The Privacy Paradox: a Construal Level Theory
perspective
Author: Marit Werksma
Student number: 10674926
Supervisor: J. Demmers 2nd Supervisor: Dr. A. Zerres
Submission: 29 June 2015, final version
Program: MSc. In Business Administration – Marketing Track Institution: University of Amsterdam
STATEMENT OF ORIGINALITY
This document is written by student Marit Werksma who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
TABLE OF CONTENTS
1. Introduction………...………...6
2. Theoretical framework………..……….………..9
2.1 Privacy Concerns and Privacy Behaviors ..………...9
2.2 Psychological Distance……….…….………...….……10
2.3 Temporal & hypothetical distance………..……...12
2.4 Construal level of benefits and costs………...14
2.5 Values and feasibility concerns…….……….…..….15
2.6 Self-conceptions & Implications……….………...16
3. Overview of the present studies………...…..20
4. Study I ……….…..21
4.1 Method……...………....21
4.1.1 Participants………...……….…………..…....21
4.1.2 Design…………..………..………...….…….……...21
4.1.3 Stimuli and measurements………..…………..……...……….…...…..21
4.1.4 Procedure….………..………….……….……....23 4.1.5 Data analysis………..……...23 4.2 Results……...………..………...……...…...24 4.2.1 Psychological distance...…...……….…………..…....24 4.2.2 Construal Level....………..………...….…….……...27 4.3 Discussion……...………...………...29 5. Study II ……….………...………….….……31 5.1 Method……...………....31 5.1.1 Participants………...……….…………..…....31 5.1.2 Design…………..………..………...….…….……...31
5.1.3 Stimuli and measurements………..…………..……...……….…...…..31
5.1.4 Procedure….………..………….……….……....33
5.1.5 Data analysis………..……...34
5.2 Results……...………..………...……...…...35
5.2.1 Normality check……...…...……….…………..…....35
5.2.2 Manipulation check.….………..………...….…….……...35
5.3 Discussion………..…37
6. General discussion……….……….…39
6.1 Findings……….……….39
6.2 Limitations and directions for future research……….…..…41
7. References……….………...44
8. Appendices………..……….………..49
8.1 Survey study I………...……….………....49
ABSTRACT
Nowadays, customer information has become the key aspect on which many businesses depend. As a result, consumers are increasingly concerned about their privacy. Nevertheless, their behavior is not always consistent with these concerns. In this paper, the so-called ‘privacy paradox’ is explained based on Construal Level Theory. In order to do this, two studies have been conducted. The first study proves that the benefits of sharing personal data online are perceived as psychologically closer than the costs of sharing them. In the second study, however, no evidence was revealed for the relation between the kind of life goals individuals hold(high-‐level idealistic vs. low-‐level, pragmatic) and the reliance on the costs vs benefits of the disclosure of information. Similarly, no evidence was found for the expectation that psychological distance is the underlying mechanism that moderates the effect of life goals on consumers’ reliance on the costs vs benefits of information disclosure.
1. INTRODUCTION
In recent years, advanced technology possibilities such as the internet have increased the availability and access to information (Fournier & Avery, 2011), which enables businesses to rely more heavily on customer information in developing their activities (Hagel & Rayport, 1997; Jenkinson, 2006; Lynn, 1998; Messner, 2004; Bessen, 1993; Blattberg, 1991; Glazer, 1991). Customer information includes personal or individual-specific data “such as names, addresses, demographic characteristics, lifestyle interests, shopping preferences, and purchase histories of identifiable individuals” (Nowak & Phelps, 1995; Phelps, Nowak & Ferrell, 2000, p. 28).
A consequence of this increased usage of customer information is that consumers are more aware of privacy issues and are concerned about their privacy (Peltier, Milne & Phelps, 2009; Lee, Im & Taylor, 2008). Several researchers state that privacy concerns will lead consumers to be more in control of the amount and depth of information that they share with companies (Hagel & Rayport, 1997; Peltier et al., 2009).
But how do consumers deal with these privacy concerns? Do they really undertake actions to be more in control of the amount and depth of information that they share with others? Several researches have demonstrated, that even though people are increasingly concerned about their privacy, their behavior is not always consistent with these concerns (Jorstad, 2000; Taddicken, 2014; Norberg, Horne & Horne, 2007). Consumers do not undertake actions to control their personal information better and they keep freely providing their data on the internet (Culnan & Armstrong, 1999; Acquisti & Grossklags, 2005; Norberg et al., 2007). This phenomenon of inconsistency between people’s privacy concerns and data disclosure behavior is referred to as the ‘privacy paradox’ (Taddicken, 2014; Jorstad, 2000; Norberg et al., 2007; Spiekermann, Grossklags & Berendt, 2001).
In attempt to explain online data disclosure behavior, different perspectives have been applied. Culnan & Armstrong (1999) assume that the decision to disclose personal information is a rational one. They assume this decision is being based on a cost-benefit calculation of the consumer. The consumer, as the so-called ‘privacy calculus’, makes a rational decision in which the perceived benefits are weighted against the perceived costs of this ‘transaction’. In this perspective, the consumer will
only disclose personal information, if the benefits outweigh the costs (Culnan & Armstong, 1999).
Expanding this thought, Acquisti & Grossklags (2005) state that the privacy decision-making process is not a fully rational one. According to them, this process is affected by other irrational factors as well. In their paper, it is argued that decisions regarding privacy are affected by incomplete information, bounded rationality and the desire for immediate gratification (Acquisti & Grossklags, 2005).
Nevertheless, the inconsistency in consumer privacy concerns and their behavior has not been fully explained so far; no overarching theory explains this disproportionate discrepancy between attitudes and behavior. Therefore, the main goal of this explanatory research is to explain this gap. In line with the irrational perspective of Acquisti & Grossklags (2005) I will propose that a Construal Level Theory perspective can be used to explain the abovementioned paradox, and this proposition will be examined. Therefore, the research question in this paper is: How does Construal Level Theory explain the inconsistency in consumer’s privacy concerns and their information sharing activities?
If Construal Level Theory can account for the explanation of this inconsistency between behavior and privacy concerns, this implies an important scientific contribution. Although several studies have been conducted to prove the ‘privacy paradox’ (Taddicken, 2014; Jorstad, 2000; Acquisti & Grossklags, 2005; Norberg et al., 2007), and although Acquisti & Grossklags (2005) have attempted to explain this paradox, a comprehensive explanation has not yet been found. In this paper, an important social theory is used to explain the inconsistency. Based on Construal Level Theory, I will propose a new perspective in explaining the abovementioned paradox. Combining and synthesising the Construal Level Theory with the Privacy Paradox literature could deliver some important new insights.
This research makes some important managerial contributions as well. Because of the great importance of customer information for companies in developing and bettering their products and services in the contemporary competitive marketplace (Hagel & Rayport, 1997; Jenkinson, 2006), organizations should understand how their customers act on their privacy concerns. If the research question of this thesis can be answered properly, this provides a clear explanation of the inconsistency between consumer’s concerns and behavior. Companies are then better able to understand the behavior of their consumers and can adjust their business
activities to this. Without this understanding, firms will not be able to design their information acquisition process around their customers and to gain the needed information (Zimmer et al., 2010).
The remainder of this paper is organized as follows. In the next chapter, a theoretical framework will explain the constructs that are included in the research. These constructs will be used to conduct different studies that aim to explain the inconsistency in people’s privacy concerns and their online data sharing behavior. Then, the method of the studies will be explained and the results will be discussed. Finally, the implications for these results will be discussed together with both scientific and managerial contributions, and some limitations of this paper will provide interesting directions for future research.
2. THEORETICAL FRAMEWORK
2.1 Privacy Concerns and Privacy Behaviors
As a result of the increased usage of customer information over the past years, consumers are becoming more aware of privacy issues involved in their information sharing activities (Nowak & Phelps, 1992; Peltier et al., 2009; Debatin, Lovejoy, Horn & Hughes, 2009; Williams, 2002). They are increasingly aware of the fact that sharing personal information does not only bring benefits, but also brings risks to the table (Hagel & Rayport, 1997; Zimmer, Arsal, Al-Marzouq, Moore & Grover, 2010) and as a result it has been found that the majority of American citizens hold privacy concerns (Phelps, Nowak & Ferrell, 2000).
Goodwin (1991) defined privacy as the “control of information disclosure and control over unwanted intrusions into the consumer’s environment” (Peltier et al., 2009) and Mothersbaugh, Foxx & Wang (2012) defined consumer online privacy concern somewhat more specific as “consumers’ concerns about the use of their revealed information for marketing purposes”.
The main privacy concerns that consumers hold are about how companies gain and use consumers’ personal data, what companies exactly know about them, and for which purposes the information is used (Phelps et al., 2000). The majority of the concerns are about personal and individual-specific data as for example; who people are; where they live; other demographic characteristics; lifestyle interests; shopping preferences; and individuals purchase histories (Phelps et al., 2000; Nowak & Phelps, 1995). According to Peltier et al. (2009), the overall privacy concern is related to the amount and types of information collected, data collection without the knowledge of consumers and about the fact that data could be sold to other organizations and will be out of the control of consumers. In addition, Taddicken (2014) states that consumers have the desire to stay in control of who collects and distributes their personal information. Elaborating this assumption, several researchers state that privacy concerns will lead consumers to be more in control of the amount and depth of information that they share with companies (Hagel & Rayport, 1997; Peltier et al., 2009). Based on previous research (Dommeyer & Gross, 2003; LaRose & Rifon, 2007; Schoenbachler & Gordon, 2002), Peltier et al. (2009) state that consumers are changing their behavior and for example refuse information
requests, subscribe themselves in no-call lists or unsubscribe themselves of direct marketing lists, in order to ensure that their vulnerability to potential privacy violations will be reduced (Peltier et al.,2009).
Although the above clearly shows that people are increasingly concerned with their privacy, several studies have demonstrated that people’s disclosure behavior is not always consistent to their privacy concerns (Acquisti & Grossklags, 2005; Acquisti & Gross, 2006; Jorstad, 2000; Norberg et al., 2007; Taddicken, 2014; Tufekci, 2008). For example, Acquisti & Grossklags (2005) found in their study that only a small number of social network users have changed their default privacy preferences to minimize the visibility of their user profiles. In another study, they found that the privacy concerns an individual holds is a weak predictor of their social media behavior; when people hold these concerns they still join social networks and share personal information (Acquisti & Gross, 2006). Also Tufekci (2008) and Taddicken (2014) found that privacy concerns bear little relationship with people’s personal information sharing behavior on social media like Facebook. In this study, I will try to explain this paradox by Construal Level Theory.
2.2 Psychological Distance
The basic assumption of Construal Level Theory (CLT) is that the more psychologically distant an event is, the more abstract and high-level construals are used by individuals to represent these events (Liberman & Trope, 1998; Trope, Liberman & Wakslak, 2007; Trope & Liberman, 2010). Thus, these construals that are used by individuals to represent events can be seen as a consequence of perceived psychological distance (Trope & Liberman, 2003).
Psychological distance can be defined as “a subjective experience that something is close or far away from the self, here, and now” (Trope & Liberman, 2010; Trope et al. 2007). Trope & Liberman (2010) state that we can only experience what is present, but we cannot experience what is in the past or in the future. Therefore, Trope & Liberman (2010) conclude that psychological distance is egocentric: the self, here, and now is the reference point, and everything that is removed from that point is different in distance. We experience less distant events close to a direct experience. According to this theory, we construe this event occurring in the ‘here and now’, to represent these events. We have substantial information available about it and we think about these direct experiences in
concrete, and low-level terms. We use all the detailed and contextualized information that is available (Trope & Liberman, 2010; Trope et al., 2007).
For more distant events, we can make assumptions, imaginations or predictions instead of direct experiences (Trope & Liberman, 2010; Trope et al. 2007). These assumptions, imaginations, predictions, etc. are based on how we psychologically distant see these events. The further we construe the events from a direct experience and the more distant we see the event, the less reliable and available information there is and the more abstract and high-level our perception will be (Trope & Liberman, 2010; Trope et al. 2007).
In previous research, several dimensions of distance have been described, such as temporal distance, spatial distance, probability distance, social distance (Trope et al., 2007; Trope & Liberman, 2003; Todorov, Goren & Trope, 2007; Liberman, Sagristano & Trope, 2002). These dimensions of psychological distance affect the construal levels that people hold (Todorov et al., 2007).
In the current research, two dimensions of distance will be used to explain the gap between current customer information disclosure behaviors and their privacy concerns. These dimensions are temporal distance and hypothetical distance and are chosen since costs and benefits involved in sharing personal data are often inherent to hypothetical and temporal distance.
Hagel & Rayport (1997) state that consumers are willing to release personal information, but only if they know that they can profit of it. In this, it is thus important for customers, when they will gain these benefits and how sure they are of gaining them. In addition, Culnan & Armstrong (1999) argue that consumers only disclose personal information, if the benefits outweigh the costs. The question whether how soon and how sure customers will gain their benefits, and how soon and how sure they will outweigh the negative aspects is here of great importance. Thus, in the current paper will be focused on the dimensions of hypothetical and temporal distance since these are expected to be the most relevant in the current research.
I argue that temporal and hypothetical distance affect the way people construe their perceptions about information disclosure and affect the way they hold their privacy concerns. The main expectation of this paper is that the reason for the inconsistency in concerns and behavior are the differences in levels of construal of costs and benefits. In the current paper, it is expected that people perceive the benefits of disclosing personal data online as less psychological distant. This, because they
will have the genuine experiences of certainly receiving immediate benefits when they share personal data, such as discount in exchange for their subscription for the companies’ newsletter.
In contrast, it is expected that people perceive the privacy losses as more distant, since their experiences, in general, will be that they do not know íf and when these losses will occur. For instance, when people share their personal data, their information could be sold to other parties, but consumers don’t know if this indeed will happen or when this event will occur. These disadvantages of sharing personal data are thus seen as delayed and probable events, in contrast to the advantages that are seen as immediate and certain events.
Thus, it is expected that consumers will, based on their genuine experiences, perceive the benefits as psychologically closer than the costs, and that they will therefore construe the benefits in a more low-level and concreter way than the costs. In the next section, these two dimensions of psychological distance will be explained more briefly.
2.3 Temporal & hypothetical distance
In several studies, temporal distance is mentioned as one of the most important influencers of psychological distance (Trope et al., 2007; Liberman et al., 2002; Trope & Liberman, 2003; Förster, Friedman & Liberman, 2004). These studies demonstrated that more distant future events are in general represented in a more abstract, high-level and structured way, and near future events are represented in a more concrete, low-level way (Trope et al., 2007; Liberman et al., 2002; Trope & Liberman, 2003; Förster et al., 2004). If these findings are applied to peoples contrasting disclosure behavior regarding their privacy concerns, it could be argued that privacy concerns are perceived as more future distant events and the benefits of disclosure as more near future events. The benefits of disclosing personal information are perceived much more concrete, since these benefits will occur in the near future, in the here and now. The privacy losses will turn into a real event in the more distant future and will be therefore perceived as more abstract.
Another important determinant of psychological distance is probability, or hypothetical distance. Several researches have been conducted on the ways in which probability instantiates psychological distance (Wakslak & Trope, 2009; Todorov et al., 2007). In these researches, it was found that the more likely it is that an event will
occur, the less distant this event is perceived (Wakslak & Trope, 2009; Todorov et al., 2007). Therefore, it is assumed that events that are less likely to occur, are perceived as more distant. Since there is a high probability that the benefits of the information sharing activity (e.g. for optimal usage of a website or application) will be delivered, the current research argues that consumers perceive benefits as more concrete and low-level; there will be reliable available information about this event for the consumer within a short period of time. In contrast, it could be argued that the privacy loss, the cost of the information sharing activity, is less probable to occur. The information shared by the customer will not necessarily be used in a detrimental manner to the consumer. Thus, receiving a benefit could be seen as something with a high probability and could be classified as a low psychological distant event, whereas the privacy loss could be stated to be less probable to occur and could be classified as a more psychological distant event.
When taking previous studies into account and combining these with current assumptions, it is expected that people will conceptualize the actions of disclosing personal data in concrete, less abstract, low-level terms, while they will conceptualize their privacy concerns in more abstract, high-level terms. As such, it is expected that consumers perceive disclosing personal data as less psychological distant and see privacy concerns as more psychological distant:
H1a: The benefits of disclosing personal data online are perceived as psychologically closer (temporally & hypothetically) than the costs of disclosure.
An important finding of Construal Level Theory is that people undervalue risks when they are more distant (Trope, et al., 2007). This means, that when psychological distance increases, people are less capable to assess the risks or costs of sharing personal data online in reality. Furthermore, Trope et al. (2007) emphasize the concept of delayed gratification. This concept contains the fact that consumers have a preference for smaller, sooner benefits over later, larger ones. Thus, consumers are impatient and overvalue the benefits on the short term.
If the current research can confirm the expectation that consumers perceive the benefits of disclosing personal data online as psychologically close, and the costs as psychologically distant, the so-called privacy paradox can be explained based on the risks being undervalued and on the concept of delayed gratification.
If people know what the risks of their sharing activities are, they are less capable to assess these risks since they perceive them as more distant and thus construe these events in a more abstract way. This leads to an undervaluation of these risks, and as a consequence, people will more easily choose the benefits.
In terms of the concept of delayed gratification, if people do indeed value their privacy and they choose to protect it, their privacy will be protected in the future and this protected privacy can be seen as delayed benefit. In line, I argue that people’s privacy concerns and their online information sharing behavior is inconsistent because people perceive the benefits as more psychologically close, and thus prefer these immediate benefits such as enabled functionalities, over the delayed benefits in the sense of protected privacy.
2.4 Construal level of benefits and costs
The more psychologically close an event is, the more concrete and low-level construals are used by individuals to represent events, and the more psychologically distant an event is, the more abstract and high-level events are represented by individuals (Liberman & Trope, 1998; Trope et al., 2007; Trope & Liberman, 2010). Construal Level Theory suggest that psychological distance is an important determinant for the kind of characteristics that are used by individuals in their evaluations, behavior and choices. The construals used by individuals to represent events can be seen as a consequence of perceived psychological distance (Trope & Liberman, 2003).
The first hypothesis indicates that benefits of disclosing personal data online are perceived as psychologically close, and that the costs of these disclosure activities are perceived as psychologically distant. If this can be demonstrated, I will propose that as a result of the perceived psychological distance of benefits and costs of sharing personal data online, the benefits are construed in a low-level and concrete manner. In contrast, the costs that come with the disclosure of personal data online will be construed in more abstract and high-level ways. These levels of construal, as a result of the perceived psychological distance, will influence people’s behavior in the privacy decision-making process.
For the benefits, the abovementioned implies that one is more able to imagine the occurrence of these benefits. For example, people are able to imagine more details of the benefits, can describe such an occasion more in context, and can provide more
information about such an event. In contrast, people construe the costs in more abstract and high-level ways, and will be, in comparison to the concrete and low-level ways, less able to provide details about such an occasion and will describe things in more general ways.
Thus, how distant people perceive events influences the way in which they construe these events (either in a high-level or low-level way) and this construal affects people’s evaluation and behavior. In applying this to the data disclosing behavior of people, I assume that Construal Level Theory is the underlying mechanism of the effect of psychological distance of benefits and costs on their disclosing behavior of personal data.
H1b: The benefits of disclosing personal data online are construed on a lower-level, concreter way, than the costs of disclosure.
2.5 Values and feasibility concerns
People have different values, affecting the way a situation is being defined and experienced, and influences the way a person perceives potential outcomes of their behavior as attractive or not (Feather, 1995).According to Feather (1995), values can be interpreted as continued abstract structures of the ideas people have of what desirable behaviors and desirable end states contain. According to Construal Level Theory, Fujita, Eyal, Chaiken, Trope, and Liberman (2008) have found that high-level features impact people’s intentions on events in the distant future more strongly. Since values have a relative abstract, structured character, people use values more in construing their intentions in future situations (Eyal, Sagristano, Trope, Liberman & Chaiken, 2009). Also Liberman & Trope (1998) state that values are more used to guide intentions in psychological distant situations. Since I expect that the costs of sharing data online are perceived as psychologically more distant, I assume here that in this distant time perspective people’s values thus will be of great importance for people in their choices to share their data or not.
Eyal et al. (2009) did not only find evidence for their assumptions that values impact intentions in psychological distant, high-level situations, but also that feasibility concerns are more impactful for near future events. Near future intentions are predictable rather by low-level feasibility concerns, than by high-level value-related aspects. I expect the benefits of data disclosure online to be perceived as more
psychologically close than the costs, therefore I assume here that feasibility concerns will be related to the benefits of sharing personal data.
According to Liberman & Trope (1998), the values of people are reflected by their desirability, whereas feasibility is about what a person has to do to reach its values. In this, I assume that whether people rely more on desirability or on feasibility depends on people’s self-conceptions.
2.6 Self-conceptions & Implications
In their article, Kivets & Tyler (2007) describe two types of self-conceptions, the idealistic and the pragmatic selves. They define the idealistic self as “a mental representation that places principles and values above practical considerations and seeks to express the person’s sense of true self”, and the pragmatic self is defined as “an action oriented mental representation that is primarily guided by practical concerns”. Which self-representation will be activated, is different in each situation and will vary per person (Bargh, 1990). According to Construal Level Theory, Kivets & Tyler (2007) propose that a distant time perspective activates the idealistic self, and a near distant time perspective shifts attention towards the pragmatic self. In line with the research of Eyal et al. (2009) and in terms of Construal Level Theory, the idealistic self is a high-level self-conception, and the pragmatic self is a low-level self-conception.
In general, the idealistic self is activated in a more distant time perspective and the individual seeks to express the true self, using principles and values to represent the situation, whereas the pragmatic self is activated in a near distant time perspective and, as a consequence, the individual is primary guided by practical concerns.
In their paper, Kivets & Tyler (2007) mention two major implications of these different self-representations for motivation and preference; maximizing identity versus instrumental benefits. Maximizing identity motives will be enhanced when the idealistic self is activated, since individuals place principles and values above practical considerations in order to express the true self. Thus, people find it important to define, pursue and express their true self in such situations, and their motivation and preference is intrinsic in its basis (Kivets & Tyler, 2007; Deci & Ryan, 1985).
When the pragmatic self is activated on the other hand, individuals have a strong focus on situational opportunities and practical concerns in their action guidance. As a result, this enhances the desirability of instrumental goals and extrinsic inducements (Kivets & Tyler, 2007; Deci & Ryan, 1985). Consequently, Kivets & Tyler (2007) state that instrumental resources like money and goods will be more influential when the pragmatic self-conception has been activated.
When taking these life goals of consumers in account and applying them to consumer’s data disclosure behavior, I argue that these life goals will influence people’s behavior in the information disclosure process. Disclosing personal data online is often rewarded with instrumental resources like discounts, vouchers, enabled functionalities, or other benefits that comport with goals that people hold, such as access to a particular website, or quick usage of an application. These instrumental resources as for instance discounts are the desired situational opportunities and practical concerns which play an important role when the pragmatic self as low-level life goal, is activated. Thus, when people hold pragmatic life goals, they are more oriented towards situational opportunities and practical concerns, and people will be oriented towards the short-term benefits of disclosing their data online. On the other hand, disclosing personal data online is often accompanied with possible privacy losses. When people hold an idealistic self-conception and more long-term life goal, they seek to express their self. Values and ideals enhance to the importance of the own identity, and people holding idealistic self-conceptions seek to express their identity. Information about their identity is important, but with the risk of privacy losses they will have the chance to loose such information about the self, and to be less able to express the self. Hence, people with more long-term, value-oriented life goals, seek to protect the self and their values, by protecting their privacy on the long-term.
To conclude, when consumers are short-term focused and hold a pragmatic, low-level life goal, their motivation and preference is extrinsic in its basis, and they will have a strong focus on situational opportunities and practical concerns such as the benefits they gain in disclosing their personal data. In contrast, when people hold more long-term focused, value-oriented and idealistic life goals, their motivation and preference is intrinsic in its basis and consumers seek to express their self and thus to protect their own privacy.
If the first hypothesis is confirmed, it can be said that people perceive the benefits of disclosing personal data online as psychologically close. Since consumers that hold low-level life goal are short-term oriented and thus focus more on psychologically close events, and the fact that their motivation is more extrinsic in its basis, I state that the degree to which consumers rely on the benefits of disclosing personal data will be enhanced if they hold pragmatic life goals.
In contrast, if the costs of sharing personal data online are indeed perceived psychologically distant, and if people are more long-term oriented when they hold high-level life goals and they are more intrinsic in its basis, the degree to which consumers depend on the costs of disclosing their personal data will be of greater influence if they hold more idealistic life goals. This leads to the following hypothesis:
H2a: Consumers’ reliance on the benefits of disclosing personal data online will be enhanced when consumers hold a low-level goal (vs. high-level goal), whereas consumers’ reliance on the costs of disclosing personal data exert a greater influence when consumers hold a high-level goal (vs. low-level) goal.
If consumers’ reliance on the benefits of disclosing personal data online is indeed enhanced when consumers hold a level life goal, I expect consumers to hold low-level life goals when in their decision making process on whether or not to share personal data online. An important finding in Construal Level Theory, is the concept of delayed gratification, emphasized by Trope et al. (2007). Consumers have a preference for smaller, sooner benefits over later, larger ones. Thus, consumers tend to be more short-term, low level oriented and are more easily to provoke for pragmatic, situational opportunities.
In this abovementioned hypothesis (H2a), the benefits of sharing personal data online were assumed to be perceived as more psychologically closer than the costs. As mentioned beforehand, the Construal Level Theory suggests that
psychological distance is an important determinant of what kind of characteristics are used by individuals in their evaluations, behavior and choices, since the construals used in the evaluations of individuals are a consequence of the perceived
Building on this theory, I expect that Construal Level Theory is the
underlying mechanism for the relationship between life goal and reliance on costs or benefits. If, in contrast to the previously proposed, the benefits of sharing personal data are perceived as psychological more distant than the costs of sharing data, people with more idealistic life goals should rely more on the benefits instead of the costs since they are more long-term and high-level oriented. In contrast, people holding more pragmatic life goals should rely more on the costs instead of the benefits when choosing whether to share data or not, since they are more short-term and low-level oriented.
Thus, the effect of life-goal on reliance on costs or benefits should turnaround, when the benefits of disclosing data are perceived as psychologically more distant than the costs of data sharing. This leads to the following hypothesis:
H2b: The effect of life goal on consumer’s reliance on the benefits or costs of disclosing personal data online is moderated by perceived psychological distance.
3. OVERVIEW OF THE PRESENT STUDIES
The objective of the present research is to explain the inconsistency in consumer’s privacy concerns and their personal information sharing behavior in terms of the Construal Level Theory. Levels of construal are a result of how psychologically distant people perceive events, which influences people’s behavior in their decision to enclose personal data. Since people’s life-goals influence the way they behave, it is expected that people’s life-goals influence the decision making process in sharing personal data as well.
Two studies were conducted to explain the gap in people’s privacy concerns and their information sharing activities using Construal Level Theory. In the first study, the expectation that the benefits of sharing personal data online would be perceived psychologically closer than the costs of sharing personal data (H1a) was tested. The analysis was conducted for two dimensions of psychological distance, which were hypothetically distance and temporal distance. Afterwards, the expectation that the benefits of disclosing personal data online would be construed on a lower, concreter level than the costs of disclosure (H1b) was tested.
The second study tested if participants would indeed rely more on the benefits of disclosing their personal data when they held a low-level life goal, and if they relied more on the costs of disclosing their personal data when they held a high-level goal (H2a), and if this effect of life goal on reliance on benefits or costs was moderated by perceived psychological distance (H2b). A survey was used to conduct an experiment, in which participants were manipulated for their life goal and for psychological distance.
A quantitative research method was used to test the hypotheses of both the first and the second study. Both studies were conducted in Dutch, since this was the native language of the vast majority of potential participants.
4. STUDY I
4.1 METHOD 4.1.1 Participants
In order to test the first hypothesis, 79 Dutch people participated in a convenience sample. People were approached by e-mail and online social networks (Facebook and LinkedIn) to participate in the study. They received an online link that redirected them to the online survey (Appendix I). From those 79 people who decided to participate in the survey, 53 completed the survey (completion rate: 67%) and therefore only their data was used. Of this sample, 17% were male (N=9) and 83% were female (N=44). The average age was 31, but approximately 68% of the participants were between 23 and 27 years old. All participants were highly educated, since they all completed a Bachelor’s degree (28,3%, N=15) or Master’s degree (71,7%, N=38). To make sure this demographic characteristics of participants could not influence the results of the tests, the test controlled for age and gender.
4.1.2 Design
The main goal of the study was to measure the perceived psychological distance (temporal and hypothetical) and the construal level of the costs and benefits of disclosing personal data online. This was done by a survey, in which participants were asked to indicate their subjective psychological distance. A within-subjects design was used to test the hypotheses.
4.1.3 Stimuli and measurements
Participants were exposed to three everyday situations in which people have to make the decision to either or not disclose their personal data (Appendix I). The first of these situations was the ‘fitness-app’-situation, containing the following description: “You are using a fitness application on your smartphone. With this application, you track how much you exercise and what you eat on a daily basis. The application asks whether you would like to share these personal data or not”. The second situation was the ‘cookies’-situation, which was described as: “When you are visiting a fashion-webshop, you are asked to accept cookies”. The last situation was the ‘Facebook
log-in’-situation, which was described as follows: “You just downloaded a new application for your smartphone. When opening the application the first time, you are asked to choose to 1) create a new account, or 2) to log-in with your existing Facebook account”. Based on these three situations, the following stimulus was developed.
Self-indicated benefits and costs of sharing personal data online. Participants
were asked to indicate as many advantages and disadvantages they associated with sharing their personal data in the particular situation they were exposed to as possible. These self-reported advantages and disadvantages, or benefits and costs of disclosing personal data online respectively, where used as stimuli in measuring the perceived psychological distance of benefits and costs and in measuring the construal level of these benefits and costs across these various settings.
To measure the perceived psychological distance, participants were asked to indicate for each particular context how far the situation felt from the self. Two dimensions of psychological distance were measured, namely hypothetical distance and temporal distance. This was done using a slide question with a scale from 1 (very distant) to 7 (very near) for hypothetical distance, and with a scale from 1 (very near) to 7 (very distant) for temporal distance. This measurement was used in previous research by Broemer, Grabowski, Gebauer, Ermel, & Diehl (2008) and Polman & Emich (2011) to measure psychological distance. The more distant participants set their perceptions, the more hypothetically and temporal psychologically distant they perceived the particular situations.
To measure the construal level of benefits and costs, the content of the answers on the self-reported advantages and disadvantages was analysed. The researcher assessed the level of concreteness/abstractness of each consequence that the participants had indicated, using the concreteness rating scale of Stöber, Tepperwien & Staak (2000) which was used in previous research by Watkin & Moulds (2005) and Stöber & Borkovec (2002). In rating each consequence using a scale from 1 (concrete) to 5 (abstract), definitions of Stöber & Borkovec (2002) were used. They defined concrete as “distinct, situationally specific, unequivocal, clear, singular”, and abstract as “indistinct, cross-situational, equivocal, unclear, and aggregated” (Stöber & Borkovec, 2002, p.92).
4.1.4 Procedure
Participants where invited online to participate in the survey. When they clicked on the provided link, they were redirected to the online survey tool where the survey could be started. The introduction contained a description of the context of the research, an indication of the amount of time it would take to complete the survey, and a description of the fact that there were no right or wrong answers. Furthermore, the confidentiality of the answers that would be provided was emphasized and respondents were thanked for their participation in the research.
Then, all participants were exposed to three situations. The sequence in which these situations were showed was randomized. All three situations contained the same procedure and questions. In the first question of the situation, participants were asked to decide upon how likely it was whether they would share their personal data or not, on a scale from 1 (not likely at all) to 5 (very likely). Afterwards, respondents were invited to indicate as many advantages and disadvantages as possible, but at least two, they associated with sharing their personal data in that particular situation they were exposed to. In the next two questions the advantages and disadvantages listed by the participant were piped into a slide question. Participants could indicate per (dis)advantage 1) the perceived probability of the event actually occurring (hypothetical distance), and 2) the timeframe within they expected each event to occur (temporal distance). Afterwards, participants were asked to indicate how advantageous or disadvantageous they perceived the items that they previously filled in, on a scale from 1 (big disadvantage) to 5 (big advantage). With this, perceived psychological distance was measured (Broemer et al., 2008; Polman & Emich, 2011).
After participants completed all three situations, some demographic questions were asked. These demographic questions included year of birth, gender and level of education. At the end of the survey, some questions were asked concerning the usage of the Internet of participants and to what extent people would value their own lifestyle as healthy.
4.1.5 Data Analysis
To analyse the data SPSS Statistics was used. The raw data were prepared for the analysis. After checking the missing values, the data were checked for counter-indicative variables and the hypothetical distance scale was reversed recoded. Furthermore, descriptive statistics were calculated to check for errors.
The reported consequences were considered as the unit of analysis, as for every consequence temporal distance, hypothetical distance, and valence (pro or con) was reported. New variables were created for valence (on the basis of self-reported valence), participant and situation. To account for the fact that consequences were independently nested within participants and situations, a cross-categorical multilevel analysis was performed.
4.2 RESULTS 4.2.1 Psychological distance
Two mixed model analyses were performed to test if the benefits of disclosing personal data online were perceived as psychologically closer (hypothetically & temporally) than the costs of disclosing personal data online (H1a). Thus, one mixed model analysis was performed to test the effect of costs and benefits and situation on hypothetical distance and a second mixed model analysis was performed to test the effect of costs and benefits and situation on temporal distance. The mixed model analysis seemed to be an appropriate choice, since this model can include the hierarchical data structure in the analysis. Assumed was that both intercepts and slopes could vary around the model, and therefore the random intercept and slope model was used.
There were two independent variables (costs & benefits and situation), one dependent variable (hypothetical distance or temporal distance), and one contextual variable (participants). Both analyses controlled for age and gender. Since all participants were highly educated and this would thus not influence the hypothesis testing, there was not controlled for education level.
Hypothetical distance
The relationship for costs and benefits and perceived hypothetical distance showed significant variance in intercepts across advantages and disadvantages. Slopes did not vary across participants or situations.
The results of the mixed model analysis for hypothetical distance are shown in table 1. As can be seen in figure 1, a significant fixed effect was found for costs and benefits on levels of perceived hypothetical distance,
F(1,59.97) = 6.15, p < .05. This
means that costs are perceived significantly more hypothetical distant than the benefits of
disclosing personal data online. The fixed effect of situation on levels of perceived hypothetical distance was significant, F(2,109.06) = 4.18, p < .05. Therefore it can be said that the perceived levels of hypothetical distance varies across the different situations tested.
A post-hoc test was conducted and no significant difference was found for the ‘cookies’ situation and the ‘Facebook log-in’ situation (MD = -.005, p = .98), but the ‘fitness-app’ situation was significantly higher in hypothetical distance than the ‘cookies’ situation (MD = .43, p < .05) and the ‘Facebook log-in’ situation (MD = .43, p < .05).
In the test, there was controlled for gender and year of birth. Gender had no significant influence on the levels of perceived hypothetical distance, F(1,91.52) = .00, p = .99. For year of birth, there was also found no significant effect, F(1,87.665) = 3.06, p = .08. Hence, age has not significantly influenced the level of perceived hypothetical distance.
There was no significant interaction effect found between cost and benefits and situation, F(2,336.29) = 2.30, p = .10. In sum, the analysis showed that the perceived hypothetical distance is significantly higher for the costs of sharing personal data online than for the benefits. Since the expectations were that the costs of sharing personal data online would be perceived more hypothetically distant than the benefits, the first hypothesis is supported for this part.
1,8 1,9 2 2,1 2,2 2,3 2,4 2,5 2,6 BeneVits Costs Hyp ot het ica l dist ance
Figure 1: Fixed effect of cost & bene4its on hypothetical
Table 1: Results of the cross-categorical multilevel analysis for hypothetical distance
Benefits Costs Total
N M (SD) N M (SD) N M (SD)
Fitness-app 60 2.22 (.99) 87 2.64 (1.39) 147 2.47 (1.26) Cookies 51 2.08 (1.21) 82 2.00 (1.24) 133 2.03 (1.22) Facebook log-in 66 1.86 (1.01) 98 2.22 (1.50) 164 2.08 (1.33)
Temporal distance
The relationship for costs and benefits and perceived temporal distance showed significant variance in intercepts across advantages and disadvantages. Slopes did not vary across participants or situations.
The results of the mixed model analysis for temporal distance are shown in table 2.
As can be seen in figure 2, there was a significant fixed effect of costs and benefits on levels of perceived temporal distance,
F(1,48.57) = 19.36, p < .01. The
costs of sharing data online were perceived significantly more temporal distant than the benefits. There was also a significant fixed effect of situation on levels of perceived temporal distance,
F(2,85.21) = 6.92, p < .01. The level of perceived temporal distance varies thus
across the tested situations.
As found in the analysis on hypothetical distance, a post-hoc analysis revealed that no significant difference was found for the ‘cookies’ situation and the ‘Facebook log-in’ situation (MD = .163, p = .39), but the ‘fitness-app’ situation was significantly higher in temporal distance than the ‘cookies’ situation (MD = .50, p < .05) and the ‘Facebook log-in’ situation (MD = .66, p < .01).
1,7 1,9 2,1 2,3 2,5 2,7 2,9 BeneVits Costs T emp or al dist ance
Figure 2: Fixed effect of cost & bene4its on temporal
The test was controlled for gender and year of birth and gender appears to have no influence on the perceived level of temporal distance F(1,46.84) = 1.43 p = .24, neither had year of birth, F(1.50.28) = .64, p = .43. In other words, for the perception of temporal distance it does not matter in which year people have been born.
Furthermore, there is no significant interaction effect revealed between costs and benefits and situation, F(2,362.95) = 1.79, p = .17. Overall, the analysis showed that the perceived temporal distance is significantly higher for the costs of sharing personal data online than for the benefits. Therefore, the findings are in line with the expectations and the first hypothesis is supported for the expectations of temporal distance. Thus, H1a is supported for both hypothetical and temporal distance.
Table 2: Results of the cross-categorical multilevel analysis for temporal distance
Benefits Costs Total
N M (SD) N M (SD) N M (SD)
Fitness-app 60 2.47 (1.76) 87 2.94 (1.81) 147 2.75 (1.80) Cookies 51 1.82 (1.35) 82 2.63 (1.61) 133 2.32 (1.57) Facebook log-in 66 1.50 (.98) 98 2.53 (1.58) 164 2.12 (1.46)
4.2.2 Construal Level
In order to test the prediction that the benefits are construed in a more concrete, low-level way compared to the costs of disclosing personal data online, and the costs of disclosure are construed in a more abstract, high-level way compared to the benefits (H1b), a mixed model analysis was performed. As mentioned before, this model can include the hierarchical data structure in the analysis, and therefore this model was chosen to be appropriate. As well as in the previous analysis, it was assumed that both intercepts and slopes could vary around the model, and therefore the random intercept and slope model was used.
Furthermore, there were two independent variables (costs & benefits and situation), one dependent variable (construal level of benefits and costs), and one contextual variable (participants). The analysis was controlled for age and gender. Since all participants completed a Bachelor’s degree or a Master’s degree and
therefore were highly educated, it was impossible to control for the participants’ education level.
The relationship for costs and benefits and construal level showed significant variance in intercepts across advantages and disadvantages. Slopes did not vary across participants or situations.
The results of the mixed model analysis for construal level are showed in table 3. As can be seen in figure 3, a
significant fixed effect was found for costs and benefits on construal level,
F(1,100.84) = 8.84, p < .01. The
benefits of sharing personal data online were construed significantly in a lower-level, concreter way than the costs of disclosure. No significant fixed effect of situation was found on construal level, F(2,387.18) = 1.35, p = .259. Thus, construal level did not vary across the different tested situations.
The analysis controlled for gender and year of birth and it appeared that both gender, F(1,95.98) = .09, p = .77, and year of birth, F(1,93.3) = .55, p = .46, had no influence on level of construal. Therefore, it can be said that neither gender nor age influences the extend to which people construe costs and benefits of sharing personal data online.
Furthermore, no significant interaction effect was found between costs and benefits and situation, F(2,337.95) = 1.45, p = .24. Overall, the analysis showed that people construe the benefits of sharing data online on a lower, concreter level, than the costs of disclosure. These findings are in line with the expectations, and support hypothesis H1b. 2,2 2,4 2,6 2,8 3 3,2 BeneVits Costs Con st ru al level
Figure 3: Fixed effect of cost & bene4its on construal level
Table 3: Results of the cross-categorical multilevel analysis for level of construal
Benefits Costs Total
N M (SD) N M (SD) N M (SD)
Fitness-app 60 2.47 (1.21) 86 2.97 (1.37) 146 2.76 (1.33) Cookies 51 2.49 (.97) 82 2.73 (1.3) 133 2.64 (1.18) Facebook log-in 66 2.53 (1.08) 98 3.10 (1.07) 164 2.87 (1.11)
4.3 DISCUSSION
The main expectations of the first study were that 1) the benefits of online personal data disclosure would be perceived as psychologically closer than the costs of disclosure and, as a result of the perceived psychological distance 2) people would construe the benefits on a lower, concreter level than the costs of disclosure.
In the first part of the study, psychological distance was measured for two different dimensions, namely hypothetically distance and temporal distance. Participants indicated possible advantages and disadvantages of sharing personal data online, in three different typical common data sharing situations. They indicated their perceived distance for the two dimensions, and the level of construal of their self-indicated advantages and disadvantages was measured.
The first finding of this study was that the benefits of sharing personal data online were indeed perceived as psychologically closer than the costs of sharing personal data online for both the hypothetical as the temporal distance dimension. As expected, participants indicated their perception of hypothetical and temporal distance and it was showed that the benefits were significantly indicated closer than the benefits. This provides initial evidence that the benefits of sharing personal data online is seen as psychological closer than the costs of information sharing activities.
In line with Construal Level Theory (Liberman & Trope, 1998; Trope et al., 2007; Trope & Liberman, 2010), evidence was found for the proposition that the benefits of data disclosure would be construed in a more low-level and concrete way, and the costs in a more abstract and high-level way. In this, perceived psychological distance of benefits and costs is thus indeed an important determinant of the kind of construals that are used by consumers to represent these costs and benefits, and these construals can be seen as consequences of the perceived psychological distance. This
finding is in line with previous research (Liberman & Trope, 1998; Trope & Liberman, 2003; Trope et al., 2007; Trope & Liberman, 2010),
If these findings are combined with previous research suggesting that people undervalue negative aspects with psychological distance and the concept of delayed gratification (Trope et al., 2007), the so called privacy paradox can be explained. As mentioned in previous research, people do hold genuine privacy concerns (Hagel & Rayport, 1997; Phelps, Nowak & Ferrell, 2000; Zimmer et al., 2009). The inconsistency between these concerns and their online information sharing behavior exists, because people perceive the benefits as more close, and as a result, consumers prefer these short term benefits over the long term ones. Furthermore, if people know what the risks of their sharing activities are, they construe these events as more abstract, which leads to a decreased ability of assessing the risks. Thus, the genuine privacy concerns are at the moment of the information sharing decision not the most decisive ones, but the benefits of sharing personal data are.
Besides, an unexpected finding of this study was that there was a significant difference in the perceived psychological distance across different situations. Participants perceived more hypothetical and temporal distance in the ‘fitness-app’ situation then in the ‘cookies’ and the ‘Facebook log-in’ situation. A reasonable explanation of this finding could be that the ‘cookies’ and the ‘Facebook log-in’ situations are more familiar to participants then the ‘fitness-app’ situation and they perceive this therefore more close as a whole.
5. STUDY II
5.1 METHOD 5.1.1 Participants
200 people participated in the second study. People where approached via e-mail and online social networks (Facebook and LinkedIn) with an invitation to participate in the study (Appendix II). An online link was provided, which redirected them to the online survey. From those 200 initial participants, 162 completed the survey (81% completion rate) and therefore only their data was used. Of this sample, 23,5% were male (N=38) and 76,5% were female (N=124). Participants were on average 28 years old. The majority of the participants completed a Bachelor’s degree (25,9%, N=42) or Master’s degree (70,4%, N=114) and it can thus be concluded that the participants were highly educated.
5.1.2 Design
An experimental design was used to test the second hypothesis. The main idea was to create a situation where participants that were manipulated for life goal and psychological distance had to choose whether they would share their data online or not and why they would do this. This provided a 3 (high-level life goal, low-level life goal, control group) x 2 (close benefits & distant costs, close costs & distant benefits) between subjects factorial design.
5.1.3 Manipulations, stimuli and measurements
Life goal. To manipulate people’s life goal, the mind-set induction
manipulation previously used by Freitas, Gollwitzer & Trope (2004) and Fujita, Trope, Liberman & Levin-Sagi (2006) was used. Participants were randomly assigned to one out of two manipulations or to the control group, which were not manipulated on their life-goal at all. First, participants in the low-level and high-level life goal group were asked to read a short introduction. This introduction was about participating in a psychological experiment and participants read about ‘how’ or ‘why’ people do things in life. Previous researchers imply that ‘why’-questions can be related to high-level construals and ‘how’-questions can be related to low-level
construals (Vallacher & Wegner, 1987). Therefore, participants assigned to the high-level life goal condition were confronted with ‘why’-questions, whereas participants assigned to the low-level life goal condition were confronted with ‘how’-questions in considering their ‘health improvements’ (Freitas, Gollwitzer & Trope, 2004; Fujita, Trope, Liberman & Levin-Sagi, 2006).
Afterwards, depending on the group to which participants were assigned, some questions were asked to make participants more aware of the high-level or low-level way of thinking (Freitas, Gollwitzer & Trope, 2004). To that end, participants assigned to the abstract/high-level condition were asked to write down three ways in which ‘improving and maintaining health’ could help meeting important life goals. Afterwards, they were asked to indicate how much they thought that the indicated ways would help to meet their life-goal of ‘improving and maintaining health’. Participants assigned to the low-level, concrete condition were asked to write down three means of how they could ‘improve and maintain health’. Afterwards, they were asked to indicate how much they thought that engaging in these means would help them in ‘improving and maintaining health’.
Then, participants were asked to think again about the activity ‘improve and maintain health’. Participants in the low-level life goal condition were asked ‘how’ they would do this, where after their answer was merged in the next question and again was asked ‘how’ they would achieve this, where after their answer was again merged in the next question and again was asked ‘how’ they would do that. Participants in the high-level life goal condition were asked to perform the same task, but in this condition ‘how’ was replaced for ‘why’.
Psychological distance. Psychological distance was manipulated by changing
the distance in which events were perceived. While hypothetical distance was kept constant, temporal distance was manipulated. This was done in two conditions; the first condition contained close benefits and distant costs; it was stated that if participants would share their personal data, they could use the app immediately, whereas there was a chance that their data were sold to a third party some time later. The second condition contained close costs and distant benefits. In this condition, it was stated that if participants would share their personal data, they could use the app later, whereas there was a chance that their data were sold to a third party immediately.
Reliance on benefits/costs. Participants were asked how likely it was that they
would share their personal data online in the particular situation that was displayed. This was measured with the answer on the single question “How likely is it that you would choose to share your personal data?” on a scale from 1 (not likely at all) to 5 (very likely). The likelihood to share data was used as a proxy to indicate the reliance on costs or benefits. It is assumed here, that if people choose to disclose data, they rely on the benefits of disclosure, and if they choose not to disclose their personal data, they rely on the costs of disclosure.
Manipulation check. A manipulation check was conducted for the life-goal
manipulations. This has been done by using an adapted version of Vallacher and Wegner’s (1989) study, which was previously used by Liberman & Trope (1998). Originally, the questionnaire of Vallacher & Wegner (1989) contained 25 activities; for the current research five activities were chosen which were expected to be most easily imaginable for the potential participants. In this questionnaire, participants were asked to match a given occasion with either a high-level (why) or low-level (how) alternative for this activity (Liberman & Trope, 1998). People in the high-level life goal condition were expected to rematch with high-level alternative whereas participants in the low-level life goal condition were expected to rematch the activities with low-level alternatives. The Cronbach’s Alpha was calculated to test the reliability, α = .65. No items could be removed to calculate a higher reliability scale. Therefore, these questions were decided to be reliable and could be used in the manipulation check.
5.1.4 Procedure
An online survey tool was used to conduct the experiment. Each participant received an online invitation for the survey, which provided a link to start the survey. First, the introduction of the survey was showed to the participants, which contained a description of the context of the research, an indication of the amount of time it would take to complete the survey, a description of the fact that there were no answer is right or wrong, and the confidentiality of the answers was emphasized. Furthermore, the respondents were thanked for their participation in the research.
Participants were randomly assigned to one out of six manipulation possibilities as displayed in table 4. First, participants were assigned to the life-goal manipulation, which could be a high-level life goal or the low-level life goal
manipulation or they could be assigned to the control group. If people were assigned to the high-level or low-level life goal manipulation, they were exposed to the
manipulation check afterwards. Table 4: Overview of the manipulations
After participants were placed into in the low-level, the high-level life goal group or the control group, they were randomly assigned to the psychological distance manipulation. This was done by exposing them to a typical everyday situation in which they were asked to share their personal data online, and possible consequences of data sharing were showed. These consequences where manipulated for psychological distance, and people could be assigned to either the close benefits and distant costs condition or to the distant benefits and close costs condition.
Participants were asked how likely it was that they would enclose personal information online in this situation. At the end of the survey, some demographic questions were asked. These included gender, age, and education level.
5.1.5 Data analysis
In order to analyse the data, SPSS Statistics was used. First, the raw data were prepared for the analysis. Missing values and counter-indicative variables were checked. For the manipulation check of life goal, the high-level and low-level answers were mixed across the first or second answer options. Therefore, the values of this manipulation check were not all representing the same life goal (either high- or low-level). Thus, the manipulation check on life goal was found to be partly counter-indicative and was reversed recoded where needed. Furthermore, the education
Life goal Ps yc ho lo gi ca l d is ta nce
High-level Low-level Control group Close benefits &
distant costs
1. High-level life goal and close benefits & distant
costs
3. Low-level life goal and close benefits &
distant costs
5. No life goal and close benefits
& distant costs
Distant benefits & close costs
2. High-level life goal and distant benefits & close
costs
4. Low-level life goal and distant benefits &
close costs
6. No life goal and distant benefits & close
