The digital age:
The relationship between privacy concerns
and self-disclosure in mobile applications
Master’s Thesis
MSc. Business Administration – Marketing
Annelies Braemer
Student number: 11408642 January 25th 2018
Thesis final version University of Amsterdam Supervisor: Alfred Zerres
Statement of originality
This document is written by Annelies Braemer who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of Contents Abstract………. 5 1. Introduction………..……… 6 2. Literature review……….……… 10 2.1 Self-disclosure………. 10 2.2 Motivation………..………. 12 2.2.1 Intrinsic Motivation……….………. 13 2.2.2 Extrinsic Motivation……….……… 14 2.2.3 Amotivation……….………. 15 2.3 Privacy concerns…….………..……….…………..….. 15 2.3.1 Privacy………. 15 2.3.2 Privacy Concerns………. 16 2.4 User trust.……….………..……….………….. 21 2.5 Conceptual Model……….…… 24 3. Method……..………...……….………..… 25 3.1 Procedure……….……….. 25 3.2 Measurements………...…...…………... 26 3.3 Sample………...………..………….. 27 4. Results………...………..……… 29 4.1 Respondents……….……….. 29 4.2 Data preparation……….……… 29
4.3 Testing the hypotheses.……….………. 32
4.3.1 Hypothesis 1……….……… 35
4.3.2 Hypotheses 2a + 2b………... 35
4.4 Overview of the hypotheses………..…. 36 5. Discussion……… 38 5.1 General discussion……….. 38 5.2 Theoretical implications………. 39 5.3 Managerial implications………. 40 5.4 Limitations……….. 41 5.5 Future research……… 42 References………. 44 Appendix………... 54
Appendix A: Measurement Scales………. 54
Appendix B: Demographics……….. 56
Appendix C: Histogram and Normal P-Plot of Regression Standardized Residual…….. 57
Abstract
This empirical study focuses on the relationship between privacy concerns and self-disclosure in mobile applications, particularly in the world of sports. Furthermore, it examines the effects of user trust and motivation on this relationship. It is anticipated that user trust plays a mediating role in that privacy concerns have a negative impact on user trust and user trust has a positive impact on self-disclosure. This research examines the moderating effect of motivation, by hypothesizing that with low privacy concerns, intrinsically motivated users will choose to disclose more information than extrinsically motivated users, and, with high privacy concerns there will be less of a difference in terms of self-disclosure between these motivation profiles. In order to test these assumptions, a survey is conducted among individuals who have downloaded a mobile sport application (N=294). The survey contains questions concerning the following topics: self-disclosure, privacy concerns, user trust, and motivation. The results indicate that user trust does indeed play a mediating role in the relationship between privacy concerns and self-disclosure in mobile applications. No evidence, however, is found for a moderating role of motivation in this study. Motivation is found to have a direct effect on self-disclosure in a mobile application, with explanatory power for extrinsic motivation.
Key words: privacy concerns, self-disclosure, extrinsic motivation, intrinsic motivation, user
1. Introduction
In 2016, consumers downloaded 149.3 billion mobile applications; the expectation is that, by 2021, this number will increase to 352.9 billion (Keith, 2017). A mobile application is a software application that can be used on a mobile device, such as a smartphone, tablet, or an iPod (Wang, Liao, & Yang, 2013). These applications can be purchased on the Internet and on distribution platforms such as the Apple App Store and Google Play. Such mobile applications function to assist users to connect to Internet services or make it more convenient for them to access the Internet using their mobile devices (Wang et al., 2013). Considering the last few years, it can be noted that mobile devices have undergone a rapid shift, as they were initially developed purely for telecommunication services. Nowadays, however, mobile devices are increasingly being used as platforms (Damopoulos, Kambourakis, Anagnostopoulos, Gritzalis, & Park, 2013). Damopoulos et al. (2013) state that, due to this shift, there is a greater risk that personal information will be misused. In order to use a mobile application, customers are often required to disclose personal information. In the current digital age, customers are increasingly willing to do so (Attrill & Jalil, 2011). Disclosing information in a mobile application can be seen as a form of self-disclosure, which is defined as ‘the process of making the self known to others’ (Jourard & Lasakow, 1958, p. 91). Self-disclosure means that people share information about themselves with others (Attrill & Jalil, 2011); this process can occur both between individuals and between individuals and organizations (Joinson, Paine, Buchanan, & Reips, 2008).
Due to this tendency, consumers have become increasingly concerned that their privacy may be violated (Norberg, Horne, & Horne, 2007). Privacy concerns refer to the assumptions regarding risks and the possible negative consequences that are associated with the sharing of information (Zukowski & Brown, 2007). Therefore, an individual’s choice to use, install, recommend or download a mobile application depends on whether he or she
considers the mobile application in question to be trustworthy (Yan, Dong, Niemi, & Yu, 2013). This implies that user trust is an important factor in the choice to download a mobile application. User trust refers to the likelihood that a user will believe that a mobile application is capable of performing the expected task (Yan et al., 2013).
Mobile applications focused on sports and health serve as an example of an area in which there is an increasing number of mobile applications being offered (Boulos & Yang, 2013). People often have various types of mobile applications installed on their mobile phones, such as those offered by their gyms, workout applications, applications that track one’s distance while running, applications that track the user’s weight, applications that counts steps, and so forth. Examples of these mobile applications are: 7 Minutes Workout, Kayla Itsines, Strava, Runtastic, Runkeeper, FitForFree, Sportcity, and so forth. To use these types of mobile applications, one is often required to disclose certain information or surrender a degree of your privacy. Examples of the information required include a user’s weight or height; applications may also require to track a user’s location or to participate in a community.
Within the world of sports, there are different reasons why people choose to work out. Pelletier, Tuson, Fortier, Vallerand, Briere, and Blais (1995) have translated into English and validated the Échelle de Motivation vis-à-vis les Sports and developed the sport motivation scale (SMS). They identified two types of motivation, namely intrinsic and extrinsic. Intrinsic motivation refers to an activity that is performed to receive inherent satisfaction, as opposed to a specific consequence; an individual engages in the activity for the fun or challenge that comes with it, not due to external pressures or rewards (Ryan & Deci, 2000). Extrinsic motivation, in comparison, refers to engaging in an activity for a specific outcome, in order to achieve an instrumental value. For example, a person who is extrinsically motivated does the activity to receive a reward or to avoid being punished (Ryan & Deci, 2000). It would
therefore be interesting to use the field of sports to measure the type of motivation. Taking into account how these different motivation profiles respond to the relationship between privacy concerns and self-disclosure, and the role that user trust plays in this relationship, the following research questions are formulated:
“What is the relationship between privacy concerns and self-disclosure in mobile
applications? What is the mediating effect of user trust on this relationship and how is this relationship moderated by motivation?”
This study has several theoretical and managerial implications. First, it contributes to the existing literature by providing new insights concerning the relationship between privacy concerns and self-disclosure in mobile applications and the role that user trust plays within this relationship. By taking into account how different motivation profiles influence this relationship, this study provides organizations with useful insights concerning how they can ensure that their mobile applications offer their customers the appropriate services. When organizations have greater awareness of what kind of motivated consumers they have (i.e. whether they are intrinsically or extrinsically motivated), they are more capable of responding to their customers’ needs. Furthermore, Acquisti, Brandimarte, and Loewenstein (2015) state that privacy is one of the most important issues in the contemporary world, as people now live in an age of information. This concern is also reflected in new legislation that will apply in the European Union as of the 25th of May 2018: Article 35 of the General Data Protection Regulation (GDPR). This new legislation implements the concept of a data protection impact assessment (DPIA) (European Commission, 2017). This concept is a process for constructing and demonstrating compliance for the citizens of the European Union, as it provides them with control over their personal data (European Commission, 2017). This study contributes to
the discussion regarding privacy. Finally, the number of mobile applications being developed is growing increasingly (Keith, 2017). It would be interesting to gain greater understanding of the role played by user trust in mobile applications; this study contributes to the understanding of this phenomenon by investigating if user trust plays a mediating role in the relationship between privacy concerns and self-disclosure. In order to test the relationship between privacy concerns and self-disclosure, the mediating effect of user trust, and the moderating effect of motivation, a quantitative study in the form of a survey is conducted.
This research is structured as follows: First, the relevant literature concerning self-disclosure, motivation, privacy concerns, and user trust is reviewed. Second, the research design and the method are discussed. Thereafter, the results are analysed. Finally, a discussion of the results is presented, the managerial and theoretical implications are discussed and suggestions for future research are made.
2. Literature review
To provide insight into previous work that has addressed the relationship between privacy concerns and self-disclosure in mobile applications, user trust, and motivation profiles, this chapter provides an overview of the relevant literature on these topics. First, the concept of self-disclosure is elaborated. Second, different motivation profiles are explained. Subsequently, a definition of privacy is provided, and the definition of privacy concerns is discussed. Finally, the concept of user trust is explained.
2.1 Self-disclosure
Providing personal information in mobile applications can be considered to be a form of self-disclosure. Self-disclosure is ‘the process of making the self known to others’ (Jourard & Lasakow, 1958, p. 91), referring to people sharing various types of information about themselves with single or multiple others. This can range from factual to personal, private, or intimate information about themselves (Attrill & Jalil, 2011). Self-disclosure has two characteristics, breadth and depth. Breadth refers to the extent to which information is revealed, while depth refers to the extent of intimacy. For breadth, frequency and duration are important factors. For depth, a person’s intent, honesty, and accuracy are important factors. This process is often related to an individual’s well-being, the creation and maintenance of relationships, and the creation of intimacy (Taddei & Contena, 2013). This means that self-disclosure can vary in extent, the level of intimacy or depth and value achieved, reliability or precision, tone, and the level of realization (Taddicken, 2014).
In this digital age, individuals are increasingly willing to disclose personal information online (Attrill & Jalil, 2011). As an example, Special and Li-Barber (2012) investigated the reasons why undergraduate students use Facebook, taking privacy and self-disclosure into account. They found that their participants had three relevant motives for using Facebook:
relationship maintenance, passing time, and entertainment. However, it is often unclear to a self-discloser with whom information will be shared and how many individuals will have access to the shared information. In addition, the audience that a self-discloser intends to reach may differ from that which is actually reached. In this manner, third parties can pass on personal information and use it in other contexts, such as advertising (Taddicken, 2014).
Knowledge can be consciously or unconsciously shared between an individual and an organization. One reason for doing this is for an organization to provide authentication services. This occurs when an organization wishes to improve its service by recognizing an individual in the future and providing him or her with personalized offers. Another reason for the sharing of knowledge between an individual and an organization is marketing purposes. This refers to the process where the individual is required to register to gain access to a website or to join an online community. Furthermore, organizations may also request users’ personal information in order to conduct academic or commercial research (Joinson et al., 2008). Self-disclosure is thus necessary for mobile applications to function effectively (Nickel & Schaumburg, 2004).
Due to the emergence of new technologies, particularly the Internet, people are increasingly required to disclose personal information online (Joinson, Reips, Buchanan, & Schofield, 2010). This can for example be necessary when an individual wishes to make online purchases, access a special service, or join an online community. Taddicken (2014) claims that it is important to distinguish between different forms of personal information. Based on the findings of his study, many users are willing to disclose factual information, such as their names and profession. Fewer users, however, are willing to disclose sensitive information, like experiences, thoughts and/or feelings. Furthermore, Taddicken (2014) found that the higher people rated the importance of the social web (i.e. Internet applications such as social networking sites, blogs, platforms for sharing pictures and videos, etc.) and the more
they are focused on specific social web applications, the more information that they are willing to disclose.
Many people have the impression that online self-disclosure reduces vulnerability when compared to person-to-person interactions, as it affords the discloser anonymity. This phenomenon is also referred to as the ‘strangers on the Internet’ phenomenon (Joinson et al., 2008). Anonymity can also relate to increased self-disclosure via changes in self-awareness processes, uncertainty-seeking behaviour, and the online disinhibition effect (Joinson et al., 2008). This effect means that people feel less restrained when it comes to disclosing personal information in cyberspace then they would face-to-face (Suler, 2004).
2.2 Motivation
Motivation implies that an individual is “moved to do something” (Ryan & Deci, 2000, p. 54). An individual who has no impulse or does not feel inspired to act in a certain manner is characterized as being unmotivated, whereas an individual who feels inspired or who is energized to achieve his or her goal is characterized as being motivated (Ryan & Deci, 2000). People can vary both in the degree of how much motivation they have and in the orientation of such motivation. The orientation of motivation indicates what type of motivation people have, the “why” behind actions taken (Ryan & Deci, 2000).
In order to understand motivational patterns, social-cognitive theories have been developed. One of these theories is the self-determination theory, which attempts to provide greater insight into motivational processes. This theory claims that people are innately and proactively motivated to understand their social environments (Mallett, Kawabata, Newcombe, Otero-Forero, & Jackson, 2007).
Within the field of sports, Pelletier et al. (1995) have translated and validated the
extrinsic motivation, which have also been identified by Ryan & Deci (2000). Furthermore, Ryan and Deci (2000) also mention a third type, amotivation. These three types are described in the following chapters. First, intrinsic motivation is elaborated upon first, followed by discussions of extrinsic motivation and amotivation.
2.2.1 Intrinsic Motivation
Intrinsic motivation is “the inherent tendency to seek out novelty and challenges, to extend and exercise one's capacities, to explore, and to learn” (Ryan & Deci, 2000, p. 70); it refers to engaging in an activity to receive inherent satisfaction, as opposed to doing so for a specific consequence. A person engages in the activity for the fun or challenge that comes with it, not because of external pressures or rewards (Ryan & Deci, 2000). From birth, humans are curious, active, and whimsical creatures who wish to learn and investigate. They do not require external incentives to do so, as this is a natural motivational tendency (Ryan & Deci, 2000). This natural tendency acts on the inherent interests that a person wants to gain knowledge and skills; it is not limited to childhood, but is rather consistent throughout all stages of a person’s life (Ryan & Deci, 2000).
As mentioned previously, the self-determination theory attempts to explain motivational processes. A sub-theory within self-determination theory is cognitive evaluation theory. This theory attempts to identify the specific factors that account for fluctuation in intrinsic motivation. Cognitive evaluation theory takes into consideration the social and environmental factors that have an effect on intrinsic motivation (Deci & Ryan, 1985). The theory holds that social factors can enhance the intrinsic motivation for engaging in an action, which implies that people must experience their behaviour as being self-determined (Ryan & Deci, 2000).
engage in an activity to without the promise of external rewards (Ntoumanis, 2001). This form of motivation is generally present in athletes who wish to learn more about their sports or who really like their sport and wish to become better at what they are doing (Mageau & Vallerand, 2003). For example, a person may exercise for the amusement and satisfaction experienced while learning, examining, or attempting to understand something; alternatively, an individual may engage in an activity due to the amusement and satisfaction experienced when developing or achieving something or to experience the stimulating sensations, produced as a result from doing so (Pelletier et al., 1995).
2.2.2 Extrinsic motivation
After early childhood, many people’s intrinsic motivation declines due to social demands and having to occupy roles in which they must take responsibility for tasks that do not motivate intrinsically (Ryan & Deci, 2000). At this point, people are likely to become increasingly extrinsically motivated. Extrinsic motivation refers to engaging in an activity in order to obtain a specific outcome, in order to receive an instrumental value (Ryan & Deci, 2000). For example, a person who is extrinsically motivated may engage in an activity in order to receive a reward or to avoid being punished (Ryan & Deci, 2000). This implies that such an individual will only be motivated by external contingencies and will not engage in a task for its own sake (Ntoumanis, 2001).
Self-determination theory proposes that extrinsic motivation can differ in the extent to which it is autonomous (Deci & Ryan, 1985). This means there is intentional behaviour, but the type of extrinsic motivation varies (Ryan & Deci, 2000). The organismic integration theory was introduced within self-determination theory. This theory was developed in order to provide a framework for understanding the different forms of extrinsic motivation and their contextual factors (Deci & Ryan, 1985).
Within sports, extrinsically motivated individuals may engage in exercise in order to avoid negative results or to receive rewards, as opposed to doing so for fun (Pelletier et al., 1995). Furthermore, people may exercise because they can have the impression that playing sports is part of their identities and is necessary to their growth and development as a person (Pelletier et al., 1995). Furthermore, athletes may believe that a training sessions should be attended not only because they value the training itself highly but also because they believe that full commitment and hard work are necessary to achieve success in other aspects of life (Mallett et al., 2007).
2.2.3 Amotivation
Amotivation refers to a state in which an individual has no intention to act. This means that the behaviour of amotivated people lacks intentionality and personal reasons for engaging in an activity (Ryan & Deci, 2000). An amotivated individual feels incompetent and out of control, and he or she does not feel intrinsically or extrinsically motivated (Pelletier et al., 1995). As this research focuses on individuals who do feel motivated, this type of motivation is not discussed further.
2.3 Privacy concerns 2.3.1 Privacy
As mentioned previously, Acquisti, Brandimarte, and Loewenstein (2015) state that privacy has become increasingly important, given that people now live in an age of information. Increasing amounts of information concerning people’s beliefs, interests, preferences, and so on are becoming available to commercial entities. Both firms and individuals can benefit from increased access to information. However, this trend can also threatens personal autonomy. But what is privacy exactly?
There are many different definitions of privacy. Van Lieshout, Friedewald, Wright and Gutwirth (2013) claim that privacy can be viewed as a social value, a public good, and an individual value. Definitions are considered to be value-based or cognate-based. Value-based definitions define privacy as either a good that has economic value or a human right that must be protected (Taylor, Ferguson, & Ellen, 2015). Cognate-based definitions describe privacy as an individual’s behaviour or a tendency to behave in a certain manner. This can be a state of mind or an affirmation of control (Taylor et al., 2015). Thus, privacy is both a trait that is a relatively constant and enduring tendency on the part of an individual and a cognate state that depends on the situation.
Two important aspects of privacy are trust and risk. Trust is the willingness of an individual to rely on the honesty, ability, and kindness of another party. This is a very important aspect for a customer when doing business in the online environment (Bansal, Zahedi, & Gefen, 2016). Risk is the possibility that a negative outcome will occur. Negative outcomes can influence an individual emotionally, materially, and physically (Norberg et al., 2007).
2.3.2 Privacy concerns
Since the introduction of the iPhone in 2007, the functionality of mobile applications has increased significantly. A mobile application is a piece of software that operates on a mobile device and can be downloaded from distribution platforms such as the Apple App Store. These applications are often created to provide consumers with entertainment, to increase productivity, and so on (Keith, Thompson, Hale, Lowry, & Greer, 2013). Moreover, mobile applications assist people to access the Internet on their mobile phones or other portable devices by connecting them to Internet services that are usually accessed using their notebook computers (Wang et al., 2013).
Nowadays, there are an increasing amounts of personal content online, which has led to the emergence of privacy concerns too (Ahern, Eckles, Good, King, Naaman, & Nair, 2007). Privacy concerns refer to the assumptions concerning the risks and possible negative consequences that are associated with the sharing of information (Baruh, Secinti, & Cemalcilar, 2017). These concerns are focused on three areas in particular: concerns about which personal information is collected, potential loss of control over personal information, and awareness of privacy practices. As a result of these concerns, customers make more strategic decisions regarding whether or not they will share their data by actively using disclosure management strategies to decide whether or not to share personal information (Zukowski & Brown, 2007).
Shibchurn and Yan (2015) claim that the choice to disclose depends on the information owner’s privacy concerns. People disclose information more readily when they have the impression that the benefits offered by doing so outweigh the risks (Shibchurn & Yan, 2015). Another reason why people may choose to disclose information more readily is that they may not even be aware of the kind of information that they are sharing. An example of this is the fact that users frequently do not read the privacy policies of service providers, instead accepting their terms of service without being aware of how their information will be treated and how they could exercise control over the ways in which information concerning them is collected, gathered, or shared (Steinfeld, 2016).
The research conducted by Keith et al. (2013) found that, when there is an increase in the perceived privacy risk of a new mobile application, the intention of an individual to disclose information over such an application decreases. Therefore, the following hypothesis is formulated:
H1. Privacy concerns influence the degree of self-disclosure of mobile application users negatively.
Cho (2007) found proof that the functional theory of self-disclosure indicates that people engage in self-disclosure for strategic reasons. Furthermore, Bazarova and Choi (2014) also claim that people pursue strategic goals while disclosing. As an individual who is motivated is characterized as being energized to achieve his or her goals (Ryan & Deci, 2000), it could be argued that pursuing strategic goals can be seen as motivation. The motives behind engaging in disclosure can vary between people: For example, Special and Li-Barber (2012) found that users disclosed information on Facebook for different reasons, including relationship maintenance, passing time, and entertainment. Furthermore, Cho (2007) found that groups who had different reasons for engaging in online chat demonstrated different degrees of self-disclosure: For example, this study found that individuals who used online chat services to make new friends were less inclined to disclose personal information than those who used online chat to share information in general.
Adopting an intrinsic-extrinsic perspective, Shibchurn and Yan (2015) conducted a study in which they examined the disclosure intentions of online social network users. The authors found that the reasons why social network users disclose information for non-monetary purposes include socialization, entertainment, and seeking information (Shibchurn & Yan, 2015). On social networks, disclosing information is an entirely voluntary activity, as individuals are not obliged to disclose information to others in a certain manner. This indicates that, when users are willing to disclose information, they likely do so because they are being motivated by the intrinsic benefits that derive from this activity, as it is intrinsically rewarding (Shibchurn & Yan, 2015; Tamir & Mitchell, 2012). When people are intrinsically motivated, they may feel that their basic need for autonomy is satisfied (Deci & Ryan, 2000).
Offering a reward can prompt extrinsically motivated people to increase their degree of disclosure over social networks. Consumers, for example, are willing to surrender a degree of privacy when they receive tangible payments such as discount coupons (Shibchurn & Yan, 2015). In contrast, Andrade, Kaltcheva, and Weitz (2002) found that offering a reward can enhance privacy concerns, which makes consumers less willing to disclose personal information.
When consumers consider providing personal information to organizations, they behave as if they are performing a “cost-benefit” analysis, also referred to as a a “privacy calculus” (Laufer & Wolfe, 1977; Culnan, 1995). This concept refers to the phenomenon that people tend to be more willing to accept a loss of privacy when an acceptable level of risk accompanies the benefits offered by doing so (Culnan & Bies, 2003). As intrinsically motivated individuals may feel that their basic need for autonomy is satisfied, they will likely be more willing to disclose information than extrinsically motivated people, who may feel that they are surrendering control (Deci & Ryan, 2000).
When people are rewarded or threatened, they tend to feel controlled and dominated, which diminishes their degree of satisfaction in terms of autonomy (Deci & Ryan, 2000). Kasser and Ryan (1996) found that individuals who are more focussed on extrinsic motivation demonstrated low levels of well-being, whereas those who placed a strong emphasis on intrinsic motivation showed high levels of well-being. Malhotra et al. (2004) claim that privacy concerns consist of three dimensions: collection, control and awareness. As discussed by Deci and Ryan (2000), individuals who are extrinsically motivated tend to feel controlled and dominated. Therefore, it could be argued that extrinsically motivated individuals with low privacy concerns tend to feel controlled and unsatisfied, which leads to the conclusion that, when extrinsically motivated people have low privacy concerns, their willingness to disclose will decrease. As intrinsically motivated people feel that their basic need for autonomy is
satisfied, they are more likely to be willing to disclose information than extrinsically motivated people, who feel like they are surrendering control (Deci & Ryan, 2000). Therefore, the following hypothesis is formulated:
H2a. If a mobile application user has a low degree of privacy concerns, those users who are intrinsically motivated will disclose more information than those users who are extrinsically motivated.
Individuals with higher levels of privacy concerns are more likely to remove their data from databases and engage in protective behaviours (Taylor et al., 2015). As a result, it could be argued that the benefits of disclosure may not exceed the risks that are considered with high privacy concerns. An individual who is intrinsically motivated discloses because doing so is intrinsically rewarding (Shibchurn & Yan, 2015). However, someone who is intrinsically motivated and has a high level of privacy concerns may not feel intrinsically rewarded when disclosing information and will therefore be less willing to do so. The same could be said for extrinsically motivated people, since such individuals who have privacy concerns tend to feel controlled and therefore unsatisfied, which may lead to a decreasing willingness to disclose (Malhotra et al., 2004; Deci & Ryan, 2000). Therefore, the following hypothesis is formulated:
H2b. If a mobile application user has a high degree of privacy concerns there will be less difference in the extent of self-disclosure and the various motivation profiles.
2.4 User Trust
Trust is the willingness of an individual to rely on the honesty, ability, and kindness of another party in situations that entail risk (Bansal et al., 2016). There are three characteristics of trust: First, there are two parties involved, the trustor and the trustee. The trustor is the user, the individual who takes a risk. The trustee is the party who provides the trustor with a product. These two parties are mutually dependent on each other to gain benefits. Second, uncertainty and risk are involved. This means that there is no guarantee that the trustee will meet the expectations of the trustor. Finally, the trustor must have faith in the honesty, ability, and kindness of the trustee (Siau & Shen, 2003).
One area that experiences difficulties in terms of gaining trust is mobile commerce. Therefore, Siau and Shen (2003) developed a framework for building trust in mobile commerce. They claim that building trust in mobile commerce is an on-going process, which ranges from creating trust to continuous trust development thereof, with mobile technology and vendors being crucial elements. In order to improve trust in mobile technology, developers should ensure that they create platforms that emphasize both usability and security in their systems. In addition, trust in mobile vendors also needs to be developed. The authors propose five methods for developing such trust: First, customer familiarity with a platform should be promoted, because people tend to trust what is familiar. Second, vendor reputation should be developed, because this reputation reflects a company’s history and can imply certainty. Third, high-quality information should be provided to the platform users. Fourth, recognition and certification of third parties should be encouraged. Finally, attractive awards should be offered.
Beyond trust in mobile commerce, there is also a phenomenon called user trust. Yan et al. (2013) refer to user trust in a mobile application as the belief on a user’s part that an application is capable of performing the task expected of it. While using a mobile application,
the users make decisions about whether or not they trust it. There are several factors that are related to the trustworthiness of mobile applications, namely dependability, security, usability, and how popular the mobile application in question is (Yan et al., 2013). Among the factors that affect usage, trust is important because it leads to trust behaviour. This means that, after a trusting subject and a trusting object have interacted for some time, credible information can be collected. With regard to mobile application usage, trust behaviour is demonstrated when a mobile user comes to rely on an application in his/her daily life and when he or she evaluates its trustworthiness by observing the consequences of its use (Yan et al., 2013).
Olson, Grudin, and Horvitz (2005) found that “people’s willingness to share depends on who they are sharing the information with” (p. 1987). As a result, within the relationship between a discloser and the receiver, the trust that one party has in the other will determine whether or not he or she discloses information. Within the online environment, privacy is often considered to be a contributor to trust, instead of an independent effect on online behaviour (Joinson et al., 2010). For example, Malhotra, Kim, and Agarwal (2004) found that trust mediates the effect of privacy concerns on behavioural intentions. Furthermore, Metzger (2004) found that, with regard to e-commerce, the relationship between privacy concerns and self-disclosure is mediated by trust. Joinson et al. (2010) found that, for situational aspects of privacy, the relationship between privacy and disclosure is mediated by trust. Furthermore, Morosan and DeFranco (2015) examined the disclosure of personal information on hotel apps. They found that, in the relationship of trust with the willingness of consumers to disclose information, trust in the mobile application was a strong predictor variable.
Stone, Gueutal, Gardner, and McClure (1983) found a negative relationship between privacy concerns and degree of control over personal information, which is an important factor of trust. When an individual feels that he or she has no control over personal
information or has the impression that his or her personal information is exposed, privacy concerns arise (Dinev & Hart, 2004). With regard to the Internet, Sheehan and Hoy (1999) found that, when privacy concerns increase, people register on websites less frequently or provide websites with insufficient amounts of information. Lin, Amini, Hong, Sadeh, Lindqvist and Zhang (2012) state that, if a person’s impression of a mobile application aligns with what that app actually does, there would be fewer privacy problems, because people could make better trust decisions about the mobile application (Lin et al., 2012). Therefore, it could be argued that privacy concerns are an important factor in user trust.
Furthermore, trust is a central factor that influences disclosure (Metzger, 2004). Trust plays a crucial factor in the privacy calculus. As explained previously, the privacy calculus refers to the tendency that people are more willing to accept the loss of privacy when an acceptable level of risk accompanies the benefits that doing so provides (Culnan & Bies, 2003). Trust is crucial in this process, as it reduces the perceived risks that are involved in revealing personal information (Metzger, 2004). This relationship also exists in the electronic environment (Culnan, 1999). This indicates that trust is crucial to disclosure in both interpersonal and online relationships. In an environment in which social cues are diminished due to computer-mediated communication, gaining trust may be more problematic, yet more important, than in an interpersonal context (Boyd, 2003). Because trust is an important factor in terms of self-disclosure, it could be argued that user trust also plays a large role. As trust plays a significant role in the relationship between privacy concerns and self-disclosure, it can be argued that user trust also has an indirect effect on the relationship between privacy concerns and self-disclosure on a mobile application. Therefore, the following hypothesis is formulated:
H3. User trust mediates the relationship between privacy concerns and self-disclosure in a mobile application such that privacy concerns have a negative impact on user trust and user trust has a positive impact on self-disclosure.
2.5 Conceptual Model
Based on the hypotheses stated above, the following conceptual model is devised (Figure 1):
3. Method
In this chapter, the empirical section of this study is presented. First, the procedure employed to gather data for this study is explained; thereafter, the measurements used for the survey and the sample are described.
3.1 Procedure
In order to answer the research question and to test the created conceptual model, a quantitative study, in the form of a survey, was conducted. The survey has a correlational design, in which two continuous variables were measured: privacy concerns and motivation. This survey was designed online in Qualtrics, and the Statistical Package for the Social Sciences (SPSS) was used to analyse the data.
The survey started with some introducing remarks, in which the participants were thanked for their participation, the purpose of the survey was explained, and their privacy was assured. The respondents were first asked to reveal their gender, age, nationality, educational level, and work situation. Thereafter, the respondents were asked whether or not they used a mobile sport application. If a respondent answered this question with “no”, he or she was immediately sent to the end of the survey.
Thereafter, the respondents were asked question concerning their feelings about self-disclosure. This section took the form of a Likert scale and contained questions such as “When I reveal my feelings about myself in a mobile sport application, I consciously intend to do so”. Next, the respondents were provided with Likert-scale questions concerning motivation, user trust, and privacy concerns. Within the motivation section, a bogus question was asked in order to filter careless responses. This bogus question was: “I have been to every country in the world (Respond with ‘strongly disagree’ for this item)” (Meade & Craig, 2012).
Finally, the respondents were thanked for their participation.
3.2 Measurements
All variables were measured using either a 7-item or 5-item Likert scale. To ensure construct validity, scales from previous studies, with some small adjustments to better fit in the purposes of this study, were used.
The scale items for self-disclosure were measured using 10 items from Wheeless’ (1978) scale and 10 items from The Jourard Sixty-Item Self-Disclosure Questionnaire mentioned by Cozby (1973). Wheeless’ (1978) scale focuses on three different dimensions: depth, honesty, and intent. The authors of the original study reported the following Cronbach’s alpha’s: 0.84 for depth, 0.85 for intent, and 0.87 for honesty. The items of Wheeless’ (1978) items were measured using a 7-items Likert scale (1=strongly agree, 7=strongly disagree). An example of one of the items used in the current study is “When I reveal my feelings about myself in a mobile sport application, I consciously intend to do so”. The items taken from Cozby (1973) were measured using a 7-item Likert scale (1=definitely will, 7=definitely not). The authors of the original study reported a Cronbach’s alpha of 0.68. The Cronbach’s alpha in this present study was 0.902.
The scale items for motivation were measured using 20 items taken from the scale developed by Mallet et al. (2007). The number of these items was reduced from 24 to 20 items by removing those items concerning amotivation, since this topic is not applied in this research. The Cronbach’s alpha reported by the authors was 0.78. The items were measured using a 7-item Likert scale (1=strongly agree, 7=strongly disagree). An example of one of the
items is “Because training hard will improve my performance”. The Cronbach’s alpha in this present study is 0.896.
Scale items for user trust were measured using the scale developed by Yan et al. (2013). The items were adapted and reduced in number to 12 (e.g. “The more times you use a mobile application, the more you trust it”). The items were measured using a 7-item Likert scale (1=strongly agree, 7=strongly disagree). The authors of the original study reported a Cronbach’s alpha of 0.73. For this present study, a Cronbach’s alpha of 0.871 is reported.
The scale items for privacy concerns were measured with the scale developed by Baek & Morimoto (2012. The Cronbach’s alpha reported by the authors was 0.86. The items were adapted and reduced in number to five; they were measured using a 5-item Likert scale (1=strongly agree, 5=strongly disagree). An example of one of the items used in the current study is “I believe that personal information is often misused”. The Cronbach’s alpha of this present study is 0.822.
A detailed list of the constructs can be found in Appendix A. For the full questionnaire, see Appendix D.
3.3 Sample
The sample used in this study consisted of English-speaking users of mobile applications intended for sporting purposes. The study was conducted using a non-probability convenience sample. The respondents were recruited via Facebook, e-mail, and Whatsapp. When approaching the respondents, it was clearly stated that they had to use a mobile sport application in order to participate in the survey. This method was appropriate for this study,
because this research strived to include as many respondents who were in possession of a mobile sport application as possible. By focussing on users of mobile sport applications of all ages, genders, educational levels, and working situations, it was possible to obtain a larger number of respondents. Furthermore, a non-probability convenience sample is an appropriate method for reaching a large group of potential respondents. The survey was closed when Qualtrics listed 556 recorded responses.
4. Results
In this section, the results of the data analysis are presented. First, the respondents are described. Secondly, the data preparation process is described. Finally, the hypotheses are tested.
4.1 Respondents
Of the 556 individuals who opened the survey, 25 respondents did not even start it. Of the 531 who started the survey, 18.8 percent answered “no” to the question “Do you use a mobile application in order to work out?” When a respondent answered this question with a no, he or she was sent to the end of the survey automatically. This led to 431 respondents. Of these 431 respondents, 349 completed the survey, which equals a response rate of 81 percent. Of these 349 respondents, 55 answered the bogus question incorrectly. Therefore, 294 respondents were considered in the data analysis. Of the 294 respondents, 35.7 percent were male and 64.3 percent was female. Ninety-five point two percent of the respondents were Dutch. The age of the respondents varied from 18 to 65, with a mean of 31.0 and a standard deviation of 12.1. One hundred and eighty five respondents were highly educated, with a WO degree (bachelor’s or master’s). Forty six point six percent were employed (full-time or part-time) and 48.6 percent were students. The other 4.7 percent consisted of individuals who were unemployed and searching for employment, individuals who were unemployed and not looking for work, and people who were retired. Detailed demographic information can be found in Appendix B.
4.2 Data preparation
There were no counter-indicative items, so no items had to be recoded. Thereafter, dummy variables were created. There were no respondents who answered “other” to the
gender question; therefore, female was recoded as 1, and male was recoded as 0. Next, a dummy variable was created for working situation, for which being an employee (both full-time and part-full-time) was recoded as 1, and all other items were recoded as 0. Finally, a dummy variable was created for whether or not a respondent was highly educated, i.e. held a WO degree (bachelor’s or master’s), for which WO bachelor’s degrees and WO master’s degrees were recoded as 1, and all other items were recoded as 0.
Afterwards, the outliers for self-disclosure were tested. No outliers were found, as the z-score was between 3 and -3. Therefore, there was no reason to exclude data (Iglewicz & Hoaglin, 1993). Finally, the mean, standard deviation, correlations, and reliabilities were calculated in SPSS. The findings of which can be found in Table 1. With respect to the normality of residuals and homogeneity of the data, the assumptions of the regression are met. The describing figures can be found in Appendix C.
The reliability of all the constructs was tested to examine the internal consistency of the scales. This resulted in the finding that all of the variables were reliable, since they were all over 0.7 (Field, 2013) (self-disclosure: α=.902, motivation: α=.896, user trust: α=.871, and privacy concerns: α=.822). These results can be found in Table 1.
Table 1. Means (M), Standard Deviations (SD), Reliability and Pearson Correlations for all Variables Variables M SD 1 2 3 4 5 6 7 8 1. Self-disclosure 3.76 1.15 (.902) 2. Gender1 .643 .480 -.007 (-) 3. Age2 31.0 12.1 .180** -.262** (-) 4. Employee3 .466 .500 .223** -.158** .648** (-) 5. WO-educated4 .629 .484 .008 .089 -.307** -.144* (-) 6. Privacy concerns (PC) 1.90 .758 -.263** .012 -.133* -.120* -.050 (.822)
7. User trust (UT) 2.93 .837 .253* -.065 .114 .134* .000 -.265** (.871)
8. Motivation (MOT)
2.80 .976 .150* .113 .098 -.012 -.080 -.030 .217* (.896)
Note. N = 294. 1Gender was coded as 1 = female and 0 = male. 2Age was measured in years. 3Employee was recoded as employed (full-time + part-time) = 1 and all other items = 0. 4WO-educated was coded as WO bachelor’s degree + WO master’s degree = 1, and all other items = 0. The Cronbach’s alphas are between brackets.
* p < .05 ** p < .01 (two-tailed)
The results of the correlational analysis (Table 1) show that self-disclosure has the highest standard deviation (SDself-disclosure=1.15, SDpc=.758, SDut=.837, and SDmot=.976). This
means there is more variance for self-disclosure than for the other variables. Privacy concerns have the highest number of “strongly agree” responses, with M=1.90.
There are correlations between the outcome variable and two control variables: age shows a significant relationship with self-disclosure (r=.180, p<.01), as well as employee (r=.223, p<.01). This correlation indicates that being an employee has an impact on how much information a respondent discloses. In addition, there is a correlation between the outcome variable and privacy concerns, user trust, and motivation: privacy concerns shows a significant relationship with self-disclosure (r=-.263, p<.01), as well as user trust (r=.253,
concerns is negative (r=-.263). Furthermore, age and employee show a significant relationship (r=.648, p<.01), which indicates that age influences whether or not a respondent is an employee. Furthermore, motivation and privacy concerns do not show a significant relationship (r=-.030, p>.05). Motivation and user trust show a positive significant relationship (r=.217, p<.05), as do user trust and privacy concerns (r=-.265, p<.01). In order to identify multicollinearity, the variance inflation factor (VIF), which indicates if a predictor has a strong linear relationship with another predictor (Field, 2013), was calculated. Hutcheson and Sofroniou (1999) indicate that a VIF statistic that exceeds 5 causes a problem with multicollinearity. Since VIF<5, there is no multicollinearity (see Table 2).
Table 2. Overview of Collinearity Statistics
Variable Tolerance VIF
Gender .902 1.11 Age .496 2.02 Employee .573 1.75 WO-educated .892 1.12 Privacy concerns .910 1.10 User trust .878 1.14 Motivation .909 1.10
4.3 Testing the hypotheses
In order to test the hypotheses, a multiple regression analysis, the ordinary least squares (OLS) regression analysis, was performed. This regression was tested with the independent variables and the following moderations: motivation and privacy concerns, user trust and privacy concerns, and user trust and motivation. The results of this analysis are presented in Table 3. As a robustness check, the variable motivation was divided into intrinsic motivation and extrinsic motivation.
Considering Table 3, Model 1 contained the control variables (gender, age, being an employee, and being WO-educated). In Model 2, the independent variable (privacy concerns) and some other variables (user trust and motivation) were added next to the control variables, and, in Model 3, possible interactions were added. The first model was significant and explained 4.4 percent of the variance (R2=.057, F=4.33, p<.01). Model 2 was significant and explained 13.9 percent of the variance (R2=.159, F=7.73, p<.01). Finally, Model 3 was significant, as it explained 13.8 percent of the variance (R2=.167, F=5.68, p<.01). Because there was no increase in the variance, it could be argued that the moderations have no explaining power.
Table 3. OLS Regression Analysis with Self-disclosure as Dependent Variable
Model 1 Model 2 Model 3
Estimates B SE β p B SE β p B SE β p Gender .095 .141 .040 .500 .041 .136 .017 .764 .038 .138 .016 .781 Age .009 .008 .094 .242 .003 .007 .030 .699 .004 .007 .040 .601 Employee .406 .173 .177* .019 .353 .164 .154* .033 .371 .165 .162* .026 WO-educated .140 .143 .059 .328 .072 .136 .030 .598 .035 .138 .015 .803 Privacy concerns (pc) -.285 .086 -.189** .001 -.636 .415 -.421 .126
User trust (ut) .202 .079 .148* .011 .272 .368 .199 .461
Motivation (mot) .199 .076 .149** .009 .335 .351 .250 .342 ut x pc .090 .098 .184 .361 pc x mot .031 .097 .081 .751 ut x mot -.067 .079 -.253 .399 R2 .057 .159 .167 Adjusted R2 .044 .139 .138 F-value 4.33** 7.73** 5.68** Note: N = 294. * p < .05 ** p < .01
From Model 2 (as depicted in Table 3), it follows that being an employee, privacy concerns, user trust, and motivation have direct significant effects on self-disclosure (βEmployee=.154, p<.05; βpc=-.189, p<.01; βut=.148, p<.05; and βmot=.149, p<.01). Privacy
concerns have a negative effect on self-disclosure, while motivation and user trust have a positive effect on self-disclosure. When considering Model 3 (in Table 3), all of the moderations were not found to be significant (βut x pc=.184, p>.05; βpc x mot=.081, p>.05; and
βut x mot=-.253, p>.05).
A robustness check was conducted. Again, a multiple regression was used, but motivation was now divided into extrinsic and intrinsic motivation. This resulted in results comparable to those obtained before motivation was divided. The first model was significant and explained 4.4 percent of variance (R2=.057, F=4.33, p<.01). Model 2 was significant and explained 13.9 percent of variance (R2=.163, F=6.93, p<.01). Finally, Model 3 was significant. It explained 13.6 percent of variance (R2=.165, F=5.61, p<.01). Furthermore, the effects are comparable: employee, privacy concerns and user trust are significant. However, there is one difference: extrinsic motivation is significant, but intrinsic motivation is not (βintrinsic=-.050, p>.05; βextrinsic=.191, p<.05). This indicates that extrinsic motivation has the
explanatory power of the significant effect of motivation in total.
Next, the mediation of user trust was tested. According to Baron and Kenny (1986), mediation needs to consider the following steps: First, the independent variable must have a significant effect on the expected mediator variable. Second, the mediator variable must have a significant effect on the dependent variable. Finally, the relationship between the independent variable and the dependent variable, which first was significant, must no longer be significant. In order to test for the mediation while taking the moderation as proposed in the conceptual model into account, Hayes’ process model 5 is used. The results of this analysis led to the creation of Figure 2.
Figure 2. Results of the Mediation Analysis Entered into the Conceptual Model
Note. N = 294. * p < .05 ** p < .01 (two-tailed)
4.3.1 Hypothesis 1
H1. Privacy concerns influence the degree of self-disclosure of mobile application users negatively.
Table 3 (Model 2) indicates that privacy concerns have a significant negative effect on self-disclosure (βpc=-.189, p<.01). Therefore, this hypothesis is accepted.
4.3.2 Hypotheses 2a + 2b
H2a. If a mobile application user has a low degree of privacy concerns, those users who are intrinsically motivated will disclose more information than those users who are extrinsically motivated.
H2b. If a mobile application user has a high degree of privacy concerns there will be less difference in the extent of self-disclosure and the various motivation profiles.
Table 3 (Model 3) indicates that there is no moderation of motivation on the relationship between privacy concerns and self-disclosure (βpc x mot=.081, p>.05). Therefore,
both hypotheses are rejected.
4.3.3 Hypothesis 3
H3. User trust mediates the relationship between privacy concerns and self-disclosure in a mobile application such that privacy concerns have a negative impact on user trust and user trust has a positive impact on self-disclosure.
Figure 2 presents the results of the mediation analysis. The standardized regression coefficient between privacy concerns and user trust is statistically significant (B=-.278, p=.000). Furthermore, the results indicate there is significant effect of user trust on self-disclosure (B=.205, p=.010) and that the effect of privacy concerns on self-self-disclosure is no longer significant (B=-.541, p=.086). The significance of this indirect effect was tested with bootstrapping procedures, as, according to Hayes and Scharkow (2013), this is believed to be the most trustworthy method for indirect effect testing. Since the 95%-confidence interval does not contain zero [-.1256, -0150], user trust can be indicated as a mediator in this relationship (Preacher & Hayes, 2008). Therefore, this hypothesis is accepted (with Effect=-.0572).
4.4 Overview of the hypotheses
Table 4. Summaries of the Hypotheses
Nr Hypothesis Result
1 Privacy concerns influence the degree of
self-disclosure of mobile application users negatively.
Accepted
2a H2a. If a mobile application user has a low degree
of privacy concerns, those users who are
intrinsically motivated will disclose more
information than those users who are extrinsically motivated.
Rejected
2b If a mobile application user has a high degree of
privacy concerns there will be less difference in the extent of self-disclosure and the various motivation profiles.
Rejected
3 H3. User trust mediates the relationship between
privacy concerns and self-disclosure in a mobile application such that privacy concerns have a negative impact on user trust and user trust has a positive impact on self-disclosure.
5. Discussion
5.1 General discussion
The first hypothesis was accepted, indicating that privacy concerns influence the degree of self-disclosure in a mobile application negatively. Shibchurn and Yan (2015) claim that an individual’s choice to disclose depends on his or her privacy concerns. People disclose information more readily when they have the impression that the benefits offered by doing so outweigh the risks. When there is an increase in the perceived privacy risk associated with disclosing on a mobile application, an individual’s intention to disclose information on that application will decrease (Keith et al., 2013). This decrease in the tendency to disclose information leads to a negative relationship between privacy concerns and self-disclosure. This is also in accordance with the findings of Dinev and Hart (2006) and Wirtz, Lwin and, Williams (2007), who claimed that information disclosure is influenced by privacy concerns. Furthermore, the moderating role of motivation on the relationship between privacy concerns and self-disclosure could not be proven. Therefore, Hypotheses 2a and 2b are rejected. A possible explanation for this is may be that this effect is not applicable to mobile sport applications. Since the survey conducted for this research focussed on sport motivation profiles and self-disclosure in mobile sport applications, it may be possible that other results would have been obtained if it the survey was not focussed on sports but instead on other areas (e.g. the intrinsic or extrinsic motivation of students or employees).
Next, it is remarkable that a moderating effect of motivation on the relationship of privacy concerns and self-disclosure could not be proven, but a direct effect of motivation on self-disclosure could. The results of this study indicated a direct effect of extrinsic motivation on self-disclosure. A possible reason for this could be that offering a reward may encourage people to increase their degree of self-disclosure (Shibchurn & Yan, 2015). Some studies, for
example, have shown that consumers are willing to surrender a certain degree of privacy when they receive one-time rewards in the form of tangible payments, e.g. discount coupons and/or bonus points (Zhang, Wang, & Chen, 2000). Taking this into account when considering mobile sport applications, an individual may be more willing to disclose information as a result of being extrinsically motivated in the form of external regulation. External regulation means that a behaviour is supervised by external sources and can lead to punishment or rewards (Ntoumanis, 2001). This implies that an individual may wish to avoid negative results or desires to receive a reward (Pelletier et al., 1995). When an organization offers, for example, a reward, it may be possible that this would have a direct effect on self-disclosure and that privacy concerns would not be considered in this process or would be taken for granted by the consumers.
Finally, the third hypothesis was accepted. This indicates that user trust mediates the relationship between privacy concerns and self-disclosure in mobile applications. As was previously discussed by Malhotra et al. (2004), Metzger (2004), and Joinson et al. (2010), trust mediates the relationship between privacy and self-disclosure. Therefore, when taking into account the relationship between privacy concerns and self-disclosure in a mobile application, it can be concluded that user trust plays a mediating role in this relationship.
5.2 Theoretical implications
Acquisti, Brandimarte, and Loewenstein (2015) state that privacy is one of the most important contemporary issues, since we live in an age of information. The findings of this study provide important insights that may contribute to the existing literature on privacy concerns, self-disclosure in mobile applications, user trust, and motivation. Zukowski and Brown (2007) previously suggested that customers make strategic decisions considering whether or not to share their data based on privacy concerns. In addition, Shibchurn and Yan
(2015) claimed that an individual’s choice to disclose information depends on his or her privacy concerns, and also Keith et al. (2013) expected a negative relationship between privacy concerns and self-disclosure. This study contributes to the literature by identifying a negative relationship between privacy concerns and self-disclosure in a mobile application.
Although the mediating role of trust in the relationship between self-disclosure and privacy concerns in e-commerce has been previously discussed in the literature (Malhotra et al., 2004; Metzger, 2004; Joinson et al., 2010), few studies have considered the mediating role of user trust in the relationship between self-disclosure in mobile applications and privacy concerns. The confirmation of the fact that user trust plays a mediating role in this relationship represents a contribution to the existing literature, since this research ascertained that trust not only plays an important role in the context of e-commerce but also in the world of mobile applications.
Finally, this study contributes on a more psychological level by exploring how motivation can influence self-disclosure in a mobile application when it is classified into two categories, namely intrinsic and extrinsic motivation. A moderating effect of motivation was not proven in this study, but a direct effect of extrinsic motivation on self-disclosure in mobile applications was.
5.3 Managerial implications
Beyond its theoretical implications, this study also has managerial implications. This research indicates the roles that privacy concerns, user trust, and motivation play with regard to self-disclosure in mobile applications. In the modern world, an increasing number of mobile applications are being produced, and this number is expected to grow (Keith, 2017). According to Nah, Siau, and Sheng (2005), mobile applications provide organizations with a many opportunities to handle business more effectively. By providing organizations with
greater insight into how customers feel about disclosing information in a mobile application, they can better respond to their customer’s needs. This is particularly important for the parties that wish to make a profits via mobile applications (Wang et al., 2013). Therefore, it is important for organizations to understand that user trust has a mediating effect on privacy concerns and self-disclosure in mobile applications. For example, Runtastic is a mobile application that supports workout activities. In order to use this mobile application to its full potential, a user must provide the mobile application with information such as his or her weight and allow the mobile application to track his or her location. This location tracking could clash with an individual’s privacy concerns; therefore, it is important for an organization to realize that it could reduce the risk of consumers not disclosing information in its mobile application (and thus not being completely satisfied with it) by ensuring that the application scores highly in terms of user trust.
Another managerial implication is that it could not be proven within this study that motivation plays a moderating role in the relationship between privacy concerns and self-disclosure. The results of this study indicated, however, that extrinsic motivation has a direct effect on self-disclosure, while intrinsic motivation does not. An organization should be aware of this distinction, particularly when considering its marketing. If a company wishes to encourage users to disclose information on its mobile application, it should invest in extrinsic motivation (e.g. by offering rewards).
5.4 Limitations
Two hypotheses regarding the moderation effect of motivation were rejected. The methodology employed may be a reason for this, but this may also indicate opportunities for future research (which are discussed in Chapter 5.5). First of all, in order to gather a large amount of data in a short timeframe, this study was conducted using a non-probability
convenience sample. A disadvantage of the use of a non-probability convenience sample is that the generalizability and the representativeness of the results can be questioned. The use of a non-probability convenience sampling resulted in an unequally distributed sample, since 58.5 percent of them were 18-25 years old, 95.2 percent were Dutch, and 64.3 percent were female. It can be argued that a more equally distributed sample would have provided other results. For example, it may be the case that older respondents would be more aware of privacy concerns and therefore would have responded to the questionnaire differently.
Second, with regard to the motivation variable, the possibility of error and biases should be considered, because it can be argued that the respondents may have answered these questions in a socially desirable manner. In order to avoid this, the need for honesty was emphasized in the introductory section of the survey.
Furthermore, the survey was quite lengthy, which may explain the high dropout percentage (19 percent) and why people answered the bogus question incorrectly (55 out of 349 respondents).
Finally, the respondents were approached via Facebook, among other channels. It could be argued that these respondents already had lower levels of privacy concerns, because they already disclosed personal information online (Special & Li-Barber, 2012).
5.5 Future research
As stated in the previous section (5.4 Limitations), with regard to the motivation variable, the possibility of error and biases should be taken into consideration. Since this study focussed on motivation in sports, it could be interesting for future research to measure motivation and self-disclosure in a dimension other than sports.
Second, this study focused on user trust. It could be interesting to add another variable, namely, the trust that an individual has in an organization. Morosan and DeFranco (2015)
