this stimulates the number of movie downloads. This is in line with the theory of Gefen and Heart (2006), which states that online trust will influence consumers’ intentions. Moreover, it corresponds with the studies of Mukherjee & Nath (2007) and McKnight et al. (2002). They both state that trust in the online vendor is important for consumers to deal with the perceived risk in an online transaction, which corresponds with the illegal download of a movie.

The moderating effects are tested with one hypothesis that is tested in the two complementary studies. Hypothesis 3 suggests that the negative relationship between legitimacy and the number of movie downloads is stronger for high online trust in the supplier than for low online trust in the supplier. This hypothesis is based on the theory of Schlosser et. al (2006)

(43)

and Gefen and Heart (2006), which state that online trust is mainly important when people engage in perceived risky, uncertain and/or vulnerable situations, such as transacting online. Moreover, according to Kim and Benbasat (2006) online trust can influence consumers’ decision-making as it reduces perceptions of uncertainty, risk and vulnerability. However, in the first study the multiple regression analysis performed did not show any statistical evidence for this hypothesis. So, online trust in the supplier did not moderate the relationship between attitude toward unethical behaviour and the number of movie downloads. Potential

explanations for the lack of this statistical evidence can be found in the theory of Schlosser et. al (2006). In their research they have demonstrated across four studies that online trust does not always has an equally strong influence on online purchase intentions (Schlosser et. al, 2006). Moreover, the fact that downloading a movie is both free and illegal might also have an influence on the unexpected outcome. And finally, this result can also be related to the perception people have of digital piracy. When people do not perceive digital piracy as illegal or risky, online trust can be less powerful in influencing consumers’ decision-making.

In the second study, we did find statistical evidence for the moderating effect of online trust in the supplier. When performing a multiple regression analysis in that study, we find a

coefficient of 0.480 that is statistically significant (p = 0.000). In other words, a difference in online trust in a country has a significant effect on the relationship between rule of law and the number of movie downloads, above and beyond the impact of the other variables. This is in line with the theory of Kim and Benbasat (2006). Online trust has an important moderating effect on the relationship between rule of law and the number of movie downloads. It is interesting to see that online trust in the supplier has a significant moderating effect on the relationship between rule of law and movie downloads, but has no significant, moderating effect on the relationship between attitude toward unethical behaviour and movie downloads.

(44)

Several questions were posed in the beginning of this research that are important to now reconsider. Based on our findings, we must question the findings of study 1. The multiple regression analysis of study 1 explains very little of the variance in the model (5.4%), therefore we should be careful with interpreting these results. While the number of respondents was sufficient (N > 200), we must acknowledge that the results are close to random white noise and we were not able to find satisfying results. The model might be biased by unusual cases (outliers), however we took precautions by excluding respondents that did not download at all or downloaded more than 100 movies in the last months. Another explanation can be found in the construct of movie downloads. In het survey, this construct is measured by one open question. Considering the fact that a limitation of a survey might be that respondents are not motivated or willing to provide accurate and honest answers, this may have led to a non-representative measure.

When linking the two studies to the main research question: What is the effect of online trust in the supplier on the relationship between legitimacy and the number of movies downloads?, we can conclude that legitimacy does not have a direct, negative effect on the number of movies people download. Neither study 1, nor study 2 found statistical evidence for this expectation. Despite these results, the main focus of this research was on the effect of online trust in the supplier. The theoretical framework showed that online trust becomes important in risky, uncertain or vulnerable situations (Schlosser et. al, 2006; Gefen and Heart, 2006). Moreover, because of online trust, consumers engage in online transactions (McKnight et al., 2002). We can conclude that online trust in the supplier has a positive, direct effect on the number of movies people download. So online trust appears to be important for people when downloading a movie. Moreover, online trust in the supplier has a partially significant effect on the relation between legitimacy and the number of movie downloads. It is significant when

The relationship between legitimacy and movie downloads and the moderating role of online trust

MSc Business Administration

Master Thesis

The relationship between legitimacy and movie

downloads and the moderating role of online

Acknowledgements

Statement of originality

Abstract

Table of Contents

1. Introduction

2. Theoretical framework

3. Method

4. Results

5. Discussion