Societal Trust and Risk in the Privacy Paradox; a Facebook Case Study

Societal Trust and Risk in the Privacy Paradox; a Facebook Case

Study

Faculty of Economics and Business

Msc. Marketing

Management track

13-01-2020

Completion date: 13-01-2020 Supervisor: Lundahl, O.

2 Abstract

In the rapidly growing world of social media and online activity, this study aims to explain the theory of the privacy paradox. With the privacy paradox studied on the individual level

already, this research explores what happens to the paradox on the societal level in times of a social media crisis. The context, the Cambridge Analytica scandal in 2018, is ideal in finding out what happens to the two key concepts of the privacy paradox, trust and risk, on the societal level. This study did so through a longitudinal frame analysis of three major British newspapers, the Guardian, the Telegraph, and the Daily Mail. Both the quantitative as the qualitative findings showed that the societal risk perceptions increased during the scandal and stayed high after the scandal. The societal level of trust in Facebook decreased during the scandal and remained low after the scandal, implying that users did not trust Facebook anymore. However, despite the low trust and high risk perception, the public continued using Facebook. Users accepted the fact that Facebook cannot be trusted anymore, and the risk became normalized.

3 TABLE OF CONTENTS 1. Introduction 4 2. Literature Review 7 2.1 Social Media 7 2.2 Privacy Paradox 8

2.3 Trust and Risk on the Societal Level 10

3. Methodology 14 4. Findings 18 4.1 Quantitative Findings 18 4.2 Qualitative Findings 20 5. Discussion 30 6. Conclusion 36

7. Limitations & Future Research 37

4 1. INTRODUCTION.

In today’s world, social media is all around us, mostly through smartphones (Chugh & Ruhi, 2018), and new technologies, apps, and websites are born daily. Social media has become an essential part of our lives, with 3,84 billion social media users worldwide, according to the We Are Social Global Digital Report 2019. What follows is the fact that social media have much power (Lund, Cohen & Scarles, 2018) and have grown to be critical channels for marketing (Kaplan & Haenlein, 2010). Of all the social media, Facebook is the largest social media platform with 1,63 billion daily users. Their size, combined with the power social media has these days, makes Facebook one of the most influential companies in the last two decades. With great power comes great responsibility, but in the most prominent online scandal ever, Facebook showed the world it was not ready, or not willing, to have that responsibility.

At the beginning of 2018, Facebook was a key player in an immense crisis in customer data privacy that shocked the online world. Cambridge Analytica, a British consulting firm, was able to collect data from almost 87 million Facebook users, without the approval of these users. The company obtained access to 320,000 user-profiles and their friends’ data, which was done using the “thisisyourdigitallife” app. The app was developed by psychologist Alexandr Kogan of Cambridge University, UK, and he sold it to the company (Schneble, Elger & Shaw, 2018). The impact of the affair was huge, and people became very concerned about their privacy protection. According to Woollacott (2019), the United Kingdom proposes new laws against social media companies like Facebook. This new set of regulations would include fines, site blocks, and prosecution of senior management if social media companies fail to protect their users.

5 privacy but, when their behavior is investigated, their actions show that privacy is not the main priority. The privacy paradox on the individual level is studied mainly in the

e-commerce and social network sector. It shows that while many people show interest in their privacy, in theory, and stay favorable to actual behavior to protect their privacy, people rarely actual behave to protect it (Joinson, Reips, Buchanan & Schofield, 2010). In the context of social network sites, more important for this study, the same pattern is observed. Young & Quan- Haase (2013) explain that people tend to utilize different privacy protection strategies on Facebook, such as limiting wall post access and sending private messages instead of open content. However, these strategies show no worry for data collection in the background by third parties. Instead of that the privacy concerns lead to the secured provision of personal information in social networks, the reverse effect is noticed with many users providing this information without any doubt (Manier & O’Brien, 2010). Despite the many studies on the privacy paradox on the individual level, Barth & De Jong (2017) show in their analysis of 35 studies on the privacy paradox, that there remains a lot of research to be done on this topic, even on the individual level.

With clear and decisive reasoning for the privacy paradox on the individual level still missing, research on this topic remains of great importance. However, the privacy paradox on the societal level is undiscovered. Whereas the Cambridge Analytica scandal affected millions of people, it affected societies as a whole, creating a need for an understanding of the privacy paradox on the societal level. Therefore, this study explores the privacy paradox on the societal level using the framework from the study of Norberg et al. (2007). In their research, Norberg et al. (2007) identify two key antecedents of the privacy paradox; trust and risk. This study will use these two critical antecedents of privacy paradox as key concepts to shift the paradox from the individual level to the societal level. On the personal level, these

antecedents will influence an individual in whether he or she says they value privacy but also acts upon it. As Norberg et al. (2007) hypothesize, the perception of risk will influence the behavioral intention, whereas the perception of trust will decide the actual behavior in

securing privacy. To shift the paradox from an individual level to a societal level, the concepts of it, trust and risk, need that shift as well. To do so, a deeper look into both the key concepts is needed and of great importance to make the shift.

6 Gordon, 2002; Hoffman, Novak & Peralta, 1999; Mayer, Davis & Schoorman, 1995), some studies looked at the societal importance of trust already. For instance, Bachmann & Inkpen (2011) state that for a trust crisis to occur, trust on a micro-level, so the individual level, is not of great importance. The macro-level level of trust, so the trust in regulations and laws, is the initiator of the crisis. In their study, they call this institutional-based trust. Fukuyama (1995) explains this institutional-based trust as societal trust and states that this kind of trust is key for a developed society, since it is practically the foundation of social capital and thus the foundation of a society as a whole.

The other key antecedent of the privacy paradox and important for data privacy concerns is risk. Perceived risk changes someone’s attitude about privacy concerns and affects one's intention to value privacy, not their actual actions (Norberg et al., 2007). Most studies on risk focused on the individual level of risk, not much on the societal level. Shifting the risk to a societal level has been studied already, following the preceding literature from Slovic, Fischhoff & Lichtenstein (1984). Important when studying risk on a societal level is the process of risk normalization. Barnett and Breakwell (2003) studied the risk amplification around the 1995 oral contraceptive pill scare and showed that risk could become normalized when being more exposed to the hazard. As well as trust, the perception of risk on a societal level is crucial and key for maintaining a developed society, according to Fukuyama (1995).

These studies show the importance of understanding risk and trust, and the balance between the two, on a societal level, especially in times of a crisis. The aim of this study is to change the scope of privacy concerns and the privacy paradox from an individual level to a societal and to see what happens to the concerns and the paradox in times of a crisis. Therefore, the research question of this study is:

‘How does the balance between trust and risk, on the societal level, change before-, during- and after a crisis?’

This study contributes to the literature on the privacy paradox by shifting the focus from an individual level to the societal level. This study uses the concepts ´trust´ and ´risk’, as

Norberg et al. (2007) stated to be the key concepts of the privacy paradox, to make this switch and therefore mainly contributes to the literature of Norberg et al. (2007). Both concepts are studied on the societal level and create a view on the privacy paradox on the societal

7 identify and explain, especially for trust and risk in times of a crisis.

To make this shift, and thus the contribution, this research studies the changes of both risk and trust on a societal level more extensively. Doing that, this study contributes to the literature on both concepts as well, by showing the development of the concepts on a societal level, during an online crisis. For trust, this study contributes to the literature from Bachmann & Inkpen (2011), by showing whether their tools for creating institutional-based trust worked for Facebook and how they have influence. Also, this study showed how different institutional-based trust could work for a different societal scandal, extending the scope of the study from Bachmann & Inkpen (2011). As for risk, this study contributes to the study by Barnett & Breakwell (2003) by showing how risk became normalized, as they explain in their research, in a different context, and using only one media source instead of different media sources.

The context of this research, social media, is carefully chosen and contributes to the literature on crisis contexts, e.g., the study by Humphreys & Thompson (2014). With the power social media have, the size of its market, and the lacking legislation for social media companies, the Cambridge Analytica scandal might not be the only social media scandal, as other scandals can happen in the future. It makes the Cambridge Analytica scandal an interesting but also topical context and it is used in this study as a crisis.

This study is based on a longitudinal frame analysis (2016-2018) of two UK newspapers, the Guardian and the Telegraph, and a UK tabloid, the Daily Mail. This research will use framing analysis to explore the changes in trust and risk before-, during- and after the crisis, after which quantitative analysis is used to increase the reliability of the results. However, to continue this study, it is important to start with reviewing the existing literature on the

importance of the context of this study, social media, and how it is of relevance for this study.

2. LITERATURE REVIEW 2.1 Social Media

To start, the context is looked at, social media, and, more specifically, Facebook. Social media is all around us and is almost an indispensable part of users’ lives through smartphones (Chugh & Ruhi, 2018). Social media can best be described as online platforms that facilitate connectivity, communication, and collaboration (Zincir, 2017). Whereas social media is the concept that the users use, the company behind the social media facilitates the social

8 people, technology, and the processes, is an online service allowing users to create a private or public profile to connect and reach out to their social connections (Boyd & Ellison, 2007). This research will focus solely on Facebook as a social media website. Facebook Inc.,

consisting of Facebook, Whatsapp, Instagram, and Facebook Messenger, has 2,2 billion daily users, according to their website, and was founded in February 2004. Facebook, on itself, as a social media platform had 1,63 billion daily users on average for September 2019. The aim of this study is solely on Facebook, not on the other social media platforms of Facebook Inc. This study does so because Facebook was a crucial player in the Cambridge Analytica scandal with, according to Isaak & Hanna (2018), Facebook selling their user data to the Cambridge Analytica company.

The Cambridge Analytica affair is considered to be one of the biggest scandals of the last decade, with 87 million people affected. Cambridge Analytica had access to data from 87 million Facebook users, without the approval of the users themselves (Schneble et al., 2018). Later, this data was used to micro-target Facebook users with political messages, swinging some important elections. The scandal showed the world the privacy risks of using Facebook and how vulnerable users but also governments are to these risks.

Privacy concerns rose among its users around the world, and many users said they would delete Facebook; there even was a #DeleteFacebook movement on Twitter (Gilbert, 2018). However, according to statista.com, the number of worldwide users is still rising every year: at the end of 2017, Facebook had 2,1 billion monthly users, and halfway in 2019, it had 2,4 billion monthly users. It seems that people said they would care about privacy but in practice did not, which is the essence of the privacy paradox. The aim of this study is to explore how the key concepts of this paradox, trust and risk, changed on a societal level with the Analytica scandal. To do so, the study continues by exploring the privacy paradox as known from earlier research on the individual level.

2.2 Privacy Paradox

9 their findings ‘showing that privacy concern was not an important predictor of negative customer behavior outcomes’ (Martin et al., 2017, p. 52). In the study of Norberg et al. (2007), where the privacy paradox was introduced, the paradox consisted of two key concepts; trust and risk.

Trust

Despite the many definitions of the concept of trust, this study finds the definition by Mayer et al. (1995) to be the most complete. That definition is: “the willingness of a party (trustor) to be vulnerable to the actions of another party (trustee) based on the expectation that the other (trustee) will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party (trustee)”(Mayer et al., 1995, p. 712). As Bhattacherjee (2002) states, within the online world, the users would be the trustors and the online companies the trustees. That is because the users (consumers) give personal

information, e.g., date of birth, e-mail address, or phone number, to the online firms and are therefore vulnerable to the behavior of the online firm. The users are limited in their ability to monitor the firm’s use of their personal information, which creates the need for trust

(Bhattacherjee, 2002). Data privacy is the issue here, whether the user’s data is kept private.

When trust is mentioned to consumers who are asked about their willingness to give personal information to marketers, negative privacy perceptions appear not to reflect their desire to provide personal information (Norberg et al., 2007). The higher the level of trust, the more willing consumers are to provide personal information (Schoenbachler & Gordon, 2002; Hoffman, Novak & Peralta, 1999). Also, Milne and Boza (1999) studied the impact of trust on a consumer’s sense of privacy, stating that creating trust is a better way to manage customer data than trying to reduce consumer’s concerns about privacy.

The results of these studies show that companies can lessen concerns about privacy from their customers by building trusting relationships with them.

Risk

10 Earlier research showed that risk does influence the perceptions of privacy (Norberg et al., 2007). White (2004) states that perceived disclosure consequences reflect one’s perception that negative outcomes can be bigger and greater than the potential advantages when personal information is disclosed. These negative perceptions can affect people emotionally,

materially, and sometimes even physically (Moon, 2000). Concern about these possible negative outcomes had a bigger impact on perceptions of privacy than the level of trust in a company or organization did (Horne & Horne, 2002). These studies do suggest that privacy concerns are influenced by risk; some argue that risk is even of a bigger influence than trust.

However, all these studies on the concepts of risk and trust are on the psychological and individual level. This study’s focus is to extend the privacy paradox model from Norberg et al. (2007) with new findings and by changing its scope from the individual level to the societal level. To be consistent with this focus, it is important to find out what these concepts mean on the societal level and how they can be influenced and changed.

2.3. Trust and risk on the societal level

Since this study aims to expand the privacy paradox from an individual and psychological level to a more societal level, it is important to explore and describe the two concepts on the societal level as well. The difference is that the scope will change from micro-level, the individual level, to the macro-level, the organizational and societal level, and that the different values and factors affecting it will change. The two concepts within this new macro-level view are explained more extensively below.

Risk on a societal level

11 Breakwell (2003) studied the risk amplification around the 1995 oral contraceptive pill scare. In their research, they showed that if people were more exposed to the hazard, with series of notifications about the hazard, the risk became more normalized. According to Lima, Barnett & Vala (2005), this work is helpful in ‘highlighting the role of mediated- rather than direct-experience in risk normalization and in positing the processes through which this occurred’ (Lima et al., 2005, p. 1231). This is of importance to keep in mind with this study. That is because the data, the news media, is mediating between the hazard, the data breaches, and the society. The reporting can lead to risk normalization and plays a role in the privacy paradox.

On the other hand, as Lewis & Tyshenko (2009) show, there is also something called social risk attenuation, whereas the risk is seen as of less importance to the public. In their study, Lewis & Tyshenko (2009) show that the public in Canada did not perceive the risk of the mad cow disease as high and, inappropriately, socially attenuated the risk of it. The social

attenuation of risk may be because the public concern was directed at other things. For instance, as Lewis & Tyshenko (2009) mention in their study, the public concern was more directed to the US-Iraq war than the mad cow disease. This leads to less attention from the media and, therefore, social attenuation of the risk. Since the numbers show no decrease in the amount of Facebook users, this may also be an important part of the risk in the privacy

paradox.

This research uses both the amplification and the attenuation aspects of risk in the analysis to find out what the scandal did to risk on a societal level. The studies mentioned before,

highlight the importance of risk, however not in the context of social media. Also, the concept of risk and its changes pre-, during and post a crisis time frame have not been studied yet. Therefore, to contribute to the literature, this study will highlight the concept of risk, pre-, during and post the Cambridge Analytica scandal.

Trust on the societal level

Trust on the societal level is important, as Fukuyama (1995) states, trust is a key characteristic for a developed society, and the foundation of our social capital, the foundation of our society. The definition of trust from Mayer et al. (1995) can still be applied to trust on a societal level. However, it can be extended more; trust on a societal level can also be described as

12 on a micro-level is affected. As they explain, a trust crisis is mostly the result of trust on a macro-level, so the trust in, most of the time, large organizations, institutional-based trust. That shows the need for more exploration of institutional-based/societal trust and how it develops with a crisis to find out how to eventually repair this trust.

In this study, Facebook in itself is not an institution, but the governance over Facebook is. That is the (worldwide) regulation for data privacy to keep things like the Cambridge

Analytica scandal from happening, which failed. In comparison. Bachmann & Inkpen (2011) did a study on how to build institutional-based trust in times of the global financial crisis. They did so by looking at four mechanisms through which institutions can, in their view, reduce risk and advance the trust-building processes in inter-organizational relationships. These four concrete mechanisms are legal provision, corporate reputation, certification of exchange partners, and norms, structures, and procedures.

In the context of this research, with the Analytica scandal and Facebook, these mechanisms are interesting to look at. In Bachmann & Inkpen (2011) their study, especially the

mechanisms of legal provision, reputation, and the norms, structures, and procedures, are interesting. The mechanism of certification of exchange partners might not (yet) be applicable to the world of Facebook since there are no certifications for social media companies, and if they were to be made, it would fall under the legal provision mechanism. The three other applicable mechanisms to be looked at are described briefly below.

The legal provision mechanism is that of laws and legal rules. It was first thought and assumed, even in academic contexts, that trust and legal provisions do not go hand in hand very well (Bachmann & Inkpen, 2011). However, newer research shows that the presence of reliable contract law does positively affect the development of trust in business relationships, as relationships mainly based on legal rules can significantly lower the risk of trust

(Arrighetti, Bachmann & Deakin, 1997). Law can be efficient in reducing risk because it aligns the expectations and behaviors way before the disagreements come up. The mechanism can be used to describe why the level of trust has or has not changed during a time period and is therefore included in this research.

13 defined by how the media write about the company. That reputation will eventually have influence on the level of trust (Bachmann & Inkpen, 2011) the users will have on Facebook. Corporate reputation is a mechanism that transfers informal behavioral norms (Bachmann & Inkpen, 2011) and has a lot of influence on the perceived trust level. It is therefore of

importance to this study because it can explain certain levels of trust in Facebook.

The last mechanism this study uses for the level of trust is the mechanism on norms, structures, and proceedings. This mechanism is, as Bachmann & Inkpen (2011) describe, about creating and having a community, mostly online. In that community, certain norms, structures, and proceedings are set regarding the behavior of the users of the community, and these norms, structures, and proceedings will make individual and collective behavior more predictable. The users are more willing to invest trust in a relationship with these norms, structures, and proceedings than to invest in a relationship without them.

Role of the media for both concepts

As briefly mentioned, the media plays a role in creating the perception of trust and risk on the societal level. As Barnett & Breakwell (2003) stated in their study on the risk amplification around the 1995 oral contraceptive pill scare, the perception of risk can become more

normalized. According to their study, the media play a big role in this in the way that, if they highlight the hazard of an event more to the people, the risk becomes normalized. So in this case, the media first amplified and later normalized the risk after a crisis.

On the contrary, Humphreys & Thompson (2014) show in their study that the media can actually fulfill a positive role as well, one that helps to recreate the old reputation. In their study on the public discourse surrounding two oil spills, they find that news media coverage has a lot of influence on the reputation of a company, negatively, but can ultimately restore the old level of trust by the right way of news coverage.

14 To continue, this research will use frame analysis for the data analysis and get a complete overview of both concepts of the privacy paradox, on a societal level. However, this study starts by first explaining the methods used for the analysis.

3. METHODOLOGY

As mentioned, this study aims to find how society’s perceptions of trust and risk change, in times of an online crisis, such as the Cambridge Analytica scandal. More specifically , the aim of the study is to answer the following question: ‘How does the balance between trust and risk, on the societal level, change before-, during- and after a crisis?’. To answer this question, this study conducted a longitudinal frame analysis on how the concepts of risk and trust have changed in the media. The period of the study is three years, from 2016-2018. This timeframe is chosen because the Cambridge Analytica scandal became world news and public at the beginning of 2018. Therefore, the first aim, and frame, is to study how the concepts ‘trust’ and ‘risk’ were in two years before the scandal, in 2016 and 2017, when the world did not know about the selling of data in that large quantities yet. The second frame is in the first half-year of 2018 during the scandal, when the newspapers, especially the Guardian, brought the news to the world and showed that Facebook sold people’s personal information as data, without their approval. The last frame is in the second half-year of 2018, after the scandal was brought to the news and people knew about it, and their reactions to it.

This study is based on a longitudinal frame analysis of three large UK newspapers, the Guardian, the Telegraph, and the Daily Mail. At first, these three newspapers were chosen because these newspapers are the leading newspapers in the UK when it comes to the number of readers they have and the reach of the papers. According to PAMCO, the Telegraph has a monthly reach of 21.509.000 adults, the Guardian has a monthly reach of 23.931.000 adults, and the Daily Mail reaches out to 25.278.000 adults every month, with adults meaning people who are 15 years or older. Compared to other newspaper brands in their segments, these numbers are among the highest. The Guardian and the Telegraph are placed in the ‘Qualities’ segment, with the Guardian being the biggest and the Telegraph as third-biggest. The Daily Mail is placed in the ‘Mid-Market’ segment, being the biggest in that segment.

15 Guardian is seen as a left-wing newspaper, described as ‘the British daily of general

circulation that is the furthest to the left of the ideological spectrum, traditionally supporting the Labour Party’ (Carvalho, 2005, p. 2). The Telegraph is the most right-wing newspaper with a conservative stand (BBC, 2009). It is well-known for its strength in covering foreign affairs (Tunstall, 1996).

Newspapers as a source of data were chosen for several reasons. Newspapers have been used much for research in the communication discipline, as they were used to represent the public opinion in this research (Gamson, Croteau, Hoynes & Sasson, 1992; Schudson, 2003). They can be used as a bundle of quotations from consumers, law enforcement, civic and cultural leaders and, most important, the newspapers can be used to represent the landscape of public opinion and semantic association on a given subject (Deephouse, 1996; Gamson et al., 1992; Gamson & Modigliani, 1989; Schudson, 2003). Also, as Humphreys & Latour (2013) write, newspapers are reliable and stable indicators of the public opinion. They form a metric for settled cultural associations that are comparable over time. Furthermore, archival, or

newspaper, analysis is able to create a representation of the opinions as they were observed in the specific time period, without the interference of individual memory that retrospective interviews on occasion produce (Bernard, Killworth, Kronenfield & Sailer, 1984; Golder, 2000; Humphreys, 2010). Lastly, newspapers belong to the media and media can play a big role in defining the public perception, as Barnett & Breakwell (2003), as well as Humphreys & Thompson (2014), showed in their studies. It works both ways; the public partially decides what the media covers and how extensive things are highlighted, but the media also defines opinions for people, in a way that people define their opinion based on what they see and read in the media.

16 newspapers, from left- and right-winged sides, this study aims to incorporate the voice of every English person and be as neutral and objective as possible.

For this study, the data was derived using LexisNexis, and at first, a pre-selection was done based on whether ‘social media’ was in the headline or the lead paragraph of an article from these three newspapers. After the general pre-selection, another criterion was added. Since this study focuses on the event of changes pre-, during- and after the Cambridge Analytica scandal, with the main actor being Facebook, this study chose to use only articles with ‘Facebook’ in it for the analysis. This is done because Facebook, as a company, faced many challenges with the scandal and is the main actor since their users were being targeted, and their data was gathered. That is why this study chose to narrow down the articles to just this selection. This resulted in a total of 2.641 articles used for this study, from the three

newspapers combined.

In this research, frame analysis is done using the constructionist approach (D’Angelo, 2002). This sort of frame analysis first involves identifying clusters of messages which have similar rhetorical strategies, which show and promote specific facts and interpretations of trust and risk perceptions of Facebook (Entman, 1993; Tucker, 1998). This study identified three frames from the data. These frames are assumed to be the general opinion of the public on the topic. Otherwise, it would be impossible to use the frames to identify the changes in the trust and risk perceptions.

The first frame is identified as ‘Facebook is connecting the world’. This frame is based on the earlier days of Facebook, where the company was seen as a more ideological company, really eager to connect the world for free. Articles focusing on the positive side of the social media website and key rhetorical strategies are the people using the website for the better. Other key rhetorical strategies include happy feelings, positive emotions, and people focusing on the good, ideological, connecting, and trustworthy side of Facebook. There is no perceived risk of using Facebook. The second frame identified is in times of the crisis, ‘Facebook is a

17 negative about Facebook as a company, and even the CEO of Facebook, Mark Zuckerberg, had to explain the actions of the company in front of Congress in the United States of America. So the rhetorical strategies were the legal aspects, Facebook as a fraudulent corporation, the (privacy) risks of using Facebook, the loss of trust in the social media website, and also the fines Facebook had to pay. Finally, the last identified frame is ‘Facebook remains’. The ideological view on Facebook disappeared, but it can continue in their way of doing and is not suffering very much from the scandal. It can also be described as ‘shrugging the shoulders’ and continue to let everything go back to normal. The key rhetorical strategies for this frame are the continuous use of Facebook, why deleting your Facebook account does not really matter, the still growing numbers of users around the world, risk of using Facebook becomes normalized and is socially attenuated, trust is slowly restored and, lastly, the legislations are still not changed.

Next to the qualitative analysis with the frames, a quantitative content analysis was done to triangulate the research, increase the validity of the research, and in order to document the trends. In this study, triangulation is done by using LIWC 2015 and SPSS to supplement the qualitative findings. This is used to increase the reliability of the findings further, and

therefore increase the reliability of the study as a whole. This study does a quantitative analysis of these articles using the Linguistic Inquiry and Word Count (LIWC) system of Pennebaker, Boyd, Jordan, and Blackburn (2015). The LIWC analyzes a set of articles or texts on a reliable level and at high speed, automatically arranging the frequencies of word usage for 100 different categories. It has been used as a valid text analyzing tool for different kinds of linguistic expression in different sorts of studies (Bulkely & Graves, 2018) and is therefore also used in this study to analyze the articles from the newspapers. The program LIWC 2015 was used first to analyze the selected articles using two dictionaries. These two dictionaries can be found in the appendix. The first dictionary is the dictionary for the concept of ‘trust’ and was created especially for this study. It was validated by independent judges and consists of concepts and words associated with the concept of ‘trust’ in general. The second dictionary is the dictionary used for the concept of ‘risk’. For this concept, the study chose to use the internal dictionary of LIWC 2015, as this study finds that dictionary sufficient to explore the concept of ‘risk’ in the selected articles. After the content analysis with the LIWC 2015, this study continued by doing a one-way ANOVA to check for significance in

differences between the time periods.

18 the difference in trust and risk in these periods are significantly different, to the values for trust and risk changed in these periods. As explained before, the three periods are pre-,

during-, and after the scandal. To capture the time before the scandal, this study uses the data, so all the articles, from 2016 and 2017 of the three newspapers. To see what happens during the scandal, this research uses the data from the first half of 2018, since the scandal happened at the beginning of 2018. Lastly, to measure the changes in risk and trust after the scandal, the data from the second half-year of 2018 is used. That is because by the end of 2018, the

scandal was over, newspapers were not writing much on the scandal anymore, and other events happened, making it less topical than it was before. With these three different data sets, representing the three different time periods, it will be interesting to see what happens to both trust and risk in times of crisis. In the next section, the findings of the study are discussed.

4. FINDINGS

4.1 Quantitative Findings

19 Graph for ‘risk’ (1,2 and 3 representing the different time periods)

The results of the quantitative analysis for the concept of risk show that a one-way-ANOVA, with the three defined time periods as the independent variable and the percentage of words from the ‘risk’ dictionary, from the dataset, as the dependent variable, was significant (F(2,15) = 6,34, p < 0,05). A post hoc test reveals that there is a significant increase in the words belonging to the ‘risk’ dictionary, from time period 1 to time period 2. There is no significant change from period 2 to period 3, implying that risk perceptions remained high and

20 Graph for ‘trust’ (1,2 and 3 representing the different time periods)

The results of the quantitative analysis for the concept of risk show that a one-way-ANOVA, with the three defined time periods as the independent variable and the percentage of words from the ‘trust’ dictionary, from the dataset, as the dependent variable, was significant (F(2,15) = 5,1, p < 0,05). A post hoc test reveals that there is a significant increase in the words belonging to the ‘trust’ dictionary, from time period 1 to time period 2. There is no significant change from time period 2 to time period 3. The results are almost the same as was the case with the risk perception, showing that the level of trust remained low after the

scandal. The reasoning behind this is explained more in the qualitative findings.

The balance between both concepts seems changed due to the scandal, and both concepts are heavily affected by it. The quantitative analysis shows some interesting results which will be elaborated more on in the qualitative findings part.

4.2 Qualitative Findings

21 ‘Facebook remains’, is observed in the whole of 2018; that is because people already expected that Facebook could just continue. The three frames and the qualitative analysis are explained next.

First frame ‘Facebook is connecting the world’

The first frame identified, ‘Facebook is connecting the world’, is about the fact that Facebook is perceived as an ideal, almost non-profit company with its main aim to connect the world. Facebook connects people, and the youth uses it to preserve contact with their friends (Wood, Bukowski & Lis, 2016). It gives a new dimension to the social world of people. Also, as Ellison, Steinfield & Lampe (2007) suggest, there is a firm association between the use of the social media platform Facebook and social capital. With social capital meaning the resources acquired through relationships among people (Coleman, 1988). As Ellison et al. (2007) also show, Facebook might provide better benefits for users of Facebook who are experiencing lower self-esteem and lower life satisfaction. This shows only positive effects of Facebook, and the rhetorical strategies used were happy feelings, positive emotions, and people focusing on the good, ideological, and connecting side of Facebook. A good example from the data is:

‘GET SOCIAL MEDIA SAVVY

First, head to Facebook. Using quote marks when searching names will narrow down results for example, Hannah Brown' won't show various Hannahs and Browns. If you find the person you're looking for, add them as a friend first. Messages from strangers land in another' folder and risk being unseen.’ (Daily Mail, 03-10-2016)

This quote shows the ideal essence of Facebook, connecting with (old) friends. It even describes how to handle connecting to friends and what to do. The quote highlights the main purpose of Facebook, connecting with friends. As also can be concluded from this quote is the fact that Facebook is seen as a trustworthy social media website to use, as it is even

mentioned as the first step in the process. It seems that there is no perceived risk in using Facebook.

22 ‘Experts blame the growing influence of social media sites such as Instagram, Facebook and Twitter for encouraging feelings of insecurity. There is huge pressure on youngsters to log on to these sites on smartphones and post photos and comments about their apparently exciting' lives. This is encouraging competition, self-loathing and even bullying, particularly among young women, experts say. The report by NHS Digital was based on a survey of 7,500 adults aged 16 and above about their mental health. But of most concern are women aged 16 to 24. Some 26 per cent of them were found to be experiencing common mental health disorder symptoms, up from 19 per cent in 1993. These include depression, anxiety, obsessive behaviour and sleeping problems - which affect only 9 percent of men of the same age.’ (Daily Mail, 30-09-2016)

The mental health issue with Facebook and other social media websites was identified in research as well. For instance, Kross, Verduyn, Demiralp, Park, Lee, Lin, Ybarra (2013) showed in their study that Facebook use could result in a decline in how people feel moment-to-moment and how satisfied they are with their lives. Also, as Blease (2015) describes in his study, too many friends might also lead to negative outcomes, even mild depression. This shows the downside of this ‘connection’ and the downside of the ideological view of

Facebook and shows that there might be risks in using the social media website. Trust is not affected by this, but the risk is there, although people might not directly experience these risks, experts see and show these risks. The risks are not experienced directly from the company Facebook or the privacy risks; these risks arise due to comparison with peers of people. Trust remains in the company, not being affected by this.

However, the main theme among the data is still positive and focused on the positive,

trustworthy, and connecting side of Facebook. This ideological side of connecting the world is perfectly shown in an article from the Guardian in 2016.

The distance between us is getting smaller, too, as more people sign up to the platform, (..) with twice as many people on Facebook today, we're more interconnected, shortening the distance between any two people in the world. (The Guardian, 05-02-2016)

23 amount of users worldwide, Facebook as a company covers almost the whole world, and this quote is almost the perfect representation of the first frame. To continue, the second frame of this study is looked at.

Second frame ‘Facebook is a fraudulent company’

The second frame, ‘Facebook is a fraudulent company’, shows what happened to the societal perceptions of risk and trust during the scandal, in the first half-year of 2018. It gives insights into how people reacted when they found out their Facebook information was secretly sold to another company. For this frame, the rhetorical strategies used are; the legal aspects,

Facebook as a fraudulent corporation, the (privacy) risks of using Facebook, the loss of trust in the social media website, and the fines Facebook had to pay. The frame highlights the articles that show the rising interest in the debate on the societal impact of technology and its risks towards the privacy of people (Kosinski, Bachrach, Kohli, Stillwell & Graepel, 2014). The frame shows the negative sides of Facebook and the consequences of the scandal.

The first quote is from the Telegraph gives a brief explanation of what happened.

‘Facebook (…) has been embroiled in controversy since it was revealed that 87 million people had their personal data harvested by political consulting firm Cambridge Analytica without their knowledge. Facebook applications, such as quizzes and games, were able to harvest data from all of a user's friends' profiles, even if the friends themselves had not used the app.’ (The Telegraph, 14-04-2018)

The quote summarizes the scandal in a good way and shows the breach of trust for users of Facebook. As the quote explains, the data was harvested even if users did not use the

application, only because their Facebook friends used the application. This shows the breach of trust in the company, since data is taken unknowingly. This harvesting of data without people knowing it shows the lack of two critical trust builders, according to Bachmann & Inkpen (2011); this taking of data leads to bad corporate reputation but also shows the lack of legal provision. Also, the risk of using Facebook is highlighted, since your personal

24 amplification of risk (Barnett and Breakwell, 2003), with the negative sides highlighted in the quote. The quote shows how both risk and trust are affected by the scandal.

The second quote is from the Guardian and shows how Facebook works, as a business, to make money.

Facebook "farms" its users for data: the more they produce - the more "user engagement" there is, in other words - the better. Consequently, there is an overriding commercial imperative to increase levels of engagement. And it turns out that some types of pernicious content are good for keeping user-engagement high: fake news and hate speech are pretty good triggers, for example. (The Guardian, 25-06-2018)

The quote gives a clear representation of how Facebook as a company works and how it makes money. The quote even makes a small comparison with Facebook and a farm, by saying that Facebook ‘farms’ it’s users for data. This study continues with that comparison as this study finds a resemblance between a farm and Facebook. The main product for Facebook, the product it farms, is the user data. As the quote also explains, the more they have, the better. To increase the crop and the harvest, so to farm more data, Facebook uses fake news and hate speech, in farming terms; ‘pesticides’. It shows how Facebook uses unethical ways to gather user data and how risky the use of Facebook is. This is already explained in research as well, for instance, as Iosifidis & Nicoli (2019) describe, Facebook has a big influence and ‘the platform has been used to promote violent acts of terrorism, distort presidential election results, influence perceptions and sway public opinion during important democratic

moments.’ (Iosifidis & Nicoli, 2019, p. 7). This highlights the risks of using Facebook to the citizens and the impact these risks have. As Iosifidis & Nicoli (2019) also explain in their study, Facebook is reluctant to remove the fake content from their platform, and there is a need for legislation in this area. The latter, the needed legislation, can also be concluded when looking at the results of Bachmann & Inkpen (2011) and will be needed to restore the trust in Facebook. That is because, as Bachmann & Inkpen (2011) describe, to gain trust among the public, one tool could be to make laws and install legal provision for companies to adhere to.

25 It is a media company with enormous influence in shaping someone's worldview about whom to trust. And it is profit-driven. "Facebook makes money if the advertiser pays, regardless of whether people's lives are being improved," says Heiferman. In May 2017, Facebook

reported that 98% of its quarterly revenue came from advertising, up from 85% in 2012. In other words, it's in the company's interests to keep our eyes glued to the screen, no matter what the content.’ (The Guardian, 11-02-2018)

This quote highlights the money-making aspect of Facebook, whereas it only cares about the payments of advertisers, not if people’s lives are improved. Whereas in the first frame, the assumption was that Facebook was not focused on making money, this quote shows that their focus has shifted to it. The more ideological view on Facebook changed to an attitude as if Facebook is money-making monster. There is also a clear link with the earlier quote on the content, saying that the company wants ‘our eyes glued to the screen, no matter what the content’, as Iosifidis & Nicoli (2019) already explained. Risks are highlighted, with the fact that Facebook does not care about whether people’s lives are improved and when a company only wants our eyes glued to a screen, the company’s reputation decreases, and therefore trust in the company as well (Bachmann & Inkpen, 2011).

The last quote of the ‘Facebook is a fraudulent company’ frame is a statement of not Mark Zuckerberg, but another founding president of Facebook.

‘In an unprecedented attack of candour, Sean Parker, the 38-year-old founding president of Facebook, recently admitted that the social network was founded not to unite us, but to distract us. "The thought process was: 'How do we consume as much of your time and conscious attention as possible?'" he said at an event in Philadelphia in November. To achieve this goal, Facebook's architects exploited a "vulnerability in human psychology", explained Parker, who resigned from the company in 2005. Whenever someone likes or comments on a post or photograph, he said, "we... give you a little dopamine hit". Facebook is an empire of empires, then, built upon a molecule.’ (The Guardian, 04-03-2018)

26 human psychology, and this highlights the increased perceived risk in using the social media platform. It says Facebook is an empire of empires, regarding the power they have as a company. Their power also became very clear with Cambridge Analytica, with even

influencing outcomes of elections (Isaak & Hanna, 2018). To tackle such a problem, laws are needed, according to Isaak & Hanna (2018). The importance of this aspect for the concept of ‘trust’ on a societal level was also addressed in the literature review as well, where Arrighetti et al. (1997) showed that the presence of a good and reliable contract does increase trust in business relationships. These privacy laws will increase trust and automatically reduce the perceived risk. However, for risk to become normalized, according to Barnett & Breakwell (2003), a series of notifications about the hazard is needed.

Third frame ‘Facebook remains’

The third frame, ‘Facebook remains’, is, in essence, about the privacy paradox as Norberg et al. (2007) discovered it on the individual level. It shows that many users may be showing theoretical interest in their privacy, even keeping a positive attitude to active protection behavior of their privacy, but not translating it into actual behavior to protect their privacy (Joinson et al., 2010). These studies were done on an individual level; however, in this study, with the qualitative analysis, the privacy paradox can be seen on a societal level as well. It shows that, despite the scandal and its consequences, Facebook can continue as a company, and people still use the social media platform. To identify this frame, the rhetorical strategies that this study used are; the continuous usage of Facebook, why deleting your Facebook account does not really matter, the risk of using Facebook becomes normalized and is socially attenuated, trust remains low, the still growing numbers of users around the world and that the legislations have still not changed.

One quote highlighting the frame perfectly is from the Telegraph and is about why people should not join the #DeleteFacebook movement.

27 Facebook at all. (...) So I'm just about OK with Facebook taking my data and doing

interesting things with it, and I still believe that the biggest and best tech firms have good things to offer the world. (The Telegraph, 25-03-2018)

This quote shows two things. First of all, it starts with stating the assumption that nothing is for free and that users should have known this beforehand. One could indeed assume that by creating a public profile, as mentioned by Boyd & Ellison (2007), you could have known that these profiles can be used for other things. However, as Isaak & Hanna (2018) state, not having Facebook did not mean your data was not used. As they explain, even websites with a Facebook logo on it are linked to it, and through other touchpoints, Cambridge Analytica was able to analyze people. As they also explain, at that time, Facebook and other social media companies were not obligated to let their users know what can and/or will happen with their user data. So, basically, the user was left in the dark about what happened at the back-end of their profile.

What is also interesting from this quote is the latter part, the part where the author finds peace in sharing data with Facebook, with the belief that tech firms have good things to offer in the world. This is almost the definition of the privacy paradox from Norberg et al. (2007), in where people first say they care about their privacy and trust and risk perceptions are affected, but after some time, everything is going back to normal under the assumption that the tech companies will still have good things to bring to the world. Trust seems restored, despite the scandal, and it seems there is almost no perceived risk. Users are mostly aware of the privacy risks in the online world and still tend to share information in return for personalized services and retail value (Acquisti and Grossklags, 2005; Barth & De Jong, 2017).

Another fact about social media platforms is written in another article from the same newspaper, the Telegraph. They write the following:

‘Social media will overtake television to become the UK's biggest advertising format in two years, according to forecasts, despite the shadow hanging over the industry from the Facebook privacy controversy.’ (The Telegraph, 02-04-2018)

28 There are still good returns on advertisements; otherwise advertising companies would not use social media websites, which is because people keep using the social network platforms, despite the high risk and low trust. The continuous use of social network platforms might be because people feel that using social media networks has more benefits than risks, like Barth & De Jong (2017) describe. Despite the high risks and low trust, the benefits outweigh these facts. The quote also displays the fact that other companies, companies that want to advertise, feed social media companies by using more and more social media platforms to advertise. This way, social media companies like Facebook are stimulated by both the users as the advertising companies, giving them all the power.

One of the reasons for Facebook being able to remain as a company and for Mark Zuckerberg to keep his head up is the fact that they do not solely own Facebook as a social media

platform but also other platforms. As this quote from the Telegraph perfectly shows, the company known as Facebook Inc. is still growing in its revenue.

‘While there remain questions over Facebook's other two major acquisitions, the $19bn takeover of messaging app WhatsApp and the $2bn purchase of virtual reality start-up Oculus, Instagram can count as nothing less than a triumph. The app is now growing faster than Facebook itself, having signed up 200m new users in the last nine months, and unlike Zuckerberg's other two major acquisitions, is contributing handsome sums to Facebook's top line. Worldwide, it is expected to pull in $8.1bn in advertising revenue this year, almost double a year ago, according to estimates from eMarketer. By 2020, this will have doubled again, to $17bn.’ (The Telegraph, 22-06-2018)

The quote shows that Facebook is just a company, not an ideological connecting firm. If Facebook, as a social media platform would fail, the company itself can remain. As mentioned in the second frame, in the quote from a founding president of Facebook, this quote confirms that Facebook Inc. is an empire of empires and has incredible power. With their acquisition of other platforms, they secure their position, with Instagram being the biggest triumph of their strategy. Also, the quote does an estimation of the advertising

29 To end this last frame, this research found a perfectly summarizing quote in the Telegraph.

"The CA (Cambridge Analytica) scandal has meant that some people have deleted their Facebook accounts." However, she also pointed out that most people would say "now that it's out it is business as usual". She added: "The power to connect may seemingly override the power to keep one's 'patterns of life' private." (The Telegraph, 03-05-2018)

This quote gives a potential reason as to why Facebook is still able to continue, namely that people feel that ‘now that it's out it is business as usual’. This study finds that an interesting reason that can be explained by, for instance, Barnett & Breakwell (2003). In their study on risk amplification, they found that when people are more exposed to the hazard, with a series of warnings about the hazard, the risk can become normalized. In this case, this can be

applicable, with the media, newspapers, TV, online, magazines, etc., extensively highlighting the scandal, the risk of the scandal can have become normalized. In the LIWC analysis, with words used to describe ‘risk’, the graph increased during the scandal but decreased little after the scandal, showing that, in this study, newspapers kept reporting the risks of Facebook. Therefore, this study assumes that newspapers were a big contributor to the normalization of risk. Also, a potential reason for the fact that there is less perception of potential risks may be that the risk is attenuated. In their study, Lewis & Tyshenko (2009) show that risk can be attenuated, instead of amplificated, for several reasons. The first reason could be the fact that in the time of the scandal, there was other major news in the media, and the scandal was not highlighted that much. The second reason for social attenuation of the risk may be the fact that the people found the scandal less important than other things happening at that time or the fact that people, in general, did not really care that much about their privacy, as was

suggested.

30 might also be because people do not specifically have trust in Facebook, but in the institutions that rule over Facebook, the governments. This study assumes that people do not trust

Facebook as a corporation anymore, explaining the quantitative findings, but trust that new legislation will come to limit Facebook in their power. According to Bachmann & Inkpen (2011), creating new and better legislation is an important tool in creating institutional-based trust. The last mechanism for institutional-based trust, proposed by Bachmann & Inkpen (2011), is that of norms, structures, and proceedings. In this case, after the scandal, Mark Zuckerberg, Facebook’s founder, was summoned to Washington to appear before Congress and to explain what happened and how this will not happen again. In his statements,

according to Wichter (2018), Zuckerberg stated that Facebook is working to avoid scandals like this in the future and securing privacy. However, statements like this are made before, and it might be that people might not believe the founder anymore, therefore having no or very little effect on the level of trust. It could create new structures and procedures within the company and could potentially lead to restored trust in Facebook, according to Bachmann & Inkpen (2011), however, no direct effect is observed yet, in both the qualitative as quantitative findings.

5. DISCUSSION

The question of this research is: ‘How does the balance between trust and risk, on the societal level, change before-, during- and after a crisis?’ What happened to both concepts on the societal level will eventually explain the privacy paradox on the societal level and lead to the contribution of this study. To do so, this study starts by first explaining the changes for both the concepts, trust and, risk, separately.

31 frame, the level of societal risk remained high, despite a small decrease in the perceived risk in both the quantitative as the qualitative analysis, and the public continued using Facebook against all the odds. This study finds some reasons for this.

The first might be the fact that the risk became normalized, as Barnett & Breakwell (2003) also state in their study. The prominent news coverage by the media, and the fact that the public was exposed to the hazards of Facebook with a series of warnings, led to the normalization of the societal risk, as explained by Barnett & Breakwell (2003).

Another reason for this decrease in societal risk perceptions might be that the risk became attenuated. In their study, Lewis & Tyshenko (2009) state that risk can be publically

attenuated when other major public events are present as well. In the time that the Cambridge Analytica scandal became public, the end of March and the beginning of April, and of course, in the aftermath of it, which endured during the whole year of 2018, other major events happened as well. For instance, the Royal Wedding of the British Prince Harry and Meghan Markle was held on May 19, 2018, and was the most-watched event of 2018 in the United Kingdom (Breech, 2018). Another big event in 2018 was the FIFA World Cup in Russia, in the summer of 2018. England managed to become fourth in the tournament (Fifield, 2018), and the tournament is always a big happening, with almost half of the world watching. With these two events happening right after the Cambridge Analytica scandal, the risk could be attenuated, just as Lewis & Tyshenko (2009) conclude in their study.

Another explanation for the continuous use of Facebook, despite the known risks, might be that people did not feel any personal risk. As Kak (2018) describes, Cambridge Analytica used the data for political persuasion of people; it was not that people their online identity was stolen. Therefore, it might be that people did not feel and see the direct personal risk of using Facebook, only the bigger picture of the risk in general, and continued using the social media platform, despite the high level of societal risk.

32 did not really care about the risk. As explained, it might be that people did not feel the

personal risk and felt no direct effect.

As for the societal level of trust, other conclusions can be made for the different frames. The societal level of trust changed almost in the same way risk changed during the frames. In the first frame, the societal level of trust in Facebook is high. Whereas the potential (small) risks of using the social media platform are known, regarding the insecurity problems among the youth, the trust in the platform is not affected. This could be because Facebook is only the facilitator and the connection, not the reason why people would bully other people. So, trust in Facebook as a company itself is not affected and very high, but the risks of joining

Facebook are known. This is a small discrepancy in the relationship between trust and risk, as stated before. In the second frame, the frame where the Cambridge Analytica scandal was discovered and brought to the news, the societal trust in Facebook decreased to a low point because of the breach of trust. The public lost faith in Facebook, and the ideal image, as stated in the first frame, disappeared. In the eyes of the people, Facebook became a big corporation like any other, violating rules and harming people. In the third frame, ‘Facebook remains’, the trust seems to restore in some qualitative findings, but, looking at the quantitative findings, the trust is not restored yet, only a tiny bit. Looking at the qualitative findings, it does show that Facebook continues, and most users just stayed using Facebook, while aware of the risks. However, this study finds that there is no real sign of trust in the sense that users really trust Facebook with their data. The societal level of trust is not restored, but the people, in general, have accepted it and do not care about it. The public feels as if it is almost inevitable to use something in the present times, without giving away some data. The basic privacy is gone, and people experience it as normal to give away some data, as if it is almost necessary. This is something Young & Quan Haase (2013) also showed in their study.

33 What also might have happened is that Facebook became the same as Primark, an Irish fast-fashion tailor. Facebook became the same as Primark in the sense that it became a big company with many people using it but being a company that cannot be trusted. As The Economist describes in 2015, even when a factory of Primark collapsed in Bangladesh in 2013 due to inadequate safety conditions, killing 1100 people, the consumers soon forgot about it. Primark itself recognized just six months before the fatal accident that the safety conditions were hard to oversee (Butler, 2013) but did not do anything about it. Despite all this, Primark its sales jumped by more than 20% the year after the accident, according to The Economist (2015). This is the same for Facebook, with a growing number of users even after the scandal and knowing that Facebook consciously sold their data to Cambridge Analytica. This study assumes that people accept the fact that some companies, like Facebook and Primark, cannot be trusted, but they became necessities for people. Therefore, the public keeps making use of the products and services of these companies.

Another explanation for the continuous use of Facebook without trust on a societal level, could be the fact that there might be no trust in Facebook as a company, but there is institutional-based trust, as explained by Bachmann & Inkpen (2011), so trust in the

government to come with new legislation. Next to that, the congressional hearings of Mark Zuckerberg can lead to new structures, norms, and proceedings, leading to more trust,

according to Bachmann & Inkpen (2011). However, looking at the quantitative findings, trust is not at that point yet, but it might be in the future.

34 with the qualitative findings, this study assumed that for the concept of ‘risk’, the lower values represent the lower societal perception of risk. For the concept of ‘trust’, the lower value means more trust, and if the quantitative values for ‘trust’ go up, the societal level of trust actually goes down. This tells us that the perception of risk was low in period 1 but increased a lot in period 2. It stayed on the high level in period 3 as well, showing the

perceived societal risk remained high after the scandal. For trust, in the first period, there was much trust in Facebook. In the second period, the line increased much, showing that there was a breach in trust and trust went down. In period 3 it remained on almost the same level as in period 2, indicating that trust was not restored yet in that period.

Looking at the qualitative findings, the balance between risk and trust is disrupted in the first frame, to a small degree. There are already some known risks of using Facebook, whereas the trust in Facebook is not affected. In the second frame, both concepts are aligned in the way as expected, with high societal risk perceptions and low societal trust. In the third frame, the balance between the two concepts changes. Whereas the risk perception goes slightly down but becomes accepted and normalized, the trust is not restored. The societal level of trust is still low, but people seem to be accepting it. The results among these three different frames combined, lead to the main contribution of this research. This study contributes to the

literature on the privacy paradox, especially on the study by Norberg et al. (2007), by shifting the privacy paradox from the individual level to the societal level. The two concepts in the model of the privacy paradox on a societal level show similar results, as Norberg et al. (2007) showed in their study with the privacy paradox on the individual level. As Norberg et al. (2007) state, their hypothesis that trust will influence behavior was not confirmed, and the effect of trust was still a question. In this study, it is shown that trust does not influence disclosing behavior as users used Facebook when there was lots of trust as much as when there was no trust. For risk, Norberg et al. (2007) showed that it affected the disclosure intention on an individual level. For the societal level, this is, as was shown, the case as well. When risk was exposed, people intended to delete Facebook and not use the social media platform anymore. However, most people kept using Facebook, so risk did not affect the actual disclosing behavior.

35 there is more trust. This study assumes that, due to these results, both concepts are related and connected with each other and can predict each other’s outcome. With a low societal trust level assuming there must be high perceived societal risks and low perceived risks leading to a higher level of trust. Ultimately, this study assumes that if there would be a fourth period, e.g., covering the whole year of 2019, the societal perception of risk and the level of trust will both be at the same level as they were in the first frame and period. For the societal level of risk, this assumption is made based on the fact that the attenuation of risk is starting already, with other events happening and the level of perceived risk going down already. The other events, people forgetting about the risk and people not feeling the direct personal risk, ultimately leads to attenuation of the risk (Lewis & Tyshenko, 2009). In time, people forget about the real hazards of using Facebook, and therefore, the level of perceived risk will return to the old level. For the level of trust, this assumption is made because after the hearings of Zuckerberg, the possible new legislation for data privacy and the promise of Facebook to make sure this will not happen again, people will start to trust Facebook again. Especially the fact that there might be new legislation in 2019 and that Facebook rebuilds their good

reputation will restore the level of trust to the old level, according to Bachmann & Inkpen (2011).

Looking at the privacy paradox in the setting from this study, this setting might be an

36 First frame

‘Facebook is connecting the world’ Second frame ‘Facebook is a fraudulent company’ Third frame ‘Facebook remains’

Low societal risk perception High level of societal trust

High societal risk perception

Cambridge Analytica scandal made public by the media

Low level of societal trust

Societal level of risk remains high, slightly lower. Risk accepted and normalized

‘Facebook is just another company like Primark’

Level of trust remains low, there is a small increase. No trust needed for Facebook

Framework of findings

6. CONCLUSION

Understanding the privacy paradox on the societal level is of importance since it shows the behavior of a whole nation. It can be of help for governments in deciding on their new laws and regulations, for companies in creating a new view on how to influence people and

maximize profits, and it shows researchers how society can respond to certain phenomena. As Fukuyama (1995) explains, societies change over time, from very developed to less developed or the other way around, and understanding a phenomenon like the privacy paradox on the societal level can help explain these changes.

37 even without trust, they continued using the social media platform. The users that do care a little about their privacy but do not want to delete their social network accounts used active profile management to limit this data sharing (Young & Quan Haase (2013). Both societal trust and risk combined in the privacy paradox show that the privacy paradox on the societal level does not differ much from the paradox on the individual level.

The results of this study show that in present times people continue to use a social media platform like Facebook, despite knowing the risks of giving away data and not trusting the company with your data. This study made some assumptions on how this is possible, but there is still a lot more to be studied in this area. The responses to a crisis were logical, but the privacy paradox present showed that it takes more than a crisis for people to actually stop using social media websites. This might be due to people’s addiction to it, people having faith in new and useful regulations made by the government or other public institutions, or maybe the fact that websites like Facebook became a part of everyday life. This study shows how the power of a social media giant like Facebook can continue, even with the scandal being

exposed to the public.

The context of the study was topical, and it gave results that can be used in the future. The results can be of importance in the future since this study suspects that the Cambridge Analytica scandal will not stay the only data privacy scandal. With other online applications and platforms, being launched on a day-to-day base, to connect the world and the

governments lacking in creating up to date laws and regulations to protect the public from these new apps, a new scandal is bound to happen. Companies, the public, and the institutions should be aware of this and prepare for the future.

7. LIMITATIONS & FUTURE RESEARCH