Amsterdam Business School, University of Amsterdam Msc. in Business Administration – Digital Business Track
Trust and social distance in the sharing economy: How to stimulate sharing intentions in a ridesharing context.
Author: Marsha de Jongh Student number: 11393394 Date of submission: 26-01-2018 Supervisor: N. Stofberg
Statement of originality
This document is written by Marsha de Jongh 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 content
Statement of originality ... 2 Table of content ... 3 List of Tables ... 5 List of Figures ... 5 Acknowledgements ... 6 Abstract ... 7 1. Intoduction: ... 8 2. Literature Review ... 13 2.1. What is the sharing economy? ... 13 2.2 Ridesharing ... 15 2.2.1 About Ridesharing ... 15 2.2.2. Barriers to ridesharing ... 17 2.3 Trust ... 18 2.4 Reputation systems ... 19 2.5 Social distance and conformity ... 20 2.6 Using SNS in the sharing context ... 21 2.7 Identity and social distance ... 24 3. Conceptual model ... 26 4. Methodology ... 29 4.1 SE platform ... 29 4.2 Research design: ... 30 4.3 The two different manipulations ... 32 4.4 Data collection: ... 32 4.5 Sample: ... 33 4.6 Pre-tests ... 33 4.6.1 Explorative pre-test ... 34 4.6.2. Pre-test 2: Qualitative pre-test ... 34 4.6.3: Pre-test 3: Quantitative pre-test ... 36 4.7 Measure development ... 37 4.7.1. Independent Variable Sharing intentions ... 38 4.7.2. Mediator Trust ... 38 4.7.3. Mediator Social distance ... 39 4.7.4. Other control variables ... 40 4.8. Data analysis: ... 41 4.8.1. Independent Variables ... 41 4.8.2. Control variables ... 41 4.9. Believability Conditions ... 42 5. Results ... 43 5.1 Correlations ... 43 5.2 Hypothesis testing ... 45 Mediation model 1: IV Friends of friends with mediator trust ... 45 Mediation model 2: IV events and mediator trust ... 47 Mediation model 3: IV Friends of Friends and mediator social distance ... 48 Mediation model 4: IV Events and mediator social distance ... 49 6. Discussion ... 53 6.1 Summary results ... 53 6.2. Discussion of the results ... 53 6.2.1. The friends of friends manipulation ... 536.2.2. The events manipulation ... 54 6.2.3. Trust on sharing intention ... 55 6.2.4. Social distance on sharing intention ... 56 6.3. Implications for theory ... 56 6.4. Managerial implications ... 57 6.5. Limitations and future research ... 58 7. Conclusion ... 60 8. References ... 61 Appendix A: The application ... 69 A-1 Original application: ... 69 A-2 Manipulation Events ... 71 A-3 Manipulation Friends of Friends ... 74 Appendix B: Survey questions ... 77
List of Tables
Table 1: Overview hypotheses………..……27
Table 2: Distribution experimental groups……….….… 31 Table 3: Pre-test overview……….…….…37 Table 4: Abbreviations………..…..41 Table 5: Believability conditions...……….…42 Table 6: Correlation matrix………45 Table 7: Overview hypothesis results………...51
List of Figures
Figure 1: Sharing economy and related forms of platform economy ………15 Figure 2: Example Skjutsgruppen application………..…23 Figure 3: Conceptual model………26 Figure 4: Adaptation based on qualitative pre-test: Before and after ………..…………...35 Figure 5: DTVP………...……….36 Figure 6: Inclusion of the other in the self scale ………..………..39 Figure 7: Overview correlations mediators with the dependent variable ..………..44 Figure 8: Mediation model 1………..46 Figure 9: Mediation model 2………..48Figure 10: Mediation model 3………..49
Figure 11: Mediation model 4………..50
Acknowledgements
First and most of all I would like to thank my thesis supervisor Nicole Stofberg. Her enthusiasm regarding the research topic really made the entire process a fun experience; especially after some setbacks her positivity was really helpful. Nicole provided me with valuable advice and I learned a lot from her. Of course I also want to thank all the respondents who completed the survey, and all the people who helped with collecting the respondents. Also I want to thank Abel for the collaboration, sharing their insights and assisting in the adaptation of their booking application.
Abstract
The sharing economy can be considered a hot topic – not only for its social and economic benefits- but also its implications for sustainability. In some sectors it has already flourished. However the act of sharing brings some challenges. One sharing sector that still has a lot of growth potential is the ride-sharing sector –and is believed to allow a reduction by 40% in co2 emissions- if adopted widely. However, to data, the market for ride-sharing is still very small, despite large technological advancements that make it very scalable. Consumers often have barriers for participating in ride-sharing, based on the perceived risk of participating. This study will explore ride-sharing, and how these perceived risks can be reduced using the variables trust and social distance, and how this will influence sharing intentions. An experiment will be conducted exploring how and if these variables can be influenced through manipulating the sharing platform. Also the influence of trust and social distance on sharing intention will be explored. The findings of a survey completed by 153 Dutch respondents confirmed the positive influence of trust in other users on sharing intentions. No effect was found of the experiment on participants their levels of trust and social distance. However it does set an opening to explore different ways to approach the matter in the future. Keywords: Sharing economy, Sharing economy platform, Peer to peer, Social distance, Trust users, Trust Platform, Sharing Intention, Ride sharing.1. Intoduction:
The past few years the sharing economy has grown from being a niche market trend, to being a large-scale industry involving millions of users across countless platforms. Sharing has the ability to save or make money, provide a novel consumer experience, reduce ecological footprints and strengthen social ties (Schor & Fitzmaurice, 2015). The most commonly known sharing platform is Airbnb, a peer-to-peer online accommodation sharing-platform. Today the website has over 3 million worldwide listings, accommodating more then 160 million guests (Airbnb, 2017). Airbnb is not the only sharing platform by far. Many platforms have made a rise across varying sectors. PwC estimates that the sharing economy is currently growing roughly 35% a year (PriceWaterhouseCoopers, 2016). The fast growing sharing economy has the potential to disrupt existing industries as sharing can shift the economy from individual ownership to shared ownership, resulting in less purchases (Belk, 2014b). Whereas a few years ago nobody had even heard from the sharing economy, today it seems to become a part of society, which will be there to stay (Botsman & Rogers, 2010)
Whilst sharing in the accommodation sector has arrived, in many sectors sharing remains a largely untapped potential (Belk, 2014b). One specific sector that still has tremendous growth potential for sharing is the peer-to-peer transportation sector.
As cities are growing urban transportation systems are under increasing pressure (Cohen & Kietzmann, 2014). An efficient ridesharing system will result in less traffic, cleaner air and shorter commutes. Ride sharing Research in New York has shown that if passengers tolerate less then 5 minutes delay per trip nearly 95% of the trips could be shared, reducing total travel time with 40 percent. This could result in lower costs, less congestion and less CO2 emission (Alonso-Mora et al., 2017).
Especially today ridesharing is a very scalable solution. In the past carpooling systems have risen and fallen. However these carpooling systems were often limited to the consumers’
direct network, and were not considered to be very flexible. The big difference between traditional carpooling systems and today’s ridesharing platforms is based on new digital solutions. The sharing economy is mediated by online platforms, expanding sharing possibilities to a much broader reach than offline sharing initiatives. You can share with people whom you have never met before. Social media and information technology are thus the drivers of the sharing economy (Heinrichs, 2013). Not only do digital solutions increase the reach of sharing, they also decrease transaction costs, allow for crowdsourcing,
reputational information and reputation, which all decrease the risks involved in sharing, making the sharing economy more attractive then before (Schor & Fitzmaurice, 2015).
Considering all the benefits and possibilities, the peer-to-peer transportation sector remains fairly small in practice. This raises the question how growth can be achieved, and what the motivations and barriers of consumers are. Many scholars have argued economic motivations to be the main driver in the sharing economy (Möhlmann, 2015). This suggests people are only willing to share with strangers to maximize their own utility (Möhlmann, 2015, Eckhard and Bardhi, 2015). Based on this growing the sharing economy could be achieved by simply lowering the costs and consequently increasing the benefits (Möhlmann, 2015, Eckhardt and Bardhi, 2015). Many platforms have implemented this advice, resulting in a rise of sharing platforms that only focus on economic benefits, completely ignoring the more social benefits. However following this price based advice, ridesharing platforms should already be growing more then they do now, as they often offer much cheaper options
compared to non-ridesharing. An example of a ride-sharing platform focusing on monetary benefits is Uberpool. They promote their service with slogans like “Carpool and save money” (Uberpool, 23-01-2017). Other ride sharing platforms did implement more social incentives. Blablacar instead puts their main focus on the social aspect, the first thing you will read on their website is “Carpooling in good company” (Blablacar, 23-01-2018). Also Airbnb as one
of the most successful sharing platforms does not stress the economic benefits but instead focuses on their ‘unique experience’ and ‘human connections’. Some scholars have also argued the importance of social and environmental motivations (Hamari, Sjöklint , & Ukkonen, 2015; Botsman & Rogers, 2010). This leaves the question what approach is the most beneficial one. Are we looking in the right direction?
The largest inhibitors to ridesharing are security perceptions concerning trust
(Furuhata et al. 2013) demonstrating the crucial role trust plays to overcome barriers to shared mobility. The matter of trust has been widely researched (Hawlitschek, Teubner & Weinhardt, 2016). Trust is important, as it is necessary for peer-to-peer transactions to occur. This is amplified by the fact that most transactions start online, without any face-to-face contact, increasing the risks involved in sharing (Schor & Fitzmaurice, 2015). Among these perceived risks are safety, inconvenience, awkwardness, availability and social exclusion (Nielsen et al., 2015). The sharing economy mitigates our inherent stranger-danger bias by facilitating trust building between individuals who interact through a platform (Möhlmann & Geissinger, 2018). Currently most trust building mechanisms are based around reputation systems. These however are often argued to be unreliable (Hu, Zhang, & Pavlou, 2009). As trust is necessary for sharing to happen (Hamari, J., Sjöklint, M., & Ukkonen, A., 2016), this raises the question how trust can be facilitated outside reputation systems.
This study asserts that a possible way in which the stranger-danger bias may be reduced, is by not making other users as ‘strange’ anymore by decreasing social distance. Social distance is the degree to which feelings of closeness and connectedness is present between individuals (Aron, Aron, & Smollan, 1992). Chaube et al. (2010) argue a close relationship is a key factor of successful ride matching. 7% would accept rides from strangers, whereas 98% would accept a ride from a friend and 69% from a friend of a friend and 69% from members of their community. One way to simulate close relationships on platforms
between individuals whom have never met, is through the use of social networking sites. Social networking sites offer the possibility to show common friends between strangers. As 69% indicated they would share a ride with a friend of a friend (Chaube, Kavanaugh, & Perez-Quinones, 2010), social networking sites could be used to decrease social distance, and in turn increase people their willingness to share their ride. Social networking also allows for verification, increasing trust (Teubner & Hawlitschek, 2017).
Furthermore common interests could also have a positive influence on sharing
intention. Common interests make people seem more similar, thus decreasing social distance (Tesser , 1988). Also people have more trust in people they are closer to or familiar with (Staub, 1978).
The sharing economy remains a ‘hot topic’. However for it to keep growing it is important to explore what factors influence whether or not people are willing to participate. Especially the transportation sector could benefit greatly form sharing. Currently, we don’t like the idea of riding our cars let alone our rides and whilst the potential benefits of ride-sharing are lauded to date in practice we opt to sit in our cars solo. Even though no consensus exists on how to best approach the issue, it does become clear the risks and discomforts associated with sharing our rides with complete strangers could prevent individuals from participating. One possible solution to this is increasing the level of trust. Also at the core of the problem of risk is the fact that the sharing parties are not familiar with each other. This could also be overcome by making sure social distance in sharing situations is decreased. However the question remains how and if these two variables, trust and social distance, can be influenced through the sharing platform in use. Based on this the following research question will be explored:
Can ride sharing platforms increase consumers’ willingness to share a ride with unknown others by increasing - their level of trust in- and reducing the perceived social distance between strangers, by highlighting communalities in terms of shared friends or interests?
Answering this question will result in several contributions. First it will explore if and how trust and social distance influence sharing intention. Second the experimental setting of this research will explore how trust and social distance can successfully be influenced through integrating different features in sharing applications. To discover which features could influence these variables successfully this study will involve a vignette experiment in which an original car-sharing platform will be manipulated in two different ways. Each manipulation will attempt to influence levels of trust and social distance. The first manipulation will
integrate a social networking site to show prospective participants the friends they have in common on with other users of the platform, and the second will include a common event to which the ride will be shared. The specifics of these manipulations and why they were chosen will be discussed in the following chapter. After this a conceptual model will follow. This model will be tested through the experiment, of which details will be discussed in the methodological chapter. After this the results of the experiment will be analyzed and discussed, after which a conclusion will follow in which the research question will be answered.
2. Literature Review
2.1. What is the sharing economy?
Sharing economy is a relatively new concept, and its’ definition is not always clear. The sharing economy includes digital platforms constituted by economic activity, which is peer-to-peer or person-to-person (Hamari, Sjöklint , & Ukkonen, 2015; Schor & Fitzmaurice, 2015). The sharing economy allows individuals to make use of their underused assets, creating increased efficiency (PriceWaterhouseCoopers, 2016). Belk (2013) describes two commonalities in all sharing consumption practices: “1) their use of temporary access
non-ownership models of utilizing consumer goods and services and 2) their reliance on the Internet, and especially web 2.0, to bring this about (Belk, 2013)”. The sharing economy is
widely applicable, and is currently happening across many sectors. The five key sectors of the sharing economy are: peer-to-peer accommodation, peer-to-peer transportation, on-demand household services, on demand professional services and collaborative finance (PriceWaterhouseCoopers, 2016).
What distinguishes sharing economy from other markets is the word ‘sharing’. The concept of ‘sharing’ has been around for hundreds of years, forming the basis of human economic activity. It is a powerful force of solidarity within communities. Sharing today is still a widespread phenomenon, from sharing within a family to sharing a library with your fellow citizens (Martin, 2016). Belk (2007) describes sharing as:
“the act and process of distributing what is ours to others for their use and/or the act and process of receiving or taking something from others for our use.”
Distinction can be made between two prototypes of sharing. The first type is sharing-in, which is an act of unconditional caring, which is mostly seen within families. This type of sharing is free from obligations. The second type is sharing-out, which is the more efficient use of joint possessions within communities (Belk, 2007). The sharing economy however
extends even sharing out, entering a new undefined reach of sharing. This new reach of sharing can also be classified as ‘stranger sharing’ (Schor, 2014). The sharing economy has created a new form of sharing which extends beyond the traditional boundaries of family, friends and community (Martin, 2016, Botsman, 2013).
Agyenman and MacLaren (2017) argue a big part of the sharing economy is about commercial sharing, which is when collaborative consumption takes place in return of monetary payment. Another side of the sharing economy is communal sharing, where goods, services and skills are donated, swapped or traded for free, or in some cases a alternative medium of exchange, like time (Agyenman and MacLaren, 2017). Although there are sharing platforms where communal sharing is still the basis, most sharing platforms are currently underlined by economic exchange, and are thereby commercial sharing platforms. Still this may not be confused with other forms of economic exchange. To create a comprehensive definition of the sharing economy, Frenken and Schor (2017) have created a clear framework of what can be considered as being part of the sharing economy. The overview of this framework can be found in figure 1. First sharing is about consumers to consumer platforms, and not about renting, which is business to consumer. Renting is an example of access to ownership, which means a consumer gains access while the provider retains ownership (Bardhi and Eckhardt, 2012). An example of ushc a business model is Greenwheels. Greenwheels offers consumers temporary access to their cars for a certain fee (Greenwheels, 2018). Second consumers provide temporary access without transfer of ownership. Otherwise it would be considered as second hand economy. Third sharing economy is about sharing physical goods. The exchange of services is not included, and can be categorized under the on-demand economy. In short sharing economy is distinguished by three characteristics: consumer to consumer (c2c), temporary access and physical goods (Frenken and Schor, 2017). Based on this the definition maintained in this study is: “consumers granting each
other temporary access to under-utilized physical assets (idle capacity), possibly for money” (Frenken et al. 2015). Figure 1: Sharing economy and related forms of platform economy (Frenken et al., 2015)
2.2 Ridesharing
2.2.1 About RidesharingRidesharing exists when two or more trips are executed simultaneously, in a single vehicle. Adapted from the definition given by Amey, Attanucci & Mishalani (2011), ridesharing will be defined as following:
“A single, or recurring rideshare trip with no fixed schedule, organized on a one-time basis, with matching of participants occurring as little as a few minutes before departure or as far in advance as desired before a trip is scheduled to take place”
Within the sharing economy ridesharing has a lot of potential to grow. Ridesharing initiatives have been around long before the rise of the sharing economy. As an example in the past many carpooling initiatives have come to pass. However just like with other sharing
practices, carpooling was often limited to a persons own network (Chan and Shaheen, 2012), and were characterized as relatively inflexible, long term arrangements (Amey, Attanucci, & Mishalani, 2011).
Today new technological possibilities have raised new opportunities for ridesharing platforms. Real-time ridesharing is now a possibility. These systems use GPS technologies on smartphones to organize ride sharing real-time. Requesters can organize a trip right before departure time. Ride matching software will then match riders and drivers with similar trips and notify each party (Chan & Shaheen, 2012). In short, through new technologies ridesharing has become much more flexible compared to the past. This greatly increases the attractiveness of ridesharing for consumers (Amey, Attanucci, & Mishalani, 2011).
Ridesharing offers many benefits compared with traditional transportation via private cars. Ridesharing results in an increase of car seat occupancy, which in turn will result in less cars being used to transport the same amount of people (Morency, 2007). Environmentally it reduces energy usage, gas emission, traffic congestion and parking demand. Also it reduces travel costs by sharing them with fellow passengers (Chan & Shaheen, 2012; Amey, Attanucci, & Mishalani, 2011).
Currently a personally owned car is perceived as the most appealing mode of transport (Nielsen, Hovmøller, Blyth, & Sovacool, 2015). Noland et al. (2006) argue that policies to increase carpooling are actually the most effective strategy to reduce energy consumption, besides banning driving altogether. The perfect solution could be changing people’s habits, making them use the public transportation system. However convincing people to change their transportation mode can be difficult. Compared to private cars, public transportation does not offer the same level of comfort. Instead ridesharing poses as a solution offering the perfect middle ground between private cars and public transportation (Morency, 2007).
Ride sharing usually entails a privately owned car in which the owner has a specific trip planned. The driver can accept giving a ride if the other person is going to a place close to his destination. Another form of ride sharing is ride sharing in taxis. Taxi drivers do not have time restrictions or a specific route. Regular taxis provide passengers with convenient, comfortable and flexible trips. However taxi occupancy is often low, and taxis make considerable demands on the road resources. Ride sharing could be a solution to the low operational efficiency of taxis. Taxi is an example of the Demand Responsive Transit (DRT) operating mode (Lin, Li, Qiu and Xu, 2012).
In case of a regular taxi service the consumer creates capacity by ordering a taxi trip from A to B which otherwise would not have been made. This case is an example of the on-demand economy. Ride sharing platforms however can be distinguished from regular taxi platforms as consumers book a seat in the car on a trip which would have been made anyway, with or without the seat being occupied (Frenken and Schor, 2017). Hence, once a ride is being shared in a taxi, we can speak of peer to peer sharing, whereas this is not the case if we consume a ride on our own (without sharing it).
Sharing a ride in a taxi has benefits compared to regular carpooling where the driver is another consumer. Ride-sharing in taxi’s is more flexible as the driver is not bound to a specific route, making taxi sharing a good solution for ride-sharing.
2.2.2. Barriers to ridesharing
Besides the benefits as described in the previous paragraph, there are several behavioural barriers to increasing ridesharing.
Eckhardt and Bardhi (2015) argue strongly that consumers are more interested in the economic benefits then building social relationships, saying companies should focus on highlighting these economic advantages. However trust is often argued to be the biggest deterring factor of sharing. This suggests more emotional approaches could be more
beneficial. Research has shown that for example emotional appeals trigger more positive responses compared to rational appeals (Fox & Amichai-Hamburger, 2001). Especially trust is an interesting variable when exploring sharing platforms, as the presence of trust is necessary for sharing to happen (Belk, 2014).
2.3 Trust
The presence of trust can be a key determinant whether or not sharing intention is present in a P2P context, such as when strangers have to share a ride (Hamari, Sjöklint , & Ukkonen, 2015). Trust can be defined as “a disposition to engage in social exchanges that involve uncertainty and vulnerability, but that are also plenty rewarding” (Bicchieri, Duffy, & Trolle, 2004). Trust is the connecting link in the sharing business (Belk, 2014). Trust is a key mediating mechanism for effective marketing. Trust has influence on consumers’ attitude, risk perception and behavioural intent, which results in willingness to participate (Bart, Shankar, Sultan, & Urban, 2005).
The social aspect of the sharing economy is what distinguishes it from other sectors, and can definitely be a success factor. Studies have identified social motivation as an important driver for sharing practices (Bucher, Fieseler & Lutz, 2016). However the personal contact involved in sharing can also be challenging, namely based on a lack of trust. The exchange of services and goods between parties who do not know each other carries a greater risk then an exchange between long-term offline relationships (Lauterbach et al. 2009) Trust is important in situations of risk, uncertainty and interdependence (McKnight & Chervany, 2001). These three are all very present in the sharing economy. The nature of sharing transactions further complicates the presence of trust. Traditional online transactions, for example in the access-economy, only involve two parties, namely the seller and the consumer. Sharing platforms however include three parties in every transaction: two peers and the platform involved. first consumers must trust in the platform offering the service. Second the
consumer must trust the peers whom they are sharing with (Möhlmann, 2016). Based on this a distinction can be made between trust in peers and trust in the platform. Both types of trust have to be present on sharing platforms. Trust in peers is important because sharing will involve interactions with peers participant did not previously meet, involving certain levels of risks (Lauterbach et al. 2009) Trust in the platform is also important as the platform mediates the relationship between the peers. Trust in the platform is a form of institutional trust. This type of trust is based on firm or intermediary specific characteristics (Zucker, 1986). Trust in other users and trust in the platform will both be explored. For the bigger picture these types of trust will be mentioned in a combined variable of trust. However during the hypothesis testing both forms will be split and measured separately.
H1: Increased levels of a) platform trust and b) peer to peer trust will lead to higher sharing intention of prospective users.
2.4 Reputation systems
Considering the importance of trust in the sharing economy, facilitating trust is a key challenge of sharing platforms. To increase levels of trust the most commonly used mechanisms are reputation systems. Reputation systems are large scale word-of-mouth networks (Dellarocas, Dini, & Spagnolo, 2006). Within these systems information about the user is collected and made visible to other users. This allows users to judge others trustworthiness based on their behaviour and feedback from others, and stimulates users to be honest. A bad reputation will prevent positive interactions in the future (Resnick, Kuwabara, Zeckhauser, & Friedman, 2000), reducing the probability of opportunistic behaviour (Dellarocas, Dini, & Spagnolo, 2006).
Literature has however questioned the effectiveness and reliability of reputation systems. Research has shown a consistent positivity bias in online reviews. A ‘J-shaped’
review distribution is often present. This means the majority of reviews express consumers’ satisfaction, while there is a smaller number of reviews which are all very unsatisfied consumers, suggesting consumers with moderate views are less likely to report their ratings. This results in a low percentage of average reviews (Hu, Zhang, & Pavlou, 2009). The ‘J-shaped’ reviews effect is present in many review sites including Amazon (Hu, Zhang, & Pavlou, 2009) and Tripadvisor (Feng, Xing, Gogar, & Choi, 2012). Airbnb for example has an average rating of 4.7 of a 5 point scale, and 95% of the Airbnb properties were rated either 4.5 or 5 stars, listings under 3.5 stars being extremely rare (Cansoy & Schor, 2016). Also reputation systems decrease the likelihood participants will form interpersonal relationships (Parigi 2014, Schor, 2014). Considering reputation systems seem to lack effectiveness, other ways of stimulating trust in digital environments may be preferable.
2.5 Social distance and conformity
Besides trust social distance also has influence on consumers their likelihood to share. Research suggests social distance and the degree of identification and communication between parties are also significant factors in building trust. When social distance decreases, the other is no longer an unknown individual, but rather an ‘identifiable victim’ (Thomas C. Schelling, 1968). Construal level theory (CLT) explains that any object can be mentally represented in different ways (Liberman, Trope, & Stephan, 2007; Liviatan, Trope, & Liberman, 2008), CLT proposes individuals use higher levels construal for distant future and lower level construal for near future events. High level construal are more simple and decontextualized representations, while low level construal’s tend to be more concrete and include more contextual and incidental features of events (Trope & Liberman, N., 2003).
Interpersonal similarity is a form of social distance, which results in similar others being perceived as socially closer to oneself then dissimilar ones (Tesser , 1988). Liviatan,
Trope & Liberman (2008) argue people construct different representations of similar and dissimilar individuals when provided with the same information. In turn these representations influence an individuals judgements on the actions of similar and dissimilar people.
Social distance describes the degree to which feelings of closeness and connectedness between individuals (Aron, Aron, & Smollan, 1992). As perceived social distance influences how consumers view their relationships with peers, perceived social distance also has an influence on intentions to share with others (Stephan, Liberman, & Trope, 2011). Rachlin & Jones (2008) for example describe the closer people are, and thus less socially distant, the more likely they are to show social behaviour. Considering the social aspect is quite present in the sharing economy, social behaviour is necessary to participate.
Humans are social beings that interact and share knowledge with their network. As a result decisions may be influences by choices made by other members of the network. People may change their decisions to match the attitude, beliefs and norms to fit their social network. This behaviour is also classified as conformity (Cialdini & Goldstein, 2004). Social distance influences how people conform to behaviour of others (Akerlof, 1997).
Social distance theory suggests people will be more likely to share when similar people are sharing too. Also low social distance will make people more likely to show social behaviour. This in turn will increase people their likelihood to share with others, as this is also considered pro-social behaviour.
H2: A decrease in perceived social distance will result in higher sharing intention.
2.6 Using SNS in the sharing context
A lot of research has shown the importance of trust and motivation in the sharing economy. However it is not clear how these can be influenced by different platform positioning
strategies. This study will research how managerial choices can influence both trust and motivation within sharing platforms.
Concerning trust it is still unclear how to achieve it in the most effective way. SNS offer a good solution, as it offers the opportunity to show how many friends people have in common. As Chaube et al. (2010) have illustrated, 69% would consider sharing a ride with a friend of a friend. This suggests another way to build trust is to associate acquaintances in efficient ways (Furuhata et al., 2013).
Some platforms have implemented social media to match rides between friends or acquaintances, through which these platforms hope to build trust among participants. Social media are Internet based applications that allow the creation and exchange of user generated content. Part of social media are social networking sites (e.g. Facebook, LinkedIn) (Mangold & Faulds, 2009). More then 70% of online users between 18 and 29 years old use SNSs (Lenhart, Purcell, Smith, & Zickuhr, 2010).
An example of this is Zimride. This platform has partnered with US and Canadian colleges, universities and companies with each their own network of users. This approach makes sure to limit the amount of social distance, as all users share a college, university and so on (Zimride, 2017). Research has shown only 3% to 10% of shared rides occur between strangers, with the rest occurring between family, friends and acquaintances. This low percentage is caused by stranger danger, which results in a low interest in sharing rides between strangers, because of safety concerns (Amey, Attanucci, & Mishalani, 2011). SNS has the possibility to decrease stranger danger, and show which connections other participants have. This way using SNS also has the ability to decrease social distance, and also increase trust. The Swedish ride-sharing company Skjutsgruppen already included SNS in their application, showing the connection between users through common friends. An example of Skjutsgruppen can be seen in figure 2.
.
Figure 2: Example Skjutsgruppen application
Verification and signalling also play an important role in establishing trust. This can be done through displaying a phone number, and signalling popularity through social accounts. Also self-presentation through profile pictures impacts trust (Teubner &
Hawlitschek, 2017). These can all be accomplished through implementing SNS. Kim, Rasouli & Timmermans (2017) found in their research people are more willing to join car-sharing platforms when more people in their network joined before.
In short increased trust and decreased social distance can be established using SNS in several ways. First of all SNS has to ability to show people which friends they have in
common. Common friends greatly increase the likelihood of sharing (Chaube, Kavanaugh, & Perez-Quinones, 2010). The expectation is this increase in likelihood of sharing can be explained by users having more trust when they have friends in common. Also SNS offers a form of verification (Teubner & Hawlitschek, 2017). Also social distance is decreased by using SNS to shown mutual friends, as the other party cannot be considered a complete stranger anymore. Based on this the following hypotheses concerning SNS are formed:
H3a: Using SNS to connect users with friends or friends of friends on a sharing platform, will increase sharing intentions.
H3b: Using SNS to connect users with friends or friends of friends on a sharing platform, will increase the amount of of a) platform trust and b) peer to peer trust.
H3c: Platform trust and b) peer to peer trust will mediate the relationship between platform condition and sharing intentions. The friends of friends manipulation of the sharing platform will increase users’ trust, which will in turn increase their sharing intentions.
H3d: The friends of friends manipulation will decrease the level of social distance. H3e: Social distance will mediate the positive relationship between the friends of friends manipulation and sharing intention.
2.7 Identity and social distance
Identity has significant influence on whether individuals perceive themselves to be socially distant or not. Social identity defines whether a person knows he belongs to a certain category or group (Abrams & Hogg, 1988). A social group consists of people with similar views on social identification and see themselves as members of the same social category. Within these groups similar people are categorized as being in-group and dissimilar people as the out-group (Turner et al., 1987). People trust people whom they feel close with, whom are similar to them, and with whom they feel familiar (Staub, 1978). This is caused by the fact that people trust those whom they share an identity with more then people whom they do not share an identity with, group membership is a strong predictor of trust (Tanis and Postmes, 2005). Considering the risks involved in participating in the sharing economy, trust remains the main deterring factor of participation in the sharing economy (Ballús-Armet, Clonts &
Weinzimmer, 2014). Overcoming a lack of trust can thus have a positive influence on sharing intention.
Considering the importance of shared identity for predicting trust, including an indicator of shared identity could increase sharing intention. Common events can be an example of a way to create a common identity. People who are interested in visiting the same event, for example because they share the same music taste, share part of their identity, this in turn decreases the level of social distance, as feelings of closeness and connectedness with others are higher (Aron, Aron, & Smollan, 1992). Also individuals trust other people more when they are similar to themselves. Thus showing common events is expected to influence both trust and social distance, and through these variables also sharing intention. The following hypotheses are formed:
H4a: The events manipulation will increase users their sharing intention.
H4b: The events manipulation will decrease users their perceived social distance. H4c: Social distance mediates the relationship between the events condition and sharing intention. The events condition will decrease social distance, which in turn will increase sharing intentions.
H4d The events manipulation will increase levels of a) platform trust and b) peer to peer trust. H4e Trust will mediate the relationship between the events manipulation and sharing
3. Conceptual model
This research will examine how social distance and trust can be influenced through different types of framing, using SNS as a trust increasing mechanism, and common events to decrease social distance. In turn it is explored if social distance and trust will influence sharing
intention. Figure 3: Conceptual model H3a Social distance Trust a) Platform and
b) peer to peer Friends of Friends manipulation Sharing Intention Events manipulation H1 H2 H3d H4b H3c H4d H4c H3e, H4c H3b H4a
Table 1: Overview Hypotheses
Hypothesis
H1 Increased levels of a) platform trust and b) peer to peer trust will lead to higher sharing intention of prospective users.
H2 A decrease in perceived social distance will result in higher sharing intention.
H3a Using SNS to connect users with friends or friends of friends on a sharing platform, will increase sharing intentions.
H3b Using SNS to connect users with friends or friends of friends on a sharing platform, will increase the amount of of a) platform trust and b) peer to peer trust.
H3c Platform trust and b) peer to peer trust will mediate the relationship between platform condition and sharing intentions. The friends of friends manipulation of the sharing platform will increase users’ trust, which will in turn increase their sharing intentions.
H3d The friends of friends manipulation will decrease the level of social distance.
H3e Social distance will mediate the positive relationship between the friends of friends manipulation and sharing intention.
H4a The events manipulation will increase users their sharing intention. H4b The events manipulation will decrease users their perceived social
distance.
H4c Social distance mediates the relationship between the events condition and sharing intention. The events condition will decrease social distance,
which in turn will increase sharing intentions.
H4d The events manipulation will increase levels of a) platform trust and b) peer to peer trust
H4e Trust in s) platform and b) peer to peer will mediate the relationship between the events manipulation and sharing intention.
4. Methodology
This paragraph will describe the applied research methods of this experiment. First the actual sharing platform on which this experiment was based will be discussed (4.1). This will be followed by a description of the research design (4.2), after which the data collection process will be described (4.3). After this the sample characteristics will be highlighted (4.4). Then the pre-test will be described (4.5). A paragraph with the measure development will follow (4.6).
4.1 SE platform
This research was originally a case study of the taxi-sharing platform Abel. Abel was willing to cooperate with this research, as they were still exploring how they could stimulate consumers to use their platform. Abel was a start-up founded in 2016. Abel had a number of cars and taxi drivers in service, and they offered taxi rides via their app.
What differentiated Abel from regular taxi platforms is the fact that customers possibly share their ride with other customers. This form of taxi sharing offers several benefits. First of all taxi sharing gave Abel the ability to offer competitive prices to their customers. Also they offered an entirely new social experience, where customers are able to meet and socialize with other customers. Lastly it is an environmentally sustainable solution. They do not only increase seat occupancy and hereby the number of rides by letting people share, they also make use of green cars.
Abel is the perfect case to study in the field of shared transportation. Because they employ their own drives, they differentiate themselves from other peer-to-peer ride sharing platforms like BlaBlaCar, where other consumers are the drivers. Still they have a much bigger ‘social’ impact then for example car sharing platforms, where peers do not necessarily interact much while sharing.
After meeting with the CEO of Abel (13-10-2017), it became clear which challenges Abel faced which might decrease participation. One of the main challenges Abel was facing is trustworthiness, transparency and privacy: people are concerned with the fact that they do not know who they are sharing their ride with. Also Social contact was considered to be both an asset as a challenge.
However during this research Abel got cancelled as their investor quit. Still as the platform is a nice example of ride sharing, this research will use their original concept like previously intended, the original Abel application is the basis of this research. However the application was adapted to a new non-existing brand. Both the logo and the colors in the application were changed to prevent respondents to have negative associated with the Abel brand, as they might be aware the company does not exist anymore. The Abel application was manipulated to four different conditions: the original condition, Friend of Friends condition and the Events condition.
4.2 Research design:
The research will be done through a vignette study and a survey questionnaire with closed questions. A vignette experiment exposes respondents to short descriptions of hypothetical situations, to elicit their judgments about these situations (Rooks, Raub, Selten & Tazelaar, 2000). The vignettes used in this study created hypothetical situations in which a consumer is looking for a ride to a certain destination. After this they were presented with a demo of the taxi booking app. Three different applications were created as vignettes, serving as the independent variable. The application was manipulated to explore the influence of the manipulations on behavior. Each respondent was allocated randomly to one of the three versions of the taxi booking application. The first version is the original application, the second and third are two different manipulations. The research has a between subject-design,
meaning the conditions were given to different groups of subjects. The distribution of the respondents across the three different groups is provided in table 2.
Table 2: Distribution experimental groups
Experimental condition N Percentage
Control group 54 35.3%
Friends of friends condition 50 32.7%
Events condition 49 32%
Total 153 100%
The goal of using an actual demonstration of the application is to make the vignette as realistic as possible. The more hypothetical a vignette is, the less likely respondents their answers will match actual behavior (Neff, 1975). Also vignettes are argued to produce unrealistic results because they are not directly comparable to real life (Faia, 1979). Showing respondents the application and making them actually navigate through it makes the experience realistic as possible, minimizing these effects.
An online survey enables large amounts of data to be collected in a standardized manner, making the data easy to analyze and interpret (Saunders, Lewis & Thornhill, 2009). The survey was used to measure the respondents their level of trust in the platform and other users, social distance, and sharing intention. Also respondents their disposition to value privacy was measured, as well as their general demographics. The goal is to explore the influence of the vignettes on the answers given in the following generic survey questions.
Screenshots of the different versions of the application used as vignettes can be found in Appendix A. The actual demo in which respondents were able to navigate through the application was made using Marvel App (www.marvelapp.com). Through coding the application was integrated directly in the Qualtrics survey. This way the respondents did not
have to click on any links to be able to see the application. By doing this it was ensured every respondents saw the application. Also it helped decreasing the chances of respondents quitting the survey before finishing it, because they were unwilling to open the link.
4.3 The two different manipulations
The two different manipulations events and friends of friends were based on several cues. First of all they were based on Abel taxi service. The CEO of Abel explained how Abel already uses the taxi service on specific events, thus making the events manipulation a very likely one. Also literature has shown common interests will increase people their sharing behaviour. Also during the first pre-test events scored high when people were asked how much they identified with people who visited the same events. During the pre-test people also felt very connected with friends of friends. More detailed descriptions of the manipulations will be discussed in paragraph 4.6.4.4 Data collection:
Respondents were gathered in several ways. First the survey was distributed across social media such as LinkedIn and Facebook. Through these platforms some respondents shared the survey on their profile to stimulate their friends to fill in the survey as well. In this case snowball sampling was applied. Also besides with direct personal networks, the survey was also shared across so called public Facebook groups. Secondly respondents were gathered in public areas such as the Amsterdam University and a shopping mall. In these public spaces people were approached asking if they could complete the survey. They could do this by filling in their mail address, after which they would receive a link in a personal e-mail. This way respondents could decide for themselves when they had the time to complete the survey.
After a week respondents also received a reminder. The sample exists entirely of a probability sample. The ideal situation for sampling would be to use non-probability sampling (Saunders et al. 2009). However due to lack of time and resources the data was collected trough probability sampling. The entire survey was in Dutch, excluding respondents who are not sufficient in Dutch. However the simulation included in the experiment was also in Dutch, so translating the survey to English was not an option.
4.5 Sample:
The total sample exists of 153 respondents (n=153). Of all respondents 39,9% is male and 60.1% is female. A wide range of ages participated in this research. Ages range from 16 to 67 years old, with an average age of 28 years old (M = 28.99, SD = 12.80, range = 51).
A big group of the respondents completed a university level education: 30.7% completed a WO bachelor, 12.4% a WO master and 1.3% a PhD. Also a big group of respondents completed a HBO education (21.6%) or MBO (8.5%). 23.5% only completed their high school education. Presumably these respondents are still in the process of obtaining a higher education degree. 2.0% only finished elementary school. The biggest group of respondents lived in Amsterdam during the survey (47.1%). 19.1% of the respondents live in the Randstad area. 17% lives in a city outside of the Randstad, and 17% lives in a village. More then half of the respondents earns less then 1500 euros a month (55.6%). The seconds largest group (22.2%) earns between 1501 and 3500 euros a month, 3.9% between 3501 and 5500 and 3.3% earns more then 5500 euros a month. 15% specified they would rather not tell their salary.
4.6 Pre-tests
The vignette part of this survey was based on the construction of three different applications. The first application is the original application of taxi company Abel, the
second and third versions being manipulations. For these manipulations the original application had to be adapted using Photoshop. Adapting an application to stimulate social distance and trust can be quite tricky. This is why several pre-tests were conducted to ensure the functionality of the manipulations. These pre-tests will be described in the following paragraph. 4.6.1 Explorative pre-test
To determine what manipulations would be fit to decrease social distance between users, a small exploratory pre-test was developed (n16). This survey suggested several groups, and respondents were asked how much they identified with each group on a scale from 1 to 10, where 1 is not at all and 10 is entirely. Based on this survey the decision to use events and friend of friends to manipulate the sharing platform to decrease social distance was made. Events had a mean of 6.7 and friends of friends a mean of 7. Based on these results and the original Abel company, the decision was made to build the manipulations on friends of friends and events.
4.6.2. Pre-test 2: Qualitative pre-test
During the design process of the Abel application several qualitative pre-tests were conducted to test the functionality of the application. Based on these comments the application was adapted several times. A concrete example of an adaptation based on the feedback received from pre-test participants in the application is illustrated in figure 4. The feedback from a respondent was: “It is not clear I am booking a ride, because it looks like it could also be for buying a entrance ticket to the event.” Based on this
to the booking button. Several similar visual and textual adaptations were made to the different screens in the demo of the application, showing it to several people afterwards to confirm it. Another comment was based on privacy, as one person commented he did not like the fact that other users or ‘friends’ could see they are using the application, especially the screen as shown in figure 5 with the locations triggered this response. Based on this the variable disposition to value privacy was included in the research. Figure 4: Adaptation based on qualitative pre-test: Before and after
Figure 5: DTVP 4.6.3: Pre-test 3: Quantitative pre-test After several rounds of qualitative pre-tests and adaptations in the application a quantitative pre-test was performed. The quantitative pre-test included four conditions. The first version of the application is the original condition, where consumers share their taxi ride, but no other stimuli are included. The second condition is the events manipulation. In this condition events serve as a common interest for the consumers involved. By emphasizing common events the assumption is consumers will feel more connected with other consumers. The third condition is the common friends condition. Here a representation of the indirect Facebook connections users share was included in the application. The idea is users will have more trust in the platform when they are presented with common friends, and they will also be more likely to feel connected with other users. The last condition was a combination of both the friends of friends
condition and the events condition, to test the interaction effect of the other 2 manipulations.
The pre test sample existed of 85 respondents, and existed of the same scales as the actual survey. An analysis was performed to check how the initial four conditions were behaving. Trust users, trust platform, social distance and sharing intention for the four conditions were compared. This was done based on the different means of the conditions, as for an actual analysis the sample size was too small to be significant. The results can be found in table 3. Based on these results the interaction condition was dropped, as it did not behave as expected. Also the description of the application in the pre-text was very socially oriented. This could possibly have biased the results, thus the introduction was changed after the pre-test to be more neutral. Table 3: Pre-test overview
Manipulation Trust Users Trust platform Social distance Sharing intention
N M SD M SD M SD M SD Condition 1 24 4.4 .630 4.7 .618 2.4 1.135 4.7 1.285 Condition 2 17 4.0 .588 4.3 .628 2.8 1.286 4.2 1.524 Condition 3 21 4.6 1.104 4.7 .871 2.8 1.286 4.8 1.288 Condition 4 13 4.5 .822 4.6 .652 1.9 .921 4.2 1.373
4.7 Measure development
For this research existing measures were used. These measures were adapted to fit the context of this particular experiment.4.7.1. Independent Variable Sharing intentions
The dependent variables sharing intentions was measured using a seven-point Likert scale developed by White et al. (2012). The scale consists of four items, including the following questions: “I would like to book a taxi with Abel”, “I would be willing to use Abel for my trip”, “I would likely make Abel one of my first choices to book a taxi” and “I would exert effort to use Abel”. The seven point likert scale ranged from “strongly disagree” (1), “neither disagree nor agree” (2) to “strongly agree” (7). The Cronbachs Alpha of the sharing intention scale is .910. Correct item-total correlations are all above .30, so all items correlate with the total score, and the Cronbachs Alpha can not significantly be increased by deleting items, thus the scale is reliable.
4.7.2. Mediator Trust
One mediating variable is trust. As trust has two different types within the sharing economy, trust in the platform and trust in peers, two different scales were used to measure each individually.
The scale used to measure trust in the platform is adapted from Gefen, Karahanna and Straub (2003) and consists of 7 items. These items are measured on a 7-point Likert scale (1=completely disagree; 7= completely agree). An example of a question used in this scale is “Blablaride is trustworthy”. The scale has a Cronbachs Alpha of .808. The Cronbachs Alpha cannot significantly be increased by removing an item, and the item-total correlation is >.30, thus the scale is considered reliable.
The scale used to measure trust in peers consists of 5 items, and is adapted from Chiu, Hsu and Wang (2006). An example of a question used to measure trust in peers is “Members of Blablaride will always keep their promises with each other”. The scale has
a Cronbachs Alpha of .812. The corrected item-total correlation is >.30 and the Cronbachs Alpha cannot be increased by removing items, so the scale is considered reliable. 4.7.3. Mediator Social distance The second mediating variable is social distance. The “Inclusion of the Other in the Self Scale” (Aron et al., 1992) was used to measure social connectedness. The scale consists
of 7 diagrams, each representing a different level of social connectedness and social distance in relations. The scale is based on a 7-point interval scale, where 1 represents the highest level of social distance whereas 7 represents the lowest level of social distance. The scale is counter indicative, so it was recoded. The scale is presented to respondents as shown in figure 6.
Figure 6: Inclusion of the other in the self scale (Aron et al., 1992)
4.7.4. Other control variables Baseded on the pre-tests the variable disposition to value privacy was measured using a scale from Xu, Dinev, Smith and Hart (2011). The scale consists of three items, measured with a 7-point likert scale (1 = completely disagree; 7 = completely agree). A example of a question used to measure disposition to value privacy is “It is very important for me to keep my information private.” The Scale has a Cronbachs Alpha of .891. The corrected Item-Total correlation is >.30, and the Cronbachs Alpha does not significantly increase by deleting items, thus the scale is reliable.
Gender, age, income level, education level, urban residence and disposition to value privacy are control variables, as these all have possible influence on the sharing economy. Education was measured using seven answering options: primary school, secondary school, MBO, HBO, WO Bachelor, WO Master and PhD. Age was measured using an open ended ratio question. Bruto income was measured as an ordinal variable with 4 answering options: below 1500 euro, between 1501-3500 euro, between 3501 and 5500 euro and above 5501 euro. As a fifth option respondents could choose not to specify their income group. Because the hypothetical situation centers around a taxi sharing application, one control question is whether the respondent often uses a taxi as a means of transportation. This could influence the outcome, as respondents who never use taxi services are less likely to have a high sharing intention.
Table 4: Abbreviations abreviation Name of the variable SI Sharing Intention TP Trust Platform TG Trust Users SD Social Distance GWH Believability DTVP Determination To Value Privacy Exp Previous experience with sharing
4.8. Data analysis:
4.8.1. Independent Variables The different manipulations (Friends of Friends and Events) were recoded into dummy variables. The first (D_FOF) represents the Friends of Friends condition versus the Events condition and the original application. Condition Friends of Friends = 1, condition Events is = 0 and the original condition = 0. Dummy variable D_EVENTS has the values events = 1, condition friends of friens = 0 and the original condition = 0. 4.8.2. Control variables Some control variables were also recoded into dummy variables for the analysis. First of all gender was recoded into dummy variable D_female, where 0 = male and 1 = female. Education was recoded into two groups, high education level = 1 and low education level = 0, the groups were seperated based on the mean level of education in this sample (M =3.98). All respondentens below the mean were coded 0 and all respondent above the mean were coded 1. The same was done for income, with low income = 0 and high income = 1, divided based on the mean (M = 2.00), after excluding the respondent who did not want to specify their income (7).
After successfully recoding these variables, a correlation analysis with the dependent variable, independent variables, mediators and control variables. After this PROCESS by Hayes (2012) is used to run a regression analysis on the conceptual model to check for any mediating effects.
4.9. Believability Conditions
The believability of the four conditions was measured using the GWH variable. In the following table the scores of the believability variable are illustrated. The GWH variable was measured with a 3 item index, based on a 7-point Likert scale. All conditions have a high mean, meaning the presented conditions were considered to be realistic by the respondents. An overview can be found in table 5. Table 5: Believability conditions MEAN SD Original 5.330 .981 Events manipulation 5.487 1.024 Friends of friends manipulation 5.224 1.1025. Results
All of the data was collected using the online survey tool Qualtrics. After finishing the data collection the data collected with Qualtrics was exported to SPSS 25. The data exists of 153 fully completed surveys.
5.1 Correlations
A correlation analysis was run to quantify the relationships present between different variables. Figure 5 provides an overview of all the scale means, standard deviations and correlations between all the variables. Firstly the events manipulation shows a significant correlation with the friends of friends manipulation (r = .478, P = -.478) which indicates that the respondents perceive both manipulations as similar. The independent variable Friends of Friends does not show a significant correlation with mediators trust users (r = .015, P = .849) and trust platform (r = -.013, p = .870), just as the events manipulation does not show any significant correlation with trust users (r = -.058, p = .478) and trust platform (r = .045, p = .582). This means both manipulations do not significantly correlate with trust, suggesting the manipulations failed to influence respondents their levels of trust. The Friends of Friends manipulation does not significantly correlate with social distance (r = .095, p = .244). The events manipulation also does not have a significant correlation with social distance (r = -.154, p = .057). This suggests the Friends of Friends and events conditions did not influence social distance at all. As expected the mediators trust users (r = .338, p < 0.01) and trust platform (r = .288, p < 0.01) are positively correlated with sharing intentions. This indicates these variables are antecedents of sharing intention, as stated in the hypotheses. Trust usersand trust platform are also strongly correlated (r = .704, p < 0.01), which is to be expected as both measure different aspects of the same variable trust. Social distance has a significant negative correlation with sharing intention (r = -.254, p = .002). This means when social distance increases, sharing intention decreases. Figure 7: Overview correlations mediators with the dependent variable Some control variables were also added to the correlation matrix. Age is positively correlated with sharing intention (r = .238, p < 0.01). Also some control variables are inter-correlated. Age positively correlates with disposition to value privacy (r = .210, p < 0.01) and negatively with gender (r = -.186, p < 0.05). Also income correlates with disposition to value privacy (r = .207, p < 0.05) and Age (r = .376, P < 0.01). An overview of all correlations can be found in table 6. However no significant correlation was measured between disposition to value privacy and sharing intention. .338**
Trust users Sharing intention
Trust platform Social distance Sharing intention .288** Sharing intention -.254**
Table 6: Correlation matrix Variables M SD 1 2 3 4 5 6 7 8 9 10 1 D_FOF .32 .468 2 D_EVENTS .33 .471 -.478** 3 SI_TOT 4.29 1.319 -.154 -.002 4 TG_TOT 4.247 .812 .015 -.058 .338** 5 TP_TOT 4.614 .711 -.013 .045 .288** .704** 6 SD_TOT 5.216 1.251 .095 -.154 -.254** -.321** .-.314* 7 DTVP_TOT 4.575 1.306 .070 .017 -.034 .125 .093 .015 8 D_Female .601 .491 .158 -.002 .017 .024 .084 -.144 -.013 9 Age 28.99 12.797 -.004 .025 .238** .055 -.021 .077 .210** -.186* 10 Ecuation_LowHigh .660 .475 .019 .117 -.144 .062 .016 -.099 .105 .092 -.012 11 Income_LowHigh .222 .417 -.030 .063 .062 -.012 -.086 .019 .207* -.014 .376** .052 **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
5.2 Hypothesis testing
In the following paragraph all hypotheses will be tested using the PROCESS by Hayes et al. (2012). Mediation model 1: IV Friends of friends with mediator trust The friends of friends condition is expected to have a positive influence on sharing intention (H3a). This positive effect is mediated by a) trust in peers and b) trust in the platform.. The friends of friends manipulation is expected to have a positive effect on a) trust in the platform and b) trust in other users (H3b). In turn higher levels of both types of trust will result in a higher sharing intention (H2). This model can be found in figure 8, and controls for age, gender, education, income and disposition to value privacy. The R2 of the model is .236 indicating this model explains 23.6% of the variance in sharingintention. The direct effect of friends of friends positioning on sharing intention is -.448 with p < .05 , t = -2.145 and a CI between -.861 and -.035, meaning the effect negative effect of friends of friends framing on sharing intention is significant. The friends of friends manipulation has no significant effect on trust platform (r = -.060, p = .632 and trust in other users (r = .001, p = .999). Trust platform also does not have a significant influence on sharing intention (r= .192 p = .320). However trust in other users does have a positive effect on sharing intention (r = .445 and p = .009, t = 2.641 and a CI between .112 and .777). Last the mediating effect of trust platform (r = -.012 and CI between -.131 and .051) and trust users (r = -.001, CI between -.137 and .157) between platform manipulation and sharing intention is not significant. With regards to the control variables age seems to have a positive effect on sharing intention r = .027, p < 0.05, t = 3.204 and CI between .010 and .043 as well as education r = -.426, P < 0.05, t = .024 and CI between -.831 and -.021. Figure 8: Mediation model 1 .192 .445* -.448* Trust users Trust platform Sharing Intention FOF manipulation -.060 *p < .05 -.001 .001 -.012
