Disclosing private information on
Facebook: The effect of perceived
gratification on risk perception
Author: Ignacio Masa González Student number: 11391340
Date of submission: 22nd of March 2018 Study: MsC. In Business Administration Specialization: Marketing Track Institution: Universiteit van Amsterdam Supervisor: Joris Demmers
Statement of originality
This document is written by Ignacio Masa González 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.
Acknowledgements
The present study would have not been possible without the invaluable help and the support of Joris Demmers. He not only helped me when I faced difficulties, but also knew how to encourage me to develop my independent thinking and solve any issues I encountered by myself.
Abstract
Although previous research has established the pillars of how people measure benefits and risks, some of the links between perceived risks and perceived benefits have not yet been fully established; is there a direct relationship between the two variables, or are they independent of each other? This study aims at investigating the extent to which perceived risks have an influence on perceived benefits and at how this relationship is established. Also, the role of emotions is evaluated when measuring this relationship, are liked companies more
trustworthy than the disliked ones? The context for the study is the online disclosure of private information on Facebook, a complex ecosystem from which companies can obtain valuable information from consumers.
In previous literature, a relationship between risks and benefits has been identified but not clearly explained. This study adds a new component to the discussion: the existence of a negative relationship between perceived risks and perceived benefits. Confirming once more human irrational behavior when making choices, results suggest that perceived risks are lower when perceived benefits are higher and vice versa. Besides this, the presence of liked
companies in a purchase environment affect risks in such a way that they decrease the more affinity the consumer has with it.
Table of contents
1. Introduction………6
2. Literature review………9
2.1 The Privacy Paradox and reasons for private information disclosure………9
2.2 Risks and benefits: a psychological perspective………...12
2.3 The affect heuristic: likeability and risk perception………..13
2.4 Facebook as a context for online private behavior………....15
2.5 Research question and gap………..………. 17
2.6 Conceptual model………. 18 3. Method………...………20 3.1 Sample………...………20 3.2 Design……….………...……...….21 3.3 Statistical method………..24 3.4 Pretest………24 4. Results………...……….27 4.1 Descriptive statistics………..27 4.2 Factorial Anova………...………..29 5. Discussion………..33
5.1 Theoretical and practical implications………..………….…...…....….33
5.2 Benefits and limitations………...……….36
5.3 Future research……....………...………38 8. References……….39 9. Appendix………...………42 9.1 Factorial anova………...…...42 9.2 Experiment Scenario 1………...………...………..…..43 9.3 Experiment Scenario 2………..44 9.4 Experiment Scenario 3………..45 9.5 Experiment Scenario 4………..46 9.6 Descriptive statistics. ………..………...……..47
List of tables and figures Tables
Table 1 The four scenarios 23
Table 2 Descriptives Dependent Variable 28 Table 3 Likeability means 29
Table 4 Perceived benefits means 29
Figures
Figure 1 Conceptual model 18
Figure 2 Effect of perceived benefits 30 on perceived risks
Figure 3 Effect of likeability on 30
1. Introduction
Social network users exchange infinite amounts of data every day. The benefits that they get from them are very different, from logging faster into an app to participating in an online contest, data flows freely everywhere (Ellison, 2007). Users know that there are some risks in sharing this information online, but if the utility of the benefits is higher they decide to go forward (Dowd & McElroy, 2007).
The role of risk perception in online behavior is an interesting topic showing that while a higher level of online privacy risk leads to a smaller willingness to disclose information online as well as to a higher level of privacy concerns, a lower level of risk perception leads to a higher trust on the online site and willingness to disclose private information (Dowd & McElroy, 2007). However, these distorted risk perceptions are often irrational as previous research has shown (Ackerman et al. 1999, Sweat 2000). In order to understand better the mental processes that trigger these behaviors, some scholars tried to establish a framework that explained them: The privacy Paradox.
The Privacy Paradox established a framework to understand better the psychology of individuals with respect to their intentions in the disclosure of private information and their actual behavior (Nordberg, Horne & Horne, 2007). Furthermore, research showed that often, users do not behave online on the way that they would be expected to do so when examining their privacy concerns (Kokolakis, 2015). When they go online, users change their privacy principles in the moment they are offered a benefit (Norber, Horne and Horne, 2007). This means that in exchange of certain benefits such as sending messages to our friends or knowing the exact location of our high school ex we adopt a risky behavior in which we disclose a lot of private information that can be used for commercial or other purposes (Kehr & Bentzel, 2014).
Consumption entails giving up on one product at the expense of choosing another one that fills our needs better (Graeber, 2011). Due to the nature of the act of consumption, consumers always accept certain risks when deciding to choose one product instead of another; they face a constant trade off (Yates and Stone, 1992). Being online consumers, social network users also face trade offs and risks when going online and behave sometimes in rather irrational ways (Sawati, 2011). More precisely, when focusing on privacy behavior, the fact that these types of consumers behave in a way that differs from what was expected of them in the first place, suggests that there must be other factors driving this irrational change of behavior and tolerance towards risks (Ackerman et al. 1999, Sweat 2000).
Facebook is a social network that in the words of his creator, Mark Zuckerberg, does not have another purpose than “helping the world to feel more connected”. Currently, almost two millions of people are part of the famous social network (Statista, 2017) and even though the popularity of the social network just seems to be increasing year by year, there is a
growing trend of a higher public concern with respect to what is being done with our data and privacy (Waldman, 2016). Interestingly, privacy concerns are higher for the youth than they are for adults, who often take less action when it comes to keeping their information secured (Grant & Bolsover, 2014).
The Privacy Calculus Theory (Culnang and Armstrong, 1999) has served as a great framework to study private information disclosure in the past. In the context of social
networks, the Privacy Calculus proofs that users who disclose information online are facing a trade off of costs and benefits (Dielin & Metzger 2016). This means that a rational person should only share private information online when the benefits of disclosing this information outweigh the costs of doing so. This theory has brought up different findings of interests when analyzing the behavior of Facebook users. These users behave differently on the social network depending on their concern of privacy and show different patterns of online
ones with a lower privacy concern (Krasnova et al. 2010). Self-withdrawal behavior (limiting one’s own information in a post to other friends on the network) helps us to understand privacy concerns as well, being the users that show this behavior more concerned (Dielin & Metzger 2016).
The gap that this research aims to fill concerns the effect of perceived gratification on risk perception when using a social network such as Facebook. So the research question of this study is aims at answering how the perceived gratification (being the independent variable) effects the perception of risk (dependent variable). This study contributes to the literature in several ways: firstly, it proofs that a link between benefits and risks perception exists; secondly, it also gives insight on the role that liked and disliked items or events have in perceived risks; and last but not least, it brings to the field a deeper understanding of how social network users measure risks and benefits when sharing private data online. Previous research on the Privacy Paradox and the Privacy calculus will serve as a great starting point for doing this project.
2. Literature review
Firstly, in order to understand better the psychology of the consumers when sharing private information online it is important to review some of the core concepts of privacy in the literature. Secondly, it’s important to understand the different theoretical streams on risk taking behavior. Lastly, the literature regarding the context of this study, Facebook, will be analyzed.
2.1 The Privacy Paradox and reasons for private information disclosure
Privacy is a claim by which individuals or groups determine the exact terms on the extent of information they are willing to share with others (Westin, 1967). Moreover, privacy is a basic right, an as a right is normally protected by the law of the State (Sharp, 2013). Scholars observed that people behaved irrationally when sharing private information online and that is how the privacy paradox theory was elaborated. To understand better the Privacy Paradox we need to look back to the year 2001 in which research showed that while online shoppers cared about their online data being misused, they had no problem in giving away information to retailers in exchange for some different benefits (Brown, 2001). So, the Privacy Paradox deals with the fact that people tend to disclose private information at the same time that they claim a real care towards privacy (Nordberg, Horne & Horne, 2007).
The authors discovered that consumers complain that companies sometimes use their information with bad purposes however they usually and freely distribute their own data. Commercial and governmental interest start flourishing with regards to this
information that goes from the private to the public ambit faster than how consumers think. The authors also claim “at the same time consumers voice concerns that their rights to control their privacy are being violated”. People are constantly facing the trade off of disclosing more information versus keeping their privacy intact. From websites requiring a profile registration in order to access content to apps that instantly require you to
The authors’ most surprising finding is that intended behavior differs a lot from real behavior. Furthermore, the authors discovered that even the people claiming to be seriously concerned about privacy were also displaying personal information when required. People reveal private information about themselves only for the instant
gratification and satisfaction of being observed by others and showing their possessions or other values regarding to public valuation (Kokolakis 2015).
As mentioned by the author, social and transactional situations have already been studied in their relation to privacy in the past. The author goes deep into the
current streams defining the privacy paradox nowadays and suggests that even there is a lot of research already done, it is such a complex process that there is a gap for
conducting synthetic studies based on theoretical models.
The author brought up some of the reasons that might explain the privacy paradox:
(1) social-based theory, according to which users forget about the hazards of private information sharing because of the emotional benefits they get from sharing this information with their friends and from being part of a community; another factor are the (2) cognitive biases that social network users have when making decisions online, such as the immediacy of a benefit which lowers risk
perception on the user (Wilson and Valacih, 2012) or the affect heuristic, which changes how much information users disclose depending on how they like or dislike something; (3) bounded rationality and incomplete information of users who cannot calculate completely all of the privacy risks; (4) the quantum theory homomorphism, which incorporates they idea that users affect their privacy statuses indeterminately at the time a decision is made and (5) the Privacy Calculus, a theory developed around the idea that consumers face a trade off of benefits and costs when disclosing private information (Kokolakis 2015).
(2) Apart from these factors previously mentioned, other research suggests that there might be other reasons responsible for the Privacy Paradox: uncertainty, context-dependency and malleability and influence (Acquisiti, Brandimarte and Loewenstein, 2015). Firstly, the authors discuss that some users disclose private information online because there is uncertainty with regards to how an action online is going to impact their privacy. According to the authors, user’s unclear preferences as long with asymmetric and incomplete information are the main forces behind users’ uncertainty.Secondly, the authors found out that users behavior towards privacy depends largely on the context. In other words, peoples’ privacy behavior varies in every situation and it is not constant over time. Thirdly, the authors discuss the malleability of diverse entities to get users to disclose their private information. Default settings as well as websites’ design are some of the mechanisms that firms use to increase information disclosure. The authors also remark that not all malleability is evil and provide the example of monitoring, in their own words: “Monitoring can cause discomfort and reduce productivity, the feeling of being observed and accountable can induce people to engage in prosocial behaviors or (for better or for worse) adhere to social norms”.
Research also proofs that our behavior with regards to online privacy has not been always the same and that it was only at the beginning of the new millennium with the mass access to the Internet that consumers started to become more and more concerned (Hamilton and Cavoukian, 2002) and research suggests that even after trying to
understand all the reasons underlying behind this irrational behavior, complete privacy might just be unattainable (Dinev and Hart, 2007).
Even though research has helped us to understand better the factors behind private information disclosure, we still face the same problem: this tendency of disclosing
privacy is increasing every single day due to all the technological advancements of our era, so it is expected that we will keep encountering problems related to this (Dinev and Hart, 2007). Research has shown that people tend to evaluate the risks and benefits of making a decision (Tversky and Kahneman, 1981). However, the literature about the Privacy Paradox has not deeply studied the relevance of what could be a very important factor: the relationship between perceived benefits and perceived risks.
2.2 Risks and benefits: a psychological perspective
When evaluating risks people tend to use heuristics, which are judgmental rules that help us to reduce the complexity of tasks by turning them into simpler ones (Slovic, Fischhoff & Linchestein, 1980). The authors state that some heuristics are of great relevance for understanding the perception of risks and benefits, such as availability or anchoring.
For example, availability is a heuristic that people use to judge an event as likely if it is easy to imagine (Kahneman & Tversky, 1973). This explains why people perceive the risk of dying on a plane crash as high, because they can actually picture it in their minds even though risks are very low (Slovic, Fischhoff & Linchestein, 1980). The authors also suggest that based on this heuristic, people tend to undervalue the risks that they cannot picture.
Anchoring is another heuristic that has been object of research due to its importance as a reason for engaging in a risk taking behavior (Dowd & McElroy, 2007). As the authors explain, an anchor represents an initial assessment of certain information that we use when evaluating other related information. So in evaluating the risk of certain activity or object, individuals will base their opinions on their own anchors, resulting in differences in the perception of risk for a same situation or object (Tversky and Kahneman, 1981).
Another concept to understand risk-taking behavior is the optimistic bias, which explains that people overvalue the benefits, and underestimate the risks they are facing themselves as compared to others when exposing to risky activities and therefore think that nothing is going to
happen to them (Heather et al.2002). Also, research shows that people are more tolerant towards risks of voluntary situations, such as bungee jumping; than they are towards the situations they have no choice over, such as living in an area with high levels of radiation (Starr, 1969). Human beings share a common way of perceiving risks and benefits: riskier activities are judged as lower in benefits and safer activities are judged as higher in benefits, hence, they are inversely related (Slovic & Alhakami, 1994).
Lastly, the Psychometric Paradigm (Slovic,Fischoff & Linchtenstein, 1980) has been studied in the past to understand better how people weigh the benefits and risks of a decision, which is measurable for them. According to the authors there are several factors that
represent individuals’ perception of risk such as dread or stigma. Authors’ main finding is that there is an unacceptable tolerance towards high risks in our society, they suggest that this is due to the fact that benefits could determine sometimes risk perception. Following this logic, the first hypothesis was formulated:
H1. Perceived benefits have a direct negative relationship with perceived risks.
2.3 The affect heuristic: likeability and risk perception
There is a difference in risk perception depending on the item people are evaluating: if the item being evaluated is liked by people, it will be perceived as less risky and more beneficial than it really is. Likeability has been defined as a construct with different components that range from cognition to affect (Alwitt, 1987). In terms of marketing, likeability is a complex psychological process that starts in the mind of the consumer and that modifies his perception of a brand or a product (Ahuvia, A.C, 2005). The authors explain that consumers have a strong preference to buy those products that come from companies they like and a rejection
towards those products that come from companies they do not like. Likeability is very
important for a firm as it can evolve into the final desired dimension: brand love (Batra et al., 2012). Likeability is strongly related to attitudes and intentions as they are the psychological drivers that initially capture the information of a brand or of a product displayed by the company and then, turn these into a preference or a rejection (Lin et al, 2009). The author states that there is a relationship between likeability and different factors such as price perception, familiarity with the product, or risk perception.
Kahneman and Tversky (1979) became a turning point in the literature of risk perception. They found out that individuals had two different behaviors: they were risk-seekers when confronting losses and risk-averse when facing gains. These studies challenged the conception of humans as rational decision-makers. The authors go further by suggesting that affect has a lot to do in these perceptions. The Affect theory is really important to better understand the role of likeability in risk perception. This heuristic refers to how people perceive the different type of risks they encounter and how they act on them depending on how positive or negative they find them (Slovic, Finucane, Peters, & MacGregor, 2002). In this article it is mentioned that the correlation of risks and benefits is negatively correlated in the mind of people.
Furthermore, there is a negative relationship between perceived risks and perceived benefits (Alhakami and Slovic, 1994). This study reinforces the idea that people judge risks of an event depending on how they feel about it: if the feelings towards that event/activity are positive they will perceive the risk as low and benefits as high. On the other hand, if those feelings arise as negative then risks will be high and benefits will be low. Finucane et al. (2000) confirmed this effect by manipulating the dimension of the benefits presented to subjects in a study. The concept of risk doesn’t have one unique direction: risk is perceived in two different ways, feelings and analysis (Slovic & Peters, 2004). While feelings are related
to the individual perception of danger, the analysis is related to the decision on how to act upon these risks. The authors claim that affect has a fundamental role in behavior and motivation and develop a model to further develop these ideas. They observe that decisions are mostly taken in an intuitive way rather than in a logic way. For evaluating risks,
individuals make decisions based on their guts and not on a logical process of thinking. All together, these findings suggest two things: risk-taking decisions are made in an irrational way and this irrationality is partly explained by the feelings that individuals have towards the item or event they are evaluating. Hence, how much people like something could determine their perception of risks. This resulted in the second hypothesis:
H2. Likeability has a direct negative relationship with perceived risks.
2.4 Facebook as a context for online private behavior
This research focuses in the privacy behavior that people have with regards to social networks, but more precisely towards Facebook. Social networks have gained an increase in importance in the last years (Ellison, 2007). Social networks are “sites as web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system. The nature and nomenclature of these connections may vary from site to site” (Ellison, 2007). So according to the author’s definition, social networks are often a mean for people to expose themselves online and create a profile in which they expose their lives and contact with other peers who they know or who might have the same interests.
There is a relationship between the tendency to get involved in risky behaviors and the tendency to display private information online (Fogel and Nehmad, 2008). The author also notes that there is a difference in gender when it comes to evaluating the risks and the
careful and concerned for their private information and have a lesser need to show
themselves off. In fact, some women are so concerned about keeping an image on the social networks that they filter every content that is posted about them carefully so that people do not think anything about them. Men, on the other hand often show a careless behavior and post the first thing that comes to their minds without thinking too much about it. The authors also showed that there are differences in the way people behave depending on the site they are using. They researched on the difference of trust displayed between different social networks, in this case Facebook and Myspace. They concluded that Facebook users trusted more the site than Myspace’ users and that this is probably due to the fact that Facebook was only used by students in the beginning, contrary to Myspace which was available for every single internet user. This suggests that this irrational behavior explained on The Privacy Paradox varies between genres and platforms and this will affect the relationship between risks and benefits.
Facebook is one of the most interesting social networks for conducting research for two reasons: mostly everyone has it and so it is a social phenomenon, and secondly because it offers a way of studying the privacy patterns of
information revelation in our population (Acquisiti and Gross, 2006). The authors themselves came to some conclusion with regards to this social network’s use. They tried to decipher those differences in the demographics of Facebook users versus Facebook non-users and the common behaviors that members of a group had. They looked for motivators of the members of these communities for
explaining their behaviors and they found that users of both groups often failed at understanding Facebook’s disclosures of information policies and they often got involved into risky behaviors without knowing the consequences. The authors also found that the concern on Facebook privacy is increasing exponentially in the last
years as consumers are more aware of what the social network does with their information.
Social networks users modify their behavior online and they tend to be more open to disclose private information than what you would normally expect from their words
(Stutzman & Vitak. 2012). The authors introduced the concept of social capital “Social capital is the total actual or potential resources individuals have access to through their social network (Bourdieu 1986). Social capital includes physical (e.g., driving a friend to the
airport), emotional (e.g., giving a friend a hug), and informational (e.g., giving a friend advice about a big decision) resources, among others.” Thus, social capital can be one of the
explanations for this irrational behavior in Facebook users. They outweigh the benefits more than the costs of disclosing this information.
Further research on privacy concerns of people when using Facebook suggests that the behavior towards privacy also depends on whether it affects you directly or not (Horn and Huges, 2009). The authors found that users who see their privacy violated tend to care more about their privacy settings, as opposite to the users who never had a problem who showed a low care on privacy concerns. Therefore, this behavior is
expected to change from user to user depending on their experience. Lastly, the authors also found that the gratifications of using Facebook weights more than the cost of using it. In this sense, people love the social network and enjoy giving away private
2.5 Research question and research gap
This study aimed at providing insights for answering the following research question: how
perceived gratification affects risk perception? As it has been examined on the previous
paragraphs and thanks to the concept of The Privacy Paradox, people display unexpected patters of private behavior incoherent to their own logic. The online world is a perfect ecosystem to analyze this effect as privacy decisions are taken all the time regarding the amount of data that we are willing to share online. In order to be able to examine this effect in a more proper way, a social network was picked as a set-up for an experiment.
Specifically, using Facebook as a context, the purpose was to understand better the role of perceived benefits and perceived risks of disclosing private information. A gap was identified on the literature regarding if perceived benefits and perceived risks are factors independent of each other. Therefore, this study aims to prove that perceived benefits and perceived risks are not independent factors but that actually exists a relationship between them.
2.6 Conceptual model
The goal of this study is to expand the literature on the topic of the benefits and risks of disclosing private information online. By using the Privacy Paradox as a background and choosing Facebook as the as a context for the study, the specific aim is to measure how perceived gratification (as the independent variable) effects the perceived risk (dependent variable) when it comes to disclosing private information online. Likeability is present on the model and acts as another factor. After reviewing the literature on likeability and affect theory, a higher likeability should lead to a lower perception of risks and vice versa.
Figure 1. Conceptual model
The relationships between both factors and perceived risks are negative, meaning that the variables are negative correlated. Firstly, as the level of perceived benefits increases, the level of perceived risks is expected to decrease. And secondly, as the level of likeability increases the level of perceived risks is also expected to decrease.
Likeability
Perceived risks
H1
H2
3. Method
The following chapters will focus on the empirical aspects of this research. They start by explaining the characteristics of the sample that was used for the study, following by the design of the study, then the selected variables that correspond to the hypothesis mentioned earlier. Lastly, the statistical methods used to analyze the data will be explained.
3.1 Sample
Data was collected via an experiment with the online tool Qualtrics. Participants were reached in different ways: Facebook, direct contact, social media share and lastly, the online platform SurveyCircle. In order to encourage people to answer, the end of the survey added an Amazon token of 25 euros.
289 participants responded to the survey, from which 212 completely finished it (response rate 71%). The questionnaire was created in English language with the purpose of having a heterogeneous and rich sample of people all around the world; most respondents had the Spanish nationality. No limits were established on the sample in terms of gender or socioeconomic status, but due to the nature of the experiment all the respondents who did not have the social network Facebook (6.7%) were moved to the end of the survey after
completing the first question (“Do you have Facebook?”) and discarded for the study
afterwards. As expected, the sample of respondents who had the social network Facebook was high (93.3%). Lastly, 185 respondents were analyzed as the other 27 had an extreme
percentage of missing answers on the experiment.
The amount of time spent of Facebook by the respondents was also relevant, as more expert users should know the platform better. The time spent was between 1-3 hours per week (27%), so it is assumed that this group of people corresponds to expert Facebook users since they spend a considerable amount of time on the platform. The age of the respondents ranged from 15 to 70 years old, being the mean 25 years old. This makes sense considering that
heavier social media users tend to be young adults in their early and late 20s (Greenwood & Perrin, 2016).
3.2 Design
Regarding the experiment, 4 different scenarios were established and respondents were assigned to each of these in a random way with the help of the randomization function of Qualtrics. A 2 (benefits: high vs. low) x2 (likeability: liked vs. disliked) between subjects experiment design was used. Two effects were measured: the effect of perceived benefits on perceived risks on the one hand and the effect of likeability on perceived risks. The survey consisted on three different sections which respondents had to answer: a section for the experiment, another section for specific questions regarding risk taking behavior online and a last part about demographics. The variables measured were perceived benefits, perceived risks and likeability.
Perceived benefits
In the current times people constantly make decisions on whether display information online or not, an in doing so they measure for each occasion the benefits of sharing this information with their friends and families (Zainab & Masrom, 2015). According to the authors,
perceived benefits are subjective and vary from person to person.
These benefits can be of any type, from more likes, to increased self-steem or to happiness. In this study, perceived benefits acted as the independent variable that was
manipulated to observe the change in the perceived risks. Fischoff (2005) established that risk became more acceptable and perceived benefits increased (r = .58), signaling a division of the effect of benefits in risks depending on how high or low they are. Therefore a dummy
variable was created to represent the perceived benefits. It had values of “0” (High benefits) and “1” (Low benefits).
Likeability
According to affect theory, whether people like or dislike something has an effect on their behavior (Slovic et al. 2005). As suggested by the authors, liked items are perceived as less risky. In this sense, likeability acted as the the second factor of the model of the model and decrypting its effect on risks and benefits was of great importance. Using previous research as a reference (Lin et al, 2009), a dummy variable was created to represent the perceived
benefits. It had values of “0” (Liked company) and “1” (Unliked company).
Perceived risks
Perceived risks are the other half of the decision making process of taking any sort of action, such as disclosing private information online or deciding on whether to buy on an online business or not (Zafat et al. 2015). Companies are optimizing their online processes in order to have access to this information and can use it in a harmful or commercial way (Jones & Hiram Soltren, 2005). In this case the perceived risks were related to exposing private information publically on Facebook and they would act as the dependent variable of the study. The variable was measured by using a likert scale (1= Strongly disagree, 7= Strongly agree). These items were based in the research by Webber and Betz (2002). The questions related to risk were very specific such as “the risks do not look important to me” or “risks of data distribution are low”. Most of the questions had a positive statement that diminished risk, meaning high values of this variable represent a low perception of risks, and low values of the variable represent a higher perception of them. These different scenarios corresponded to the hypothesis mentioned earlier and to measure these effects the experiment consisted on the following idea.
Runtastic is an app that tracks your physical activity such as running or biking, it gives the option to register via Facebook (Option 1) or via a questionnaire (Option 2) that takes more time and it is less convenient. The questionnaire was presented as a less risky option
because respondents did not need to give as much private information of their Facebook profiles as in the other option. However, the Facebook option included different benefits regarding time, money saving and a chance to enter a contest and earn a last generation smartphone. After this, the experiment asked the participant to consider that Runtastic is part of their most loved or most disliked company.
Table 1 represents the four scenarios created for the experiment: Scenario 1. High benefits, liked company
• Time saved: 40 hours per year • Free yearly subscription • Contest: 1/10
• Imagine liked company
Scenario 2. High benefits, disliked company • Time saved: 40 hours per year • Free yearly subscription • Contest: 1/10
• Imagine disliked company Scenario 3. Low benefits, loved company
• Time saved: 20 minutes per year • 95 cents yearly subscription • Contest: 1/1000000
• Imagine liked company
Scenario 4. Low benefits, disliked company • Time saved: 20 minutes per year • 95 cents yearly subscription • Contest: 1/1000000
• Imagine disliked company
Table 1: The four scenarios
Scenario 1. High benefits, liked company. (Appendix 1)
The benefits of signing up via Facebook were really high in this case. Total time saved per year was of 40 hours. Financially, it was also more attractive offering a free yearly
subscription with Option 1 versus 50 euro payment with Option 2. The chance of winning an smartphone was of 1/10. In this case Runtastic was part of a liked company of the respondent.
Scenario 2. High benefits, disliked company. (Appendix 2)
The benefits of signing up via Facebook were also high in this case. The information regarding the benefits (money, time and contest) was replicated. However, in this case
Runtastic was part of an disliked company of the respondent.
Scenario 3. Low benefits, loved company. (Appendix 3)
The benefits of signing up via Facebook were low in this case. Total time saved per year was of only 20 minutes per year. Financially, it was less attractive than in the previous two scenarios by only offering a 95 cents yearly subscription with Option 1 versus a 99-cents payment with Option 2. The chance of winning an smartphone on the contest was of 1/1000000. In this case Runtastic was part of a loved company of the respondent.
Scenario 4. Low benefits, disliked company. (Appendix 4)
The benefits of signing up via Facebook were also low in this case. The information regarding the benefits (money, time and contest) was replicated. In this case Runtastic was part of a disliked company of the respondent.
To ensure the reliability of the data, two check questions regarding the benefits of signing up on Runtastic via Facebook were added right after the experiment. Apart from this, a timer was placed on every of the 4 experiments so that respondents could not move to the following page without spending at least 20 seconds reading the
experiment.
3.3 Statistical method
In the first place, counter indicative items were searched and corrected if needed. These are items that have an inverse order in answers and must be corrected on SPSS in order to have an accurate study. Secondly, a reliability test was executed to if the questions of the
questionnaire related to the dependent variable were reliable. Finally, Factorial Anova was the main tool to analyze the different effects between the variables.
3.4 Pretest
Prior to the experiment a pre test was carried out with the purpose of confirming an existing relationship between perceived benefits, likeability and perceived risks. Without this pretest
and this relationship unconfirmed, the experiment would have not made so much sense. For this pretest, 30 respondents participated and completed the experiment. They were introduced to the same scenarios and type of questions as in the real phase. The main objective was to analyze whether perceived benefits affect perceived risks and how that is affected by likeability.
Four different scenarios were presented to participants: high-benefits and liked company, high-benefits and disliked company, benefits and disliked company and low-benefits and liked company. Participants were presented randomly with one of these four scenarios and they answered the different questions related to the risks of sharing their private data on Facebook. First, they were presented with different types of benefits from signing up into the app Runtastic via Facebook, and then they were asked to imagine that Runtastic had a partnership with either their most liked or disliked company. By doing this, the
above-mentioned relationships could be measured and established a starting point for the whole study.
The results from the pre-test initially confirmed that it existed a relationship between risks and benefits and that likeability contributed to this effect. The method of Factorial Anova was used for the pre test.There was a significant main effect of perceived benefits on perceived risks F(1, 30) = 3.04, p < .05, η² = .03. Also, there was a significant main effect of likeability on levels of perceived risks F (1, 30) = 13.07, p < .05, η² = .05. Finally, there was a non-significant interaction effect between likeability and perceived benefits on risk perception F(1, 30) = 1,03 p =.3, η² = 0.
Besides the main effect, it was interesting to also know at early stages of the study whether higher benefits decreased the perception of risks more than lower benefits. The results show that people in the higher benefits groups (M=5.5, SD=1.2) perceived risks lower than in the low benefits groups (M=4.2, SD=1.2) so this negative
correlation between benefits and risks was confirmed. Regarding likeability, liked companies (M=3.4, SD=1.3) represented a more important decrease in risk perception than disliked companies (M=3.0, SD=1.3).The results of the pretest showed that the higher the benefits, the lower the risks perceived and that for liked companies a lower level of perceived risk was shown by participants.
4. Results
4.1 Descriptive statistics
Firstly, in order to find out more about the risk perception of the respondents regarding their general views on risks and benefits of sharing private information online a few questions were asked to go check the validity of the results and to generate a complete and relevant
discussion.
Descriptive statistics were used to measure the extent of these effects as it can be observed on the Appendix 5.Respondents stated in the first place that that the benefits of having social networks outweighs the risks of having them (M=4.15, SD=1.5) and the importance of signing up to apps via Facebook was also high, so it was concluded that the benefits are really important when analyzing the risks of logging in online. Regarding likeability, respondents said that what they like feels safer than what they do not like
(M=4.49, SD=1.4) and that sharing information with friends feels safer, (M=4.87, SD=1.5). However when asked about risks, respondents saw both risks in having social
networks (M=2.86, SD=1.4) and in posting in Facebook (M=3.09, SD=1.6). Note that these were counter indicative items and low numbers represent a high value of risk perception. Besides providing valuable information, these two questions were included in the experiment with the purpose of checking the accuracy of the data. Higher values on these questions would have meant that respondents were answering all questions in a positive way, without actually reading the questions. Finally, respondents showed a high privacy concern with regards to the risks of sharing private data online (M=5.85, SD=1.1) and they also claimed that they really care about what firms do with their data (M=5.54, SD=1.5).
Facebook was seen as an unsafe place to post personal information and respondents showed a high concern of what information about them is online on the social network (M=3.05, SD=1.4).
Reliability test was carried out with the purpose of verifying that the scale used was reliable. The perceived risks scale has a high reliability, with Cronbach’s Alpha = .954. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale (all above .30). If any of the items were deleted, none of them would affect reliability.
As seen on Table 2 the risk represented by every variable was different on every scenario: Risk Scenario A (M=4.12, SD=.2), Risk Scenario B ( M=3,58 , SD=.18), Risks Scenario C (M=3.61 , SD=.16) and Risks Scenario D (M=3.17 , SD=.16).These initial differences on the mean already show a difference between the scenarios with a liked company (A&C) and scenarios with a hated company (B&D).
Perceived risks
N Minimum Maximum Mean Deviation Variance Std. Statistic Statistic Statistic Statistic
Std.
Error Statistic Statistic RiskTOTA 43 1,18 6,91 4,1266 0,20900 1,37048 1,878 RiskTOTB 41 1,18 6,18 3,5835 0,18726 1,19906 1,438 RiskTOTC 58 1,00 6,45 3,6108 0,16005 1,21890 1,486 RiskTOTD 48 1,00 5,36 3,1769 0,16724 1,15870 1,343 Valid N (listwise) 0
4.2 Factorial Anova
As it can be seen on the tables below, several Factorial Anova analyses were performed in order to understand the effect of each of the factors (perceived benefits and
likeability) on the dependent variable (perceived risks).
PerceivedRisks * Likeability
PerRisks
Likeability Mean N
Std.
Deviation Minimum Maximum Variance
% of Total N Liked company 3,4303 70 1,35715 1,00 7,00 1,842 49,6% Disliked company 2,8955 71 1,25251 1,00 6,00 1,569 50,4% Total 3,1610 141 1,32823 1,00 7,00 1,764 100,0%
Table 3. Likeability means
PerRisks * PbX
PerRisks
PbX Mean N Std. Deviation Minimum Maximum Variance % of Total N High benefits 3,4086 64 1,42504 1,00 7,00 2,031 45,4% Low benefits 2,9553 77 1,21323 1,00 6,71 1,472 54,6% Total 3,1610 141 1,32823 1,00 7,00 1,764 100,0%
Table 4. Perceived benefits means
Depending on the correct or incorrect answer from the respondents to the check questions, three different results section are presented:
First division of results refers to all the respondents who completed the survey. Whether these subjects answered correctly or incorrectly the check questions was not taken into account. The output is attached on the Appendix 1. Factorial Anova provided the following results: there was a significant main effect of perceived benefits on perceived risks F(1, 185) = 7.01, p < .05, η² = .04. Figure 2 shows that the relationship between perceived benefits and perceived risks is negative. As mentioned earlier, low
values for the variable Risk represented a higher risk perception and higher values represented a lower risk perception.
Figure 2 Effect of perceived benefits on perceived risks
Secondly, there was a significant main effect of likeability on levels of perceived risks F (1, 185) = 7,49, p < .05, η² = .04. The interaction effect between likeability and
perceived benefits on risk perception was not significant F(1, 185) = 0, p =.93, η² = 0. As Figure 3 shows this relationship is again negative.
Figure 3 Effect of likeability on perceived risks
Next part of the results refers to those respondents who completed the check questions in a positive way. The four groups of the experiment were significantly reduced (N= 113). The analysis was performed in the same exact way ending up with the following
results.There was a significant main effect of likeability on levels of perceived risks F (1, 119) = 12.20, p < .05, η² = .07. Surprisingly, the effect of perceived benefits on risks was not significant with p>.05. Again, there was a non-significant interaction effect between likeability and perceived benefits on risk perception F(1, 113) = 0,4, p =.51, η² = 0, however in this case there was a higher relationship between the variables.
Lastly, a third scenario includes those who passed one of the check questions. As expected, this group was of a size in between the two previous groups (N=137). Once again, there was a significant main effect of likeability on levels of perceived risks F (1, 137) = 5.94, p < .05, η² = .04. In this case, there was also an effect of perceived benefits on perceived risks F (1, 137) = 4.50, p < .05, η² = .04 Perceived benefits had an effect on perceived risks on 2/3 scenarios. This means that perceive benefits has an effect on perceived risks and therefore we can accept H1. Surprisingly, the effect was not significant on the group of respondents who answered correctly both check questions.
Overall, an effect of likeability on perceived risks was confirmed. In the study people whose experiment involved a company they liked were willing to disclose more private information online, and the people whose experiment involved a company that they hated were less willing to disclose private information online. This effect is the strongest in the second analysis, were respondents answered right both check questions. Therefore, we accept
H2 and conclude that likeability has an effect on the relationship between perceived risks and
perceived benefits.
Once the general result for the study was concluded it was important to be aware of the differences between the groups. Therefore, following the general analysis, a mean test was made to analyze the differences between the means of both subgroups of the two independent variables: likeability and perceived risks.
Regarding Likeability, on Table 2 it can be seen that liked companies had a higher mean (M=3.4, SD=1.3) than disliked companies (M=2.8, SD=1.2) confirming that the more people like something the less risks they perceive. Subjects on the scenarios 1 and 3, in which a liked company came into play, tended to adopt riskier attitudes than subjects of the
scenarios 2 and 4, in which a disliked company came into play.
The effect for perceived benefits was similar as it can be seen on Table 3. The effect was also confirmed in this case. People in the high benefit conditions perceived risks as being lower (M=3.4, SD=1.4) than people in the low benefits condition (M=2.9, SD=1.2). Subjects in the scenarios 1 and 2, in which the benefits were more attractive (a free yearly subscription, a big chance to win an smartphone and 40 hours of time saved per year), showed a lower perception of risks than subjects on the scenarios 3 and 4, in which the benefits were less attractive (paid subscription, a small chance of winning an smartphone and 20 minutes saved per year).
These results together explained that both perceived benefits and likeability have a direct effect on perceived risks. However this effect was not the same in the categories of the two independent variables, being the individuals who liked the company that Runtastic had a partnership with and the individuals that were presented with the higher benefits, the ones who perceived the risks of sharing information as lower.
5. Discussion
5.1 Theoretical and practical implications
The purpose of this whole study was to answer the question: how perceived gratification affects risk perception? In previous literature a relationship between risks and benefits had been observed but it had not been clearly explained. This study adds a new component to the discussion: the existence of a negative relationship between perceived risks and perceived benefits.
The setting of this research is the moment in which a person decides whether to sign up or not in a sports app called Runtastic by providing his Facebook data. At this point, the user is expected to think of the risks and benefits of sharing this private data online. The study starts with the assumption that people care about their privacy but for some reason show a contradictory behavior when it comes to their actions (Nordberg, Horne & Horne, 2007). Based on this, the study put this assumption into test with the objective of measuring how social network users behave with respect to their privacy online by assuming the risks and the benefits. A second factor, likeability comes into play as a determinant of risk assessment following the core concepts of the Affect Heuristic as examined by Daniel Khaneman.
On the study few different conclusions were shown. Firstly, people behave in an irrational way when it comes to privacy online, as subjects of the study had no problem in sharing their data online when it comes to obtaining the benefits while claiming that they really care about what companies do with it and how the online network is an unsafe invornment.
Secondly, how an individual perceives risks varies from subject to subject, but, thanks to the study, a tendency to perceive certain risks as lower when the benefits of one action or event are higher came into light. As respondents were offered benefits such a free use of the app for
attitude with their actions online. It appears that people become risk-seekers when benefits are significantly high and risk-averse when these are not as important.
Thirdly, people also asses their risks depending on how they feel about them, meaning that the more they like something the less risky it looks to them and the less they like it the more risky it seems. This logic further supports the literature on affect theory and risk perception discussed earlier in the paper. Specifically, on this study, individuals showed a changing risk perception when Runtastic had a partnership with their favorite company versus when this partnership was with a company they disliked. So these positive feelings triggered a decrease in the worrying process of the subjects towards the possible problems of being open about you online.
As results showed, both perceived risks and likeability have a strong and relevant effect on risk perception. The relationship between perceived benefits and perceived risks is negatively correlated, meaning that the more benefits a person sees in an item, event or action the less risks associated to them this person will tend to perceive. However, the less benefits this person sees the higher he will perceive the risks. Apart from this, the more a person likes this event, item or action the less risks he will perceive meaning that disliked events or items might be perceived as less risky when compared to like ones, even though the risks might just be the same from an objective point of view. This study expands the literature on risks and benefits by providing with a new view on the direct influence of one variable over the other. Last but not least, the Privacy Paradox is further confirmed by observing irrational behaviors and incoherence in the subjects of the study.
The study has implications both from a theoretical and a managerial point of view. For theorists it might be important to revise the relationship between risks and benefits at a deeper level. Previous literature suggested that this relationship was non-existent and that the
this relationship. Besides this, the direct negative effect of likeability on perceive benefits not only the main ideas of the Affect Theory, but also puts them into practice at a deeper level by establishing a clear behavioral effect on people facing risky situations with items that they like.
From a managerial point of view, the study adds different points in which they could rely on for future business practices. Regarding the studied factor perceived benefits some conclusions are of special importance: consumers might be more willing to share private data to companies when the benefits that they see in exchange are more relevant to them in the first place. This could translate into always looking to stress the benefits of the product and always make them present in packaging, advertising and such. A consumer is always facing tradeoffs when doing a purchase, advantages and disadvantages (benefits and risks) come to his mind, so by letting them know how wonderful the product is and being really specific of the features, consumers might forget of the risks of the purchase (such as money wasted, remorse and so on).
Also the conclusions regarding the other factor studied, likeability, could be useful for companies. From the results of the study, it was clear that Facebook users saw less risk in sharing their data online when a liked company was involved. Consumers lose their rationality in assessing risks when they have positive emotions towards that company.
Therefore, companies should aim at generating as much brand love as possible and to work in their reputation, as this can distort the reality for consumers making them believe that
something is less risky. Following this line of thought, a company that both puts effort on having a good brand reputation and that knows how to offer attractive features (perceived as benefits) to consumers will be perceived as a safer option to buy that competitors.
Besides the above-mentioned implications, the study also focused in a specific
as it was seen on the study, most of the time they are not aware of what companies are making with their information. Even though this is extremely beneficial for companies pursuing as much consumer insights as possible, maybe the focus should change. Companies could be more transparent and specific about what data they are using, how they are using it, when they are obtaining it and so on. Not only this would lead to fairer world, but could also make huge contributions to generate a positive brand image.
5.2 Benefits and limitations
The design of this study was an experiment, as it seemed like the logical way to study how the changing conditions of an independent variable affect a dependent variable. By creating four different scenarios conclusions could be drawn into how perceived risks are affected by perceived benefits and how this relationship is moderated by likeability. If the methodology had been just a survey, it would have been inconclusive, as no modifications of the
independent variable could have been tested. Developing these four scenarios in which respondents had to put themselves into the real situation of disclosing private information online in exchange of certain benefits was an accurate way to measure how people perceive these risks and how they respond to them.
The structure of the experiment provides a solid vision on behavioral patterns in risk taking situations. For purposes of the study, all the respondents who did not have the social network Facebook were automatically expelled from the experiment, which benefits the accuracy of the results at the expense of having to search deeper for respondents.
Having a sample with different ranges of ages (16-70) and of more than 15 different countries makes the results of the study valid at an international level so these effects can be applied to very different cultural contexts.
However, the study might have been affected by some determined bias. Firstly, the study might have suffered from the acquiescence bias, as a few answers to the experiment
were extremely positive and respondents seemed to agree with every situation presented to them. Secondly, some people answered the questions related to the experiment in less than 3 minutes, being the average time for a complete answer 10 minutes. Given this information, it is likely that the study suffered from habituation bias and respondents tried to be done with the experiment as soon as possible.
Thirdly, a selection bias might be present. Respondents were aware that they were answering an experiment as it was mentioned at the start of it. This hypothetical situation could have affected their answers if they knew that the information they had regarding the benefits and risks of displaying information online could belong to a positive or to a negative scenario. Fourthly, the research setting involved the application Runtastic which is very famous between the people who practice sports but not as famous for the people who have more of a sedentary life. Interest from the people who practice sports is expected but not so much from the ones who don’t. Non-sportive respondents might have respond in an
uninterested way harming the accuracy of the data. However, this seemed like a valid option to measure risk perception, as the sign up process of every app is practically the same.
Last but not least, it is important to mention the unexpected result of the impact of perceived benefits on perceived risks, being not significant when respondents answered both check questions correctly. The check questions were present in the study with the sole purpose of observing if respondents who theoretically had payed more attention to the experiment, responded differently that the ones who answered the experiment fast.
Surprisingly, the effect of perceived benefits was not significant when respondents were more focused on the experiment. So even the results measured these effects as significant in two scenarios, there is one where it was different and this represents a limitation of the
5.3 Future research
Based on the research from this study, further analysis could be done on the exact
psychological factors that drive this behavior, irrational from a classical standard economics point of view. How is it possible that a real and objective risk is perceived differently
depending on which benefits are associated to it? Without doubt, a relevant discussion could be set on. Also, the strength of this relationship could also be measured, is this relationship infinite? It is also still not clear if a threshold exists, meaning that up to a certain level of benefits people will accept certain risks but not beyond that level. This could be texted by doing an experiment with more scenarios: more levels of the variable perceived benefits that would cover a bigger range of options with the purpose of establishing this threshold.
Also, the study could be replicated at a bigger scale: with a higher budget and a bigger amount of time better more accurate results could be drawn. Counting with a sample of 298 respondents from which a few of them did not complete the questionnaire, this effect might be a bit short and more subjects in the sample would make the conclusions more applicable. Lastly, a new method of research could be a good idea since the experiment has been carried out online and that could have led to biases and errors that in a more controlled setting would never had happened.
References
Acquisti A. Privacy in electronic commerce and the economics of immediate gratification. Proc. ACM Electronic Commerce Conf. (EC '04) (2004) (ACM Press, New York) 21–29 Acquisti, Alessandro & Brandimarte, Laura & Loewenstein, George. (2015). Privacy and human behavior in the age of information. Science (New York, N.Y.). 347. 509-14. Acquisti, Alessandro & Gross, Ralph. (2006). Imagined Communities: Awareness,
Information Sharing, and Privacy on the Facebook. Lecture Notes in Computer Science. 4258. 36-58.
Adler, P., & Kwon, S. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27(1), 17–40.
Ahuvia, Aaron. (2005). Beyond the Extended Self: Loved Objects and Consumers' Identity Narratives. Journal of Consumer Research. 32. 171-184.
Alwitt, Linda & R. Prabhaker, Paul. (1970). Functional and Belief Dimensions of Attitudes to Television Advertising. Journal of Advertising Research. 32. 30-42.
Aral, S. (2004). Sexual risk behaviour and infection: epidemiological considerations. Sexually Transmitted Infections, 80(Suppl 2), ii8–ii12.
Batra, Rajeev & Ahuvia, Aaron & Bagozzi, Richard. (2012). Brand Love. Journal of Marketing. 76. 1-16.
Bourdieu, P. (1986). The forms of capital. In: Richardson, J., Handbook of Theory and Research for the Sociology of Education. Westport, CT: Greenwood: 241–58.
Boyd, D. M, Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13(1), article 11.
Cavoukian, A., & Hamilton, T. J. (2002). The privacy payoff: How successful businesses build customer trust. Toronto: McGraw-Hill Ryerson. Chicago (Author-Date, 15th ed.) Daniel Kahneman and Amos Tversky (2013) Choices, Values, and Frames. Handbook of the Fundamentals of Financial Decision Making: pp. 269-278.
Debatin, Bernhard & Lovejoy, Jennette & Horn M.A, Ann-Kathrin & N. Hughes, Brittany & Horn, Ann‐Kathrin. (2009). Facebook and Online Privacy: Attitudes, Behaviors, and
Unintended Consequences. Journal of Computer‐Mediated Communication. 15. 83 - 108. Dinev, Tamara & J. Hart, Paul. (2004). Internet Privacy Concerns and their Antecedents— Measurement Validity and a Regression Model. Behaviour & IT. 23. 413-422.
Finucane, Melissa & Alhakami, Ali & Slovic, Paul & M. Johnson, Stephen. (2000). The Affect Heuristic in Judgments of Risks and Benefits. Journal of Behavioral Decision Making
Fogel, Joshua & Nehmad, Elham. (2009). Internet social network communities: Risk taking, trust, and privacy concerns. Computers in Human Behavior. 25. 153-160.
Graeber, David (2011) Consumption. Current Anthropology, 52 (4). pp. 489-511. ISSN 0011- Jones, Harvey & Soltren, Hiram. (2005). Facebook: Threats to Privacy.
Kahneman, Daniel & N. Tversky, Amos. (1979). Prospect Theory: An Analysis of Decision Under Risk. Econometrica. 47. 263-91.
Kehr, Flavius & Kowatsch, Tobias & Wentzel, Daniel & Fleisch, Elgar. (2015). Blissfully Ignorant: The Effects of General Privacy Concerns, General Institutional Trust, and Affect in the Privacy Calculus. Information Systems Journal. 25.
Kokolakis, S.: Privacy attitudes and privacy behaviour: a review of current research on the privacy paradox phenomenon. Comput. Secur. 2011, 1–29 (2015)
Kokolakis, Spyros. (2015). Privacy attitudes and privacy behaviour: A review of current research on the privacy paradox phenomenon. Computers & Security.
Larsson, Anders. (2017). The News User on Social Media: A comparative study of interacting with media organizations on Facebook and Instagram. Journalism Studies. 10.
McElroy, T. & Dowd, K. (2007). Susceptibility to anchoring effects: How openness-to-experience influences responses to anchoring cues. Judgment and Decision Making, 2(1): 48-53. (Feb 2007) Published by the Society for Judgment and Decision Making (ISSN: 1930-1970)
Norberg, P.A., Horne, D.R., Horne, D.A.: The privacy paradox: personal information disclosure intentions versus behaviors. J. Consum. Aff. 41, 100–126 (2007)
Siddiq Alhakami, Ali & Slovic, Paul. (1995). A Psychological Study of the Inverse Relationship Between Perceived Risk and Perceived Benefit. Risk analysis : an official publication of the Society for Risk Analysis. 14. 1085-96.
Slovic, P., Peters, E., Finucane, M. L., & MacGregor, D. G. (2005). Affect, risk, and decision making. Health Psychology, 24(4, Suppl), S35-S40.
Slovic, Paul & Finucane, Melissa & Peters, Ellen & MacGregor, Donald. (2007). The Affect Heuristic. European Journal of Operational Research. 177. 1333-1352.
Slovic, Paul & Fischhoff, Baruch & Lichtenstein, Sarah. (2005). Facts and Fears: Understanding Perceived Risk. Policy and Practice in Health and Safety.
Smith, H & Dinev, Tamara & Xu, Heng. (2011). Information Privacy Research: An Interdisciplinary Review. MIS Quarterly. 35.
Starr, Chancey. (1969). Social Benefit Versus Technological Risk. Science (New York, N.Y.). 165. 1232-8.
Stutzman, Fred & Vitak, Jessica & Ellison, Nicole & Gray, Rebecca & Lampe, Cliff. (2012). Privacy in Interaction: Exploring Disclosure and Social Capital in Facebook.
Trepte, Sabine & Reinecke, Leonard & B. Ellison, Nicole & Quiring, Oliver & Yao, Mike & Ziegele, Marc. (2017). A Cross-Cultural Perspective on the Privacy Calculus. Social Media + Society. 3.
Tversky, Amos & Kahneman, Daniel. (1973). Availability: A Heuristic for Judging Frequency and Probability. Cognitive Psychology. 5.
Tversky, Amos & Kahneman, Daniel. (1981). The Framing of Decisions and The Psychology of Choice. Science (New York, N.Y.). 211. 453-8.
Västfjäll, Daniel & Peters, Ellen & Slovic, Paul. (2014). The affect heuristic, mortality salience, and risk: Domain-specific effects of a natural disaster on risk-benefit perception. Scandinavian Journal of Psychology. 55
Waldman, Ari Ezra (2016) Privacy, Sharing, and Trust: The Facebook Study, 67 Case W. Res. L. Rev. 193
Wilson, D.W. & Valacich, Joseph. (2012). Unpacking the Privacy Paradox: Irrational Decision-Making within the Privacy Calculus. International Conference on Information Systems, ICIS 2012. 5.
Wilson, D.W. & Valacich, Joseph. (2012). Unpacking the Privacy Paradox: Irrational Decision-Making within the Privacy Calculus. International Conference on Information Systems, ICIS 2012. 5. 4152-4162.
Yates, J. F., & Stone, E. R. (1992). The risk construct. In J. F. Yates (Ed.), Wiley series in human performance and cognition. Risk-taking behavior (pp. 1-25). Oxford, England: John Wiley.
Appendix
Appendix 1 – Factorial anova tables
All respondents
Both check questions correct
Appendix 2
Appendix 3
Appendix 4
Appendix 5
Appendix 6
Descriptive statistics on privacy behavior questions
Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
- The things I like feel
safe. 138 1 7 4,49 1,461
- The benefits of having social networks outweighs the risks.
138 1 7 4,15 1,552
- I see no risks in having social networks.
139 1 7 2,86 1,492
- I see no risks in posting on Facebook.
140 1 7 3,09 1,669
- I care about my privacy online.
140 1 7 5,85 1,199
- I enjoy signing up in apps faster thanks to Facebook.
140 1 7 4,14 1,955
- I care about what firms
might do with my data. 140 1 7 5,54 1,500 - I am aware of how
firms use my private data. 139 1 7 4,17 1,817 - I feel it is less risky to
share private data with people I know.
140 1 7 4,87 1,526
- I specifically know the risks of sharing private data online.
140 1 7 4,50 1,677
- I like others to know what I am doing.
140 1 7 3,00 1,738
- Facebook feels like a safe environment to post.
140 1 6 3,05 1,476
