Why sharing should be about caring again : a vignette study measuring the impact of platform positioning, trust, perceived closeness and familiarity on car owners’ car sharing intentions

(1)

Why sharing should be about caring again

A vignette study measuring the impact of platform positioning, trust, perceived closeness and

familiarity on car owners’ car sharing intentions

Master thesis MSc Business Administration – track Marketing

Amsterdam Business School – University of Amsterdam

J.M.S. Straathof

Student number: 10535861

26-01-2018

Final Version

Word Count: 18.025

Supervisor: N.O. Stofberg

(2)

Statement of Originality

This document is written by Jolien Straathof 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.

(3)

List of Tables

Table 1 Applied Vignettes ... 28

Table 2 Distribution of participants across vignettes ... 29

Table 3 Pearson Correlations between variables of Pre-test 1 ... 35

Table 4 Pearson Correlations between variables of Pre-test 2 ... 36

Table 5 Pearson Correlations for final manipulation check ... 45

Table 6 Pearson Correlations, Means and Standard Deviations ... Error! Bookmark not defined. Table 7 Coefficients and significance levels of indirect, direct and total effect ... 51

Table 8 Coefficients and significance levels of effects on Trust and Sharing Intentions ... 53

Table 9 Overview of hypotheses: rejected or supported ... 55

List of Figures

Figure 1: The four Relational Models and commitment and interdependence ... 18

Figure 2: Conceptual Model ... 26

Figure 3: Vignette 1 = high Platform Positioning / high Perceived Closeness ... 37

Figure 4: Vignette 2 = high Platform Positioning / low Perceived Closeness ... 38

Figure 5: Vignette 3 = low Platform Positioning / high Perceived Closeness ... 39

Figure 6: Vignette 4 = low Platform Positioning / low Perceived Closeness ... 40

Figure 7: PROCESS Model 21 ... 50

(5)

Acknowledgements

By handing in this Master Thesis I finish my Master in Business Administration at the University of Amsterdam. Whilst this official submission feels as a great relief, I really enjoyed working on this thesis during the last couple of months and can say that I am very proud of what I have accomplished.

I would like to thank my thesis supervisor Nicole Stofberg for her personal supervision. With her unconditional enthusiasm, critical view and deep knowledge of the sharing economy she inspired me to reach my true potential.

Furthermore, I would like to thank my friends, for being my friends. Without their unconditional support, pleasant companion during study breaks and great ideas on how to spend the weekend I would not have been able to focus on studying whenever I needed to. They have made this final year as a student a pleasure.

Lastly, I would like to thank the University of Amsterdam for giving me the opportunity to develop myself academically and preparing me for the professional work field.

(6)

Abstract

This paper addresses motivations that drive participation in the car sharing economy. Where previous research focused solely on the impact of potential participants’ own motivations on their intentions to lend out their car, this research examines the influence of the other party’s perceived motivations on one’s own intentions to participate. The assumptions that people act always out of self-interest and that financial arguments are therefore the main drivers of participation, are challenged this way. Building on Relational Model Theory, this thesis reveals the importance of relational interaction within a sharing context. It asserts that car owners are more easily persuaded in sharing their car when communal gains are emphasized rather than self-gains. This because emphasis on communal gains forms a relationship based upon Communal Sharing, which instills expectations of mutual respect and mindfulness of one another’s property, by which peer to peer trust increases. In contrast, when managers position their platform as being centered around financial gains only, this reinforces the belief that both parties act out of self-interest, which in turn limits the development of trust.

A vignette experiment (N=351) was conducted in collaboration with the Dutch leading car sharing platform SnappCar. By showing respondents different versions of the SnappCar landing page this research examined the influence of a platform positioning that either triggers Communal Sharing relationships or Market Pricing relationships on car owners’ car sharing intentions. Trust was hypothesized as a mediator, and familiarity and perceived closeness were expected to moderate the mediation effect. The results showed platform positioning to affect sharing intentions indirectly through trust, whilst a direct effect remained absent. Familiarity did not moderate the mediating relationship, but did have a direct effect on sharing intentions. Perceived closeness neither moderated the mediating relationship, but did affect trust in a positive way.

Keywords:

* Sharing economy * Car sharing * Sharing motivations * Relational Model Theory * Communal Sharing * Market Pricing * Platform Positioning * Trust * Perceived Closeness * Familiarity

(7)

Chapter 1: Introduction

‘’ Sharing is to ownership, what the iPod is to the eight-track, what solar power is to the coal mine’’ (Levine, 2009, p.1)

Fast technological developments (Hamari, Sjöklint & Ukkonen, 2016), together with a decline in the value consumers attach to the possession of goods (Chen, 2009; Hwang & Griffiths, 2017; Botsman, 2010) and an increase in consumer awareness regarding sustainability (Verbeke, Vanhonacker, Sioen, Van Camp & De Henauw, 2007) have led to the development of a new way of consumption: the act of sharing goods between consumers through online sharing platforms. Whereas sharing on itself has according to Belk (2010) been ‘’the most basic form of economic distribution in hominid societies for several hundred years’’(Belk, 2010, p.715), traditional sharing is now being reinvented with the use of digital developments. With the potential to reduce climate change significantly, the rise of the

‘sharing economy’ is more welcome than ever. Characterized by the temporal access of goods instead of true ownership, the sharing economy is a socio-economic system, built around the sharing of human and physical resources (Matofska, 2017). The sharing economy, or by other scholars called ‘collaborative consumption’, ‘anti-consumption’ or ‘access-based consumption’ (Belk, 2010; Habibi, Kim & Laroche, 2016; Gorenflo, 2013; Hartl, Hofmann & Kirchler, 2016) can be defined as the ‘‘the peer-to-peer-based activity of obtaining, giving, or sharing the access to goods and services,

coordinated through community-based online services’’ (Hamari, Sjöklint & Ukkonen, 2016, p. 1). It is a powerful economic and cultural force that reinvents not what we consume but how we consume (Botsman, 2010). Lamberton (2016) states that consumers benefit from this collaborative

consumption not only in the way that costs of ownership are avoided, but social relationships are strengthened as well (Lamberton, 2016).

Five years ago, Forbes claimed the growth of this new way of consumption to be unstoppable, mentioning growth rates of 25% (Geron, 2013). However, its impact is currently still far below its potential due to participation rates that in some industries stay below expectations (European Commission, 2016). The specific industry of car sharing for example, experiences low car owner registrations (Boztas, 2017), whilst cars are very shareable goods (Demailly & Novel, 2014).

(8)

Attitudes towards sharing are generally positive but participation rates are staying behind, which indicates an attitude behavior gap (Hamari et al., 2015). As founder of the Dutch leading car sharing platform Victor van Tol admits: ‘’I expected way faster growth […] it is hard to bring supply and demand together, since car owners need to be persuaded’’ (Daalder, 2018). Since the success of the sharing economy depends on obtaining a critical mass (Stofberg, in press), this is unfortunate. The main reason why car owners refrain from lending out their car seems to be a lack of trust (Boztas, 2017; Schor & Fitzmaurice, 2015).

Since car sharing is a relatively new phenomenon, little research has been conducted regarding drivers of participation, and the existing scientific articles on car sharing motivations present contradicting findings (e.g. Bucher et al., 2016; Bardhi & Eckhard, 2012). Both economic and social motivations, just as other traits as familiarity, enjoyment, community belonging, seem to play a role (Möhlmann, 2015). More interesting, since sharing is obviously an act between two people, it is remarkable is that the potential impact of car user’s participation motivations on car owners’

intentions to lend out their car has not received any attention in research yet. Especially with trust being the most important barrier for participation in mind. Therefore, this thesis aims to expand the knowledge about car sharing by examining the impact of the other party’s motivation on consumer intentions to share out their car. Since owner and user usually have never met before the act of sharing takes place, the way a sharing platform positions itself online is key in this peer’s mental

representation of his sharing partner(s motivations) and the relationship they establish together. The relationship between Platform Positioning and a car owners’ intentions to share their car is therefore studied, taking the role of trust into consideration as well.

With trust playing a central role in the sharing economy (Schor & Fitzmaurice, 2015) other factors linked to trust as the level to which a car owner feels him- or herself to be connected to a potential car user and people’s level of familiarity with the sharing economy are taken into account as well. This perceived closeness is examined since it is expected to reinforce the effect of a social platform positioning on trust, whilst familiarity is take into account because it is expected to lower other participation barriers (e.g. effort). The research question is therefore as follows:

(9)

‘‘How are a car sharing platform’s positioning (social versus financial) and car owners’ intentions to share their car related, and what role do trust, familiarity and perceived closeness play?’’

Through a vignette study in collaboration with the leading Dutch sharing platform SnappCar, this research is the first to address this question in a realistic setting.

This thesis is divided in six chapters. The following chapter (Chapter 2) elaborates on the existing theory regarding (car) sharing and formulates hypotheses with corresponding theoretical foundation. Chapter 3 describes the methodology of the experimental research. Chapter 4 presents the results, after which Chapter 5 discusses these results, gives limitations of the research and gives suggestions for future research. The final chapter (Chapter 6) provides a general conclusion based upon the research outcomes.

(10)

Chapter 2: Literature review

This chapter discusses relevant theory of the (car) sharing economy and provides a theoretical framework for the proposed hypotheses.

2.1 The sharing economy

2.1.1 The rise of the sharing economy

In the past years, a significant amount of sharing platforms have popped up in society. This upcoming way of consumption which enables people to consume peer-to-peer (P2P) through the use of online platforms, challenges the business-to-consumer (B2C) way of consumption and is called the ‘sharing economy’. Examples of these peer-to-peer platforms are Airbnb, Thuisafgehaald, BlablaCar, Peerby and SnappCar, through which houses, meals, rides, everyday utensils and cars are shared. True ownership is avoided, and characterizes this innovative way of consumption. The sharing

phenomenon is enabled by the rise of technology that caused and still causes easier contact between consumers (Hamari, Sjöklint & Ukkonen, 2016; Cornet, Mohr, Weig, Zerlin & Hein, 2012). With those technological developments we’ve ‘‘wired our world to share’’, and are ‘‘moving from a culture of me, to a culture of we’’ (Botsman, 2010).

Since there is no clear consensus about what the boundaries of sharing are and the sharing economy has quite some positive connotations, many parties try to benefit from its popularity, resulting in companies claiming to be part of the sharing economy whilst they are not (Belk, 2014a). A clear definition of what ‘true sharing’ entails is necessary since important misconceptions regarding what motivates participation stem from definitional issues (Frenken & Schor, 2017). As Habibi, Kim and Laroche (2016) explain, many new sharing practices have arisen and because of the sharing economy’s beneficial image they all present themselves as ‘true sharing’. What for example distinguishes SnappCar (a peer to peer car sharing platform) from Greenwheels (a B2C platform) is that for Greenwheels the cars are owned by the platform and a private experience in which members never meet, nor interact with one other is offered. SnappCar in contrast offers a very personal experience. Users interact every time a transaction takes place by handing over the car keys of privately owned cars. To avoid confusion and set borders for ‘true sharing’, in this research,

(11)

commodity exchanges are only to be considered part of the sharing economy if four conditions are fulfilled. Corresponding to Frenken and Schor (2017), the shared assets must be physical, under-utilized (‘idle capacity’), the access must be temporary, exchange should be between consumers and may be for money. This means that for example the second hand economy, on-demand economy and product-service economy are not considered sharing, but rather ‘pseudo-sharing’. Greenwheels is not considered ‘true sharing’ but ‘smart rental’ (Frenken & Schor 2017) because the exchange does not occur between consumers, but B2C, and does not come from idle capacity. Interestlingly, Habibi et al. (2016) show that consumers do not perceive such a B2C sharing experience as true sharing either. SnappCar does in contrast awake a sense of true sharing amongst consumers through the very personal experience it offers. The interaction between peers results in a much higher degree of actual sharing going on. The misuse of the term ‘sharing’ by many firms (amongst which Greenwheels), has according to Habibi et al. (2017) ‘‘resulted in confusion over semantics that result in detrimental outcomes for managers and practitioners as they might misallocate firm resources’’ (Habibi et al., 2017, p.115). Specifically, this research believes that many true sharing platforms allocate a lot of their marketing budget in emphasizing the ‘financial gains’ they mistakenly draw from the ‘best practices’ of ‘pseudo-sharing’ practices, which are very different from their own. They that way fail to build a strong community and fail to emphasize the social aspects which are of crucial importance in a ‘true sharing’, or ‘peer to peer’ context.

2.1.2 The relevance of the sharing economy

The sharing economy is not only beneficial for consumers in the way that it is financially and socially attractive, but it could have a significant positive influence on the environment as well. In 1987, the World Commission on Environment and Development called ‘sustainable development’ ‘’the development that meets the needs of the present without compromising the ability of future

generations to meet their own needs’’ (World Commission on Environment and Development, 1987), which is something to strive for. It is no secret that the way people consume nowadays can in no way guarantee the ability of future generations to meet their own needs in the future. In today’s globalized world where consumers have access to products from all over the world, the environment suffers from

(12)

non-sustainable excess production and -purchase. Consumers live in a state of ‘hyper consumption’ that will on the long term destroy the environment, the economy and in the end themselves as well (Albinsson, Wolf & Kopf, 2010). Ninety seven percent of climate scientists state that global warming in the past and upcoming centuries is caused by human activity (NASA, 2013). The American Physical Society states: ‘’If no mitigating actions are taken, significant disruptions in the Earth’s physical and ecological systems, social systems, security and human health are likely to occur. Emissions of greenhouse gases must be reduced, beginning now’’ (APS, 2007, p1). Unfortunately political leaders fail to significantly change the situation by the means of for example regulations (van de Glind, 2013). Therefore, the consumption pattern of consumers has to drastically change in order to fight climate change. The act of sharing goods between consumers has the potential to significantly reduce climate change, since sharing is simply less resource intensive than purchasing outright (Schor, 2016; Rifkin, 2014). Consumption and thus production and waste can be reduced. For example, over 33% of just household waste could be prevented if shareable household goods would be shared (Demailly & Novel, 2014) and one shared car could replace four (Demailly & Novel, 2014) up to fifteen (Bonderová & Archer, 2017) ‘regular’ cars. For cars in specific this is relevant for two reasons. Firstly, eighty percent of the people who participate in car sharing drive fifteen to twenty percent less kilometres than before they participated (Nijland, Meerkerk & Hoen, 2015; de Jong & Jakobs, 2009; Chatterjee, 2013). Secondly because the production of a new car creates just as much carbon pollution as driving it (Beners-Lee & Clark, 2010). Chatterjee (2013) also finds the often heard

counterargument that car sharing would mainly trigger a shift from biking and the use of public transport to shared cars (Efthymiou, Antoniou & Waddell, 2013) is untrue in reality. Users of shared cars generally increase their use of public transport and walk more often (Chatterjee, 2013).

The sharing economy has a clear potential in fighting climate change and many scholars believe in the sustainable ‘‘movement from individual getting and spending towards a rediscovery of collective good’’ (Botsman, 2010). According to Levene (2009) sharing relates to ownership just as solar panels relate to coalmines (Levene, 2009). However, many sharing platforms are still in their infancy (Belk, 2014b) and participation rates in many industries are staying behind (European

(13)

in at least the Netherlands – 51 percent of adult inhabitants knows Airbnb - and attitudes towards this platform are mostly positive, only 4,9 percent of the adult habitants of the Netherlands has ever used the platform until May 2017 (Newcom, 2017). There thus is a potentially high impact possible of the sharing economy on the environment, but consumers participate in a smaller extent than necessary to make the environment benefit. Researching these consumer considerations on whether or not to ‘join’ the sharing economy is highly relevant for at least two reasons. Firstly because it enables us to understand motivations behind the upcoming sharing economy better. Secondly because we have the moral responsibility to the environment and future generations to exploit this opportunity to increase sustainable consumption.

2.1.3 Motivations behind sharing participation

Even though the sharing economy could ‘‘provide a new pathway to sustainability’’ (Heinrichs, 2013), environmental motivations in reality do not directly lead to actual behavioural sharing intentions (Hamari et al., 2016), nor to an increase of likelihood to using a sharing economy option again after use (Möhlmann, 2015). This whilst general attitudes towards consumption have radically changed due to broader knowledge of ecological, societal and developmental impact: people have positive attitudes towards sharing (Hamari et al., 2016). Other motivational factors that seem to do affect consumers’ sharing intentions to a bigger extent are economic and social factors (Schor, 2016). Moreover, according to other research, trust, familiarity, community belonging and service quality play a role as well (Möhlmann, 2015), just as enjoyment of the activity and economic gains (Hamari et al., 2016). The motivations that do seem to affect sharing intentions can broadly be placed under the umbrella terms ‘social’ and ‘economic’.

2.2 Car sharing in specific

According to Demailly and Novel (2014) the car-sharing economy has a high potential in making the world more sustainable, and cars are considered ‘highly sharable goods’. Car sharing between peers meets all conditions to be considered true sharing, and cars stand 92% of their time idle, on a parking space (Demailly & Novel, 2014). Moreover, attitudes towards car sharing are generally positive. Thus far, cars seem the ideal good to share. However, existing car sharing platforms have problems

(14)

acquiring new ‘cars to share’: the amount of car owners that are persuaded to truly lend out their car is staying behind (European Commission, 2016). The Dutch leading car sharing platform SnappCar recently even took measures by starting a, hopefully temporary, collaboration with a leasing company through which customers can lease a car for a lower price if they lend out their car through SnappCar twice a month (Daalder, 2018). This indicates the ‘problem’ is real: even though cars are highly sharable goods and consumers have positive attitudes towards car sharing, something clearly discourages car owners to participate. This research tries to find an explanation for this lacking growth, and aims to expand the existing knowledge on factors that do or do not incentivize car owners to participate in car sharing. These participation intentions will in this research be called ‘Car owners’ Car Sharing Intentions’, or in short ‘Sharing Intentions’ or ‘Intentions to Share’. They can be

understood as the extent to which car owners have the intentions to share their own car with others through a car sharing platform.

2.2.1 Car sharing motivations

As mentioned earlier, people have generally positive attitudes towards (car-)sharing (Hamari et al., 2015), regardless their motivations (economic or social). In reality however, car sharing platforms are way behind house-sharing platforms as Airbnb, participation rates are rather low (European

Commission, 2016). This is in line with the attitude-behaviour gap that Hamari, Sjöklint and Ukkonen (2015) found: people perceive sharing as positive, but this does not always lead to participation. Data shows that even though 75% of respondents of a survey of the European Commission mentioned that they would like to try sharing, only 17% of European inhabitants actually participated in some sort of sharing (European Commission, 2016; Owyang, Samuel & Grenville, 2014). For car sharing in the Netherlands participation is only 1,7 percent of adult inhabitants (Newcom, 2017). As mentioned before, the main reason why car sharing platforms are staying behind is because of the low numbers of car owners that join such a platform (Boztas, 2017).

In an attempt to make car owners join the sharing economy, platforms mainly play into the economic benefits that come from sharing your own car with others. They try to persuade car owners by emphasizing the money that can be earned through sharing: £3000 a year (easyCar Club, 2018),

(15)

$10.000/year (Getaround, 2018) or £650/week (HiyaCar, 2018). In a similar trend the Dutch car sharing platform SnappCar tries to convince car owners: ‘‘Rent out your car twice a month and earn rapidly €1000 a year. What will you earn with your car? Enter your license plate number and find it out’’ (SnappCar, 2018). This is in line with claims of for example Bardhi & Eckhardt (2012) that participation in the sharing economy is almost solely driven by personal, self-centred motivations (Bardhi & Eckhardt, 2012). However, the sharing their research examines is considered pseudo-sharing since Zipcar is a B2C company. This indicates that ‘true’ car pseudo-sharing platforms currently try to persuade consumers with arguments that work very well in a different field of practices than their own. In addition, in their own industry of ‘true sharing’, social motivations have even been found to be a necessary requirement for participation (Bucher, Fieseler & Lutz, 2016). Also from actual participation rates we can assume that the current strategies of car sharing platforms are not

resonating with car owners. There is a clear gap between theory and practice: some researchers claim financial and other self-centered motivations to be key drivers for sharing participation, whilst that type of arguments in practice do not seem to truly ‘do the job’: participation is limited. The confusion about and misuse of the term ‘sharing economy’ plays a role here. True sharing and practices of pseudo-sharing require very different levels of trust. The shared cars in ‘true sharing’ initiatives are privately owned, which indicates that lending them out comes with a risk for car owners and therefore a high level of peer to peer trust is required for them to actually lend out their car. In many pseudo-sharing contexts, the shared cars are owned by the company itself: consumers do not experience trust barriers since their risk is low.

2.2.2 Trust as a participation barrier

As SnappCar founder Victor van Tol says in an interview with The Guardian himself: ‘‘The most challenging thing for a business like ours is to get car owners share their car’’(Boztas, 2017). This because car owners do not believe the potential monetary benefits outweigh the potential costs of participation. They feel like renting out their car to strangers comes with a risk. Sheppard & Sherman (1998) call this the ‘risk of unreliability’: ‘‘a concern that another will not behave as expected’’ (Sheppard & Sherman, 1998, p.424). Whilst this lack in trust might concern small risks as receiving

(16)

your own car back covered in dog hair or with a smell of cigarettes, Van Tol takes this risk even a little further: ‘‘You can do nasty things with a car. You can take a car and rob a bank, or go abroad and never come back’’ (Boztas, 2017). People perceive damage to one’s own personal goods not only as material loss but might even experience it as a ‘lessening of the self’(Belk, 1988), because

‘‘knowingly or unknowingly, intentionally or unintentionally, we regard our possessions as parts of ourselves’’ (Belk, 1988, p. 139). Moreover, a report on trust of BlablaCar (2012) that gave insights into the extent to which people trust one another showed that whereas family and friends score 4.7 (on a 0 to 5 scale) on the degree of trust given, a stranger online only scores 1.9 out of 5 (Mazzella & Chronos, 2012). In other words, car owners generally do not trust strangers that are willing to rent their car, and the lacking growth of participating car owners shows that a financial benefit of roughly €40,- per day (SnappCar, 2018) does not outweigh this perceived risk for them. By a report on sharing behavior by Bardhi and Eckhart (2012) it firstly seems that these trust issues might be justified .Users admit they would ‘‘double park the car real quick, or park it in a tighter spot than I would do with my own car’’ (Eckhard & Bhardi, 2012, p.889) and state: ‘‘If I destroy the suspension, so be it!

Somebody will fix it, not me!’’ (Eckhard & Bhardi, 2012, p.891). However, this research interviewed users of ZipCar, a B2C pseudo-sharing initiative.

2.2.3 The influence of the other party’s motivations

Since trust as ‘‘the belief that someone is good and honest and will not harm you, or that something is safe and reliable’’(Trust, n.d.) is based upon the form and depth of relationships between people (Sheppard & Sherman, 1998), it is remarkable that only scant attention has yet been paid to the influence of the other party’s motivations. Since sharing involves two parties, the motivations of the second party may influence the first party’s participation intentions to a bigger extent than previously thought. The design of current websites of most car sharing platforms (focusing on financial benefits) triggers the car owners’ extrinsic motivations, based on money. Potential renters notice this as well, which makes them perceive the sharing activity as purely a market transaction for the car owner – in which people mainly act out of self-interest (Fiske, 1991, 1992). As a consequence, renters may feel like acting out of self-interest as well, resulting in a justified low trust relationship from both sides.

(17)

Even though many platform founders’ argument that paid services are necessary in order to scale sharing practices might be true (Bouma, 2015), it is highly questionable whether their marketing should be focused on this financial aspect.

Based upon the same principle, it might be the case that whenever potential car users believe the car owners’ motivations are mainly intrinsic – in other words: the owners acts out of other-centred motivations – these potential renters will do so as well. In this case there arises a very different relationship between the two involved parties, with different behaviour as a consequence. Both parties will feel more socially connected to one another and users will treat the shared good as if it is their own. From this view, ‘’trust is not an irrational, but a manageable act of faith in each other’’ (Sheppard & Sherman, 1998, p. 422) and can in the case of car sharing be manipulated through a platform’s communicated arguments. According to Sheppard and Sherman (1998), risks associated with interdependence can be mitigated when properly managed (Sheppard & Sherman, 1998). This idea of ‘managing trust’ is built upon Fiske’s Relational Model Theory (1991, 1992).

2.3 Fiske’s Relational Model Theory

According to Fiske’s Relational Models Theory (RMT) (1991, 1992), people act differently in different social relationships due to different norms and values per type of relationship. The existing empirical support for his theory is extensive (Haslam & Fiske, 1999).

2.3.1 Relational Model Theory, commitment and interdependence

In the RMT, Fiske names Market Pricing (MP), Equality Matching (EM), Authority Ranking (AR) and Communal Sharing (CS) as the four elementary cognitive models in terms of which social relationships are constructed, represented, evaluated and comprehended (Haslam & Fiske, 1999). People use these relational models unconsciously and act in a way that corresponds with the type of model used. The four relational models have distinct features that define social interaction:

‘’Market Pricing (MP) organizes relationships with reference to a common scale of ratio values such as money; social transactions are reckoned as rational calculations of cost and benefit’’ (Fiske & Haslam, 1999, p.242). MP relationships exist for example between consumers and sellers at a market, and are characterized by calculations of costs/benefits in terms of mostly money (Sheppard

(18)

& Sherman, 1998). Equality Matching (EM) organizes relationships with reference to their degree of balance or imbalance; it is manifested in distributions of equal shares and tit-for-tat retaliation. People are distinct but equal (Sheppard & Sherman, 1998). Communal Sharing (CS) organizes relationships in terms of collective belonging or solidarity. Members of an in-group are treated as equivalent elements of a bounded set, and unity, community, collective identity and kindness are central elements of this type of relationship (Sheppard & Sherman, 1998). Such relationships occur mainly among close kin. Authority Ranking (AR) organizes asymmetrical relationships, in which parties are hierarchically ordered and decision making is based upon relative status (Haslam & Fiske, 1999). An overview of key features of the relational models by Bridoux and Stoelhorst (2016) is presented in Appendix A.

According to Heyman and Ariely (2004), Fiske’s four relational models can be divided into two categories: economic exchanges (MP) and social exchanges (CS, EM and AR), in which CS is considered most social of the three (Heyman & Ariely, 2004). Depending on the type of relational model applied, people are confronted with different levels of ‘commitment’ and

‘interdependence’(Sheppard & Sherman, 1998). Commitment implies the level to which a person puts effort into maintaining the relationship (Blois & Ryan, 2012), whilst interdependence shows to what extent ‘’the completion of one’s own consequential activities depend upon the prior actions or ongoing cooperation of another’’ (Sitkin & Roth, 2006, p.298). As Figure 1 shows, AR and MP involve low commitment, whereas EM and CS involve high commitment. In other words: in EM and CS relationships, people tend to put a lot of effort in maintaining the relationship, whereas people do not do this in MP and AR relationships. Regarding interdependence, people in AR and CS

(19)

way. People in MP or EM relationships can vary in their level of interdependency.

From this model we can derive that people who unconsciously apply the Communal Sharing model are most likely to have other-centered motivations for their behavior. This because they put most effort in maintaining the relationship and can only reach their own goals if there is a good

collaboration between the two parties. This interdependency and mutual commitment increase mutual trust (Sheppard & Sherman, 1998). People who are in a shallow-interdependence Market Pricing relationship are most likely to act out of self-interest. They do not feel like putting effort in the existing relationship and are neither dependent on the other party to obtain their own goals. As Preece (2000) points out: “When there is trust among people, relationships flourish; without it, they wither” (Preece, 2010, p. 191).

2.3.2 Platform Positioning as a relational model trigger

As mentioned before, current car-sharing platforms focus mainly on financial motivations for potential participants that own a car. These financial motivations trigger a relationship based on Market Pricing (Sheppard & Sherman, 1998; Heyman & Ariely, 2004). This means, according to Blois and Ryan (2012), there is low commitment and shallow interdependence, mostly resulting in the application of self-interest norms and values - which in its turn causes self-interest behaviour

(Bridoux & Stoelhorst, 2016). A consequence is that participants do not care about the social relationship, resulting in a low trust for both parties. This is confirmed by Möhlmann (2016), who argues that many consumers act out of self-interest in the sharing economy by stealing and damaging wilfully, resulting in decreased trust (Möhlmann, 2016). However, where some scholars conclude that sharing is thus not social at all – just similar to economic market exchanges –they did not take

relational models into consideration. In the light of the RMT, the misbehaviour might be caused by the triggered Market Pricing relationship since most car sharing platforms as SnappCar, HiyaCar and easyCar Club currently communicate purely economic motivations. If a different type of relationship would be triggered, in particular the Communal Sharing relationship since it is the most social one, the behaviour is expected be very different. This is in line with Schaefers, Wittkowski, Benoit and Ferraro (2016) who conclude that (mis)behaviour in the sharing economy is dependent on specific

(20)

social norms and values amongst the customer group (Schaefers et al., 2016). Since kindness,

collective belonging and solidarity are key in CS relationships, a website that shows these elements as key motivations will trigger those norms and values through the activation of a Communal Sharing relationship. This way, misbehaviour would decrease whilst trust would increase. Moreover,

Schaefers et al. (2016) found that another reason for the ‘better behaviour’ in environments that show social motivations lies in the lower level of anonymity and establishment of more personal

relationships between peers (Schaefers et al, 2016). In a similar way, Acquier, Daudigeos and Pinkse (2017) conclude that misbehavior in the sharing economy occurs mainly, if not solely, in weak community-based platforms (Acquier et al., 2017). In conclusion, built upon the RMT, the level of trust and behavior in the sharing economy are expected to be dependent on the context: the

relationship between peers. Sharing platforms would this way be able to ‘manage’ the level of trust and behavior by triggering specific types of relationships amongst peers.

Since a monetary transaction will still be present in a socially presented car sharing process, the desire of car owners to financially benefit from sharing is still fulfilled but just not emphasized. Moreover, due to the presence of a financial transaction, a webpage built to trigger a CS relationship might potentially trigger an EM relationship (tit for tat) instead of solely a CS relationship as well. Since both relationships, CS and EM are the most social of the four, this is not necessarily a problem. Relationships based on Authority Ranking are characterized by ‘’asymmetry among people who are linearly ordered along some hierarchical social dimension’’ (Boer, Berends & Baalen, 2011, p.87) and this is clearly not the case between members of a car sharing platform, this relational model is left out of further account.

CS and MP are both extremes regarding the socialness of the relationship based upon

commitment and interdependence: ‘‘CS and MP are poles on a bipolar dimension of warmth (and EM and AR are poles on a bipolar dimension of inequality)’’ (Haslam, 2004, p.30). This paper will call the way in which a sharing platform’s website triggers either MP or CS ‘Platform Positioning’. The variable can be understood as high Platform Positioning triggering a Communal Sharing relationship, and low Platform Positioning triggering a relationship based upon Market Pricing. To be clear, the monetary transaction is in both cases the same, the difference lies in the way in which the landing

(21)

page emphasizes this monetary transaction or focuses on social interaction.

In sum, more personal relationships and a higher communal identification amongst

participants are expected to increase trust between car owners and users. Since a lack of trust is the main participation barrier for car owners and social Platform positioning will probably reduce this barrier to a minimum, Platform Positioning is expected to affect participation intentions directly. Belk (2010) was probably very right, sharing can be a powerful communal act that creates feelings of solidarity and belonging. But this depends on the context: only when the other party’s motivations are perceived as mainly other-centred. Therefore, the following hypothesis is formulated:

Hypothesis 1: A high level of Platform Positioning (Communal Sharing) leads directly to higher Sharing Intentions than a low level of Platform Positioning (Market Pricing).

As described in the precedent theory, a high level of Platform Positioning is expected to increase Sharing Intentions because it increases interpersonal trust. This happens because relationships based upon Communal Sharing models go hand in hand with trust (Belk, 2010). Such relationships are characterized by deep interdependence and high commitment (Sheppard & Sherman, 1998), which causes members of a group to place the interest of the group above their own and therefore trust one another (Stofberg, 2016). In MP relationships on the other hand, self-centered behavior is the norm, which creates interpersonal reserve amongst participants and limits relational depth and trust (Bridoux and Stoelhorst, 2016: 237). This mediating role of Trust leads to the following hypotheses:

Hypothesis 2: The positive effect of Platform Positioning on Sharing Intentions is mediated by Trust, such that high Platform Positioning (Communal Sharing) leads to higher Trust than low Platform Positioning (Market Pricing), and Trust positively affects Sharing Intentions.

Hypothesis 2a: There is a positive relationship between Platform Positioning and Trust.

(22)

2.4 Other factors that play a role

Other factors might play a role in this relationship of Platform Positioning and Trust on Car owners’ Sharing Intentions. In this paragraph the role of Perceived Closeness and Familiarity are elaborated on.

2.4.1 The influence of Perceived Closeness

Since potential car owner and user have not met in real life before the actual sharing activity occurs, they generally build a first impression of each other purely based upon the other’s online presentation. Besides the communicated motivations which are discussed in the previous paragraphs, all other present characteristics may play a role in this first impression as well. It is for example already widely known that members with a complete online profile are trusted more than ‘anonymous’ complete strangers (Mazzella & Chronos, 2012). Simply ‘having’ a complete online profile in the first place thus affects Trust in a positive way, but the content this online profile displays is just as important (Zhao, Lu, Wanh, Chau & Zhang, 2012). To elaborate on this, if such a profile enhances the

‘perceived closeness’ between two people, this further increases trust (Moran, 2005). This closeness can be defined as ‘‘The extent to which personal familiarity exists in a relationship’’(Moran, 2005, p.1135). The author states that people generally have the feeling that others who are close to themselves are less likely to betray them (Moran, 2005). In addition Moran states that ‘‘people are more likely to offer information, know-how or aid [as lending out their car] to others who are close, than those more distant’’(Moran, 2005, p.1135). Specified to car sharing this would mean that when a car owner perceives the potential car user as close to him- or herself, this further enhances trust. Similarity is a form of perceived closeness, similar others are being perceived as closer to oneself than dissimilar ones (Liviatan, Trope & Liberman, 2008). Similarity is ‘‘the extent to which members perceive to be sharing common characteristics such as shared goal and vision with other members’’ (Zhao et al., 2002, p.576). This similarity can regard a wide variety of characteristics, ranging from having the same sexuality to sharing hobbies, interests, friends, religion, education or even similar physical appearance (Zhao et al, 2002). Just as Moran describes a positive relationship between closeness and trust, research by Zhao et al. (2002) and by Luo (2002) find a clear positive

(23)

relationship between similarity amongst members of virtual communities and P2P trust. This is logical when keeping in mind that Similarity is a form of Perceived Closeness (Liviatan, Trope & Liberman, 2008). This research will further focus on the term Perceived Closeness in order to mitigate confusion.

In the light of this research, the Perceived Closeness of a potential car user could function as a ‘lever’ for the effect of Platform Positioning on Trust. This would work in such a way that if a car owner perceives him- or herself to be very similar/close to a potential user, the effect of Platform Positioning on Trust would be larger than when a low level of Closeness is perceived. This because in a social Platform Positioning the aspect of helping others is emphasized, and people who perceive to be close to one another are more likely to help one another (Moran, 2005). A high level of Perceived Closeness and a high Platform Positioning are therefore expected to reinforce each other. This leads us to hypothesis 3.

Hypothesis 3: The positive effect of Platform Positioning on Trust is moderated by Perceived

Closeness, such that the effect is stronger for high levels of Perceived Closeness than for low levels of Perceived Closeness.

In order to test for the influence of Perceived Closeness, the presence and content of a car user’s online profile could be added to a sharing website. A practical way to increase Perceived Closeness would be showing the amount of mutual Facebook friends between car owner and user, since having friends in common increases similarity (Luo, 2002) and thus Closeness. This way Perceived

Closeness could be manipulated.

2.4.2 The influence of Familiarity

The sharing economy is for many consumers a relatively new and upcoming phenomenon - only 17 percent of European inhabitants participated in any type of sharing (European Commission, 2016), and only 1,7 percent of Dutch inhabitant has ever used SnappCar (Newcom, 2017). This unfamiliarity may play a role in car owners’ sharing intentions as well. Moeller and Wittkowski for instance state that some consumers might be reluctant to use a service as car sharing for the first time, just because they do not have any experience with it yet (Moeller & Wittkowski, 2010). When the sharing

(24)

economy is new to consumers, participation involves higher ‘transaction costs’ than when people are highly familiar with the way it works, with transaction costs being ‘‘the mental and physical energy required for the actual process of obtaining (and retaining) the desired product’’ (Moeller &

Wittowski, 2010, p.181). This is according to Möhlmann (2015) one of the reasons why high

familiarity with the sharing economy increases likelihood of using a sharing option again: transaction costs are already reduced to a minimum (Möhlmann, 2015). By familiarity, this research means the extent to which a person already has experience with the sharing economy by lending out goods through (other) sharing platforms (e.g. Airbnb, Peerby, Thuisafgehaald) in the past.

Moreover, Gourville (2006) states that consumers in general have an initial sceptical view on innovation for which (slight) behaviour change is required. Customers tend to overvalue their status quo, and innovative products or services therefore need to be more than ‘better’ than their status quo: they need to compensate for the behavioural change as well. This required behaviour change has a substantial impact on intentions to switch to innovative products and services (Gourville, 2006). Since the sharing economy is an innovative way of consumption, consumers need overcome a change of behaviour when they first participate (from buying B2C to lending P2P or from storing goods to lending them out P2P). People who already have lent out goods through any type of sharing platform in the past might share their car more easily than unfamiliar car owners: they have already overcome the change of behaviour.

In addition to this, if familiar people had a pleasant experience with sharing out goods, this experience might logically trigger them to participate in the sharing economy again. Participation in the sharing economy could create ‘hedonic value’ (Hwang & Griffiths, 2017). Since this is the case for most familiar ‘suppliers’ in the sharing economy – 95% of Airbnb users for example rate either 4.5 or 5 out of 5 stars, with ratings lower than 3.5 stars being very rare (Bridges & Vásquez, 2016) – being familiar with sharing goods would also affect Sharing Intentions through a self-generated positive feeling with the sharing economy.

In short: the benefits of familiarity with lending out goods through other sharing platforms (Airbnb, Peerby, Thuisafgehaald etc.) is three-fold: it lower transaction costs, limits required

(25)

on the other hand comes with required behaviour change and transaction costs for ‘newcomers’ that would try out sharing for the first time. In the light of this research: even when Trust has increased to a high level, high transaction costs (effort) and required behaviour change may still restrain unfamiliar car owners from lending out their car. Mutual trust might be a necessary, but not always sufficient condition for participation. Familiarity is therefore expected to moderate the positive relationship between Trust and Intentions to Share. The positive effect of Trust on Intentions to Share is expected to be stronger for highly familiar car owners than for car owners who are unfamiliar with the sharing economy. This leads to the following hypothesis:

Hypothesis 4: The positive effect of Trust on Sharing Intentions is moderated by Familiarity, such that the effect is stronger for high levels of Familiarity than for low levels of Familiarity.

In order to validate the several hypotheses stated in this chapter, these hypotheses should be empirically tested.

2.5 Hypotheses and Conceptual model

The following hypotheses will be tested:

H1: A high level of Platform Positioning (Communal Sharing) leads directly to higher Sharing Intentions than a low level of Platform Positioning (Market Pricing).

H2: The positive effect of Platform Positioning on Sharing Intentions is mediated by Trust, such that high Platform Positioning (Communal Sharing) leads to higher Trust than low Platform Positioning (Market Pricing), and Trust positively affects Sharing Intentions. H2a: There is a positive relationship between Platform Positioning and Trust.

H2b: There is a positive relationship between Trust and Sharing Intentions.

H3: The positive effect of Platform Positioning on Trust is moderated by Perceived

Closeness, such that the effect is stronger for high levels of Perceived Closeness than for low levels of Perceived Closeness.

(26)

H4: The positive effect of Trust on Sharing Intentions is moderated by Familiarity, such that the effect is stronger for high levels of Familiarity than for low levels of Familiarity.

These hypotheses can be visualized in the conceptual model shown in Figure 2.

(27)

Chapter 3: Data and Method

This chapter elaborates on the applied methodology of the research. First, the choice for the car sharing platform SnappCar will be explained in paragraph 3.1. Then, paragraph 3.2 explains the research design, paragraph 3.3 elaborates on the pre-tests, followed by the discussion of data collection in paragraph 3.4. 3.5 elaborates on the relevant characteristics of the sample, followed by an indication of the applied measures in 3.6. The chapter finishes with the an elaboration on the used analyses for the testing of the hypotheses.

3.1 Choice for car sharing platform: SnappCar

In order to make the research come closest to reality as possible, a collaboration with a real life car sharing platform was desirable, and therefore a collaboration with SnappCar had been established. SnappCar is chosen as an appropriate car sharing platform for this research mainly because it is the biggest car sharing platform in the Netherlands. SnappCar currently struggles with the acquisition of new ‘cars to share’, and mainly focuses on financial motivations in its strategy. The problem intended to tackle in this research is thus a real life problem, or opportunity, for this car sharing platform. This means SnappCar is not only relevant to this research, this research very is relevant to SnappCar as well. Moreover, the P2P consumption SnappCar enables, fulfils all four requirements that are needed in order to be considered true part of the sharing economy as mentioned in Chapter 2. In addition, SnappCar is a Dutch car sharing platform, which makes the research easy to apply to Dutch speaking participants in the study.

3.2 Research design

The hypotheses have been tested in a quantitative, survey based research using a 2X2 between-subjects design, through the means of Experimental Vignette Methodology (EVM). The two manipulated variables are Platform Positioning (high/low) and Perceived Closeness (high/low) two variables of which the exact manipulations are elaborated on in paragraph 3.4. The study was

(28)

(Aguinis & Bradley, 2014). Vignettes are ‘‘short stories about hypothetical characters in specified circumstances, to whose situation the interviewee is invited to respond’’ (Finch, 1987, p. 105). EVM is used because this methodology comes closest to reality and because the technique makes it possible to develop questions within survey format of a very concrete kind (Finch, 1987). Several studies have found that as long as the message presented in the vignette appears both real (Finch, 1987) and relevant (Neff, 1979), the vignette responses are very closely related to how people respond in real life situations (Hughes, 1998). Moreover, Caro, Ho, McFadden, Gottlieb, Yee, Chan and Winter (2012) found that vignettes that are audio and/or visual based, like the (existing but adapted) landing page of SnappCar, make participants engage even more fully than vignettes that contain written information alone (Caro et al., 2012).

The manipulation of the two variables Platform Positioning (High/Low) and Perceived Closeness (High/Low) result in the four vignettes shown in Table 1. The development of these vignettes will be further elaborated on in paragraph 3.4.1.

Table 1

Applied vignettes

Platform Positioning = High

a

Platform Positioning = Low

b

Perceived Closeness = High

c

Vignette 1

Vignette 2

Perceived Closeness = Low

d

Vignette 3

Vignette 4

Note.

a_{=Communal Sharing,}b_{= Market Pricing,}c_{=4 mutual friends,}d_{=nothing shown}

The hypothetical situation used in this study was the same for the four vignettes and stimulated people to imagine they own a car if they do not do so in real life. Since this hypothetical situation uses everyday people and everyday events, the vignettes are both real and relevant and are therefore highly probable to trigger outcomes that are similar to actions in reality (Hughes, 1998).

The participants were firstly informed about their participation being anonymous, the expected duration of the survey and there being no right or wrong answers. Moreover, they were

(29)

informed about SnappCar as it being a car sharing platform that enables consumers to rent out their car P2P online. Each participant in the study was randomly allocated to one of the four vignettes: one of the four manipulated website landing pages of SnappCar. They observed this landing page of SnappCar for at least 50 seconds, after which they could continue if they were finished reading. Then research specific questions were asked, followed by demographics. The full survey is attached in Appendix A.

The distribution of participants across the four vignettes is presented in Table 2.

Table 2

Distribution of participants across vignettes

Vignette

N

Percentage (%)

1: High PP

a

/ High PC

b

88

25.1 2: High PP / Low PC

86

24.5 3: Low PP / High PC

88

25.1 4: Low PP / Low PC

89

25.4 Total

351

100 Note.

PP = a_{Platform Positioning,}b_{PC = Perceived Closeness}

3.3 Measures

This paragraph elaborates on the way the variables are measured. Since the applied scales were all originally in the English language, a technique called back-translation has been used in order to translate the scales to Dutch. Back-translation is a widely applied, reliable technique to translate scales in academic research (Brislin, 1970).

3.3.1 Car owners’ Sharing Intentions

(30)

Likert scale (1. I strongly disagree – 7. I strongly agree), with items mixed from previously used scales by Pavlou and Geven (2004) (α= .94) and White, MacDonald and Ellard (2012) (α= .84). These scales are mixed in order to increase the amount of items that form the construct, since this leads to better construct representations (Little, Lindenberger & Nesselroade, 1999). One example of the five items is ‘‘If I would have the possibility, I would consider sharing my car through SnappCar in the future’’ (‘‘Als ik de mogelijkheid zou hebben, zou ik overwegen om mijn auto in de toekomst te delen via SnappCar’’). The Cronbach’s alpha (α) in this current research is .95. The corrected item-total correlations give an indication that all the items have a good correlation with the total score of the scale (they are all > .30). Moreover, Cronbach’s alpha would not increase with Δ <.10 if one of the items would be deleted. Since this scale is formed by combining two separate scales, a principal axis analysis (PAF) was conducted as well in order to make sure these items complement each other well, without forming two separate groups. The assumptions for the test were sufficient for execution: the Kaiser-Meyer-Olkin measure verified the adequacy of the sampling for the analysis (KMO = .892), and Barlett’s Test of Sphericity was significant (χ² (10) = 1757.073, p <.000), indicating homogeneity of variances. A factor analysis was run to obtain Eigenvalues for each existing component in the data. Only one component had an Eigenvalue >1, in combination with 82,3% of the variance explained. In addition, the scree-plot showed a clear levelling off after the first factor. These results show that the items within Sharing Intentions complement each other very well, without forming two separate factors. The created scale for Sharing Intentions is considered reliable and can be used for analysis.

3.3.2 Trust

The mediating variable Trust is measured by five items on a 7-point Likert scale previously used by Möhlmann (2015) and Pavlou and Gefen (2004) (α= .93). One of the items that is considered

representative is: ‘‘The peers on SnappCar are generally reliable’’ (‘‘De deelnemers op SnappCar zijn over het algemeen betrouwbaar’’). A high reliability is indicated for the measure Trust, since

Cronbach’s alpha in this research is .88, which would not increase with Δ <.10 if one of the items would be deleted and all items have a good correlation with the total score of the scale (all > .30).

(31)

3.3.3 Perceived Closeness

The moderating variable Perceived Closeness is measured by a single item on a 5-point Likert scale previously used by Moran (2015) and Baer (2012). The measuring item is ‘‘How connected do you feel with other SnappCar members?’’ (‘‘Hoe verbonden voel jij je met andere leden van SnappCar?’’) with 1= ‘not at all’, to 5 = ‘very connected’. Even though this scale consists of only a single item, according to Moran (2015) it is a common measure in network analysis. Since Moran’s article is published in the Strategic Management Journal, which is ranked one of the top strategy and

management journals in the field (SCImago, 2007), this scale is assumed to have a high validity and reliability.

3.3.4 Familiarity

The moderator Familiarity is measured by three items on a 7-point Likert scale previously used by Möhlmann (2015) and Lamberton and Rose (2012) (α= .92). One of the items is: ‘‘I have experience lending out goods through sharing platforms myself’’ (‘‘Ik heb zelf ervaring met het uitlenen van spullen via deelplatformen’’). The Cronbach’s alpha in this research is reported to be .81, all items correlate well with the total score (corrected item-total correlations >.30) and the Cronbach’s alpha would not increase with Δ <.10 with deletion of any item. Therefore, the measure Familiarity is considered reliable.

3.3.5 Control variables

In a research commissioned by the European Union, Andreotti et al. (2017) found that age, gender, level of education and urbanity are strongly correlated with interest in the sharing economy (Andreotti et al., 2017; Newcom, 2017). Therefore, the variables age, gender, level of education, and urbanity are included as control variables. Also car possession and the possession of a driver’s license are

controlled for. The control variable age is measured on a ratio scale, as an open ended question. Gender is measured as a nominal variable with three options (man / woman / other), urbanity with seven options (Rotterdam, Amsterdam, Den Haag, Other city in Randstad, other city outside

Randstad, village). Level of education is measured as an ordinal variable with eight answering options (primary school / high school / MBO / HBO / WO Bachelor /WO Master / PhD/ Other). Car - and

(32)

driver’s license possession are measured as a nominal variable with two options, ‘‘yes’’ and ‘‘no’’. Finally control questions and regarding credibility and believability of the vignettes were added.

3.4 Pre-tests and manipulations

To test the validity of the experiment, an exploratory quantitative pre-test was developed and executed before the actual experiment was executed. The pre-test checked whether the adapted SnappCar landing pages (vignettes) displayed what they were supposed to display. In other words: a test was done to find out whether people truly perceive SnappCar through the Market Pricing landing page as a platform that facilitates transactions similar to those that occur in business relationships; whether people truly perceive SnappCar through the Communal Sharing landing page as a platform that facilitates transactions characterized by a high level of generosity, on which people have the feeling to belong to the same group (CS), or as a platform that facilitates transactions based upon equality and reciprocity, on which people keep an healthy balance concerning the level to which participants benefit (EM). This was tested on a 7-point Likert scale (1= not at all, 7 = very much), asking

respondents to what extent they think SnappCar is a platform with certain characteristics. In specific, respondents were asked to what extent they think that SnappCar is: ‘‘a platform on which transactions are characterized by a high level of generosity. On this platform people have the feeling that they belong to the same group and have a lot in common’’ (CS), ‘‘an online sharing platform that

facilitates transactions based upon equality and reciprocity. People on this platform try to maintain a healthy balance in terms of the benefits everyone gets on the platform (money or stuff)’’ (EM) and ‘‘a platform is one that manages transactions based on individual utility, where people believe they are entitled to a good return on their investment (money or stuff). This is a little bit like a business relationship” (MP).

Moreover, this pre-test also tested whether the manipulation of Perceived Closeness worked. In other words: whether showing four mutual Facebook friends (between car owner and user) makes car owners feel closer to a potential car user than when zero mutual Facebook friends are showed. This was done by asking ‘‘How connected do you feel with other SnappCar members?’’ on a 7-point Likert scale. At last, the credibility of the landing page was tested by asking ‘‘How credible do you

(33)

think the message of SnappCar is?’’ on a 7-point Likert scale (‘Not at all’ – ‘Very much’). The full pre-test survey can be found in Appendix B.

3.4.1 Applied manipulations

In order to trigger Market Pricing or Communal Sharing relationships and to make participants feel more or less connected to potential car users, the SnappCar landing page has been adapted and four vignettes have been created. The difference between high and low Platform Positioning (CS/MP) was created by following both recommendations of Habibi, Davidson and Laroche (2017) about

communicating social/financial ‘sharing’ (Habibi et al., 2017), and recommendations of Bridoux and Stoelhorst (2016), concerning words that trigger either CS or MP. This mainly means that the CS version of the website had to emphasize socialization, focus on community building and had to avoid calculations or references to money, whilst the MP version of the website had to deprioritize

community building and emphasize the calculated monetary and utilitarian benefits (Habibi et al., 2017). Moreover, the CS version should use words as ‘we’ and ‘us’ rather than ‘I’ and ‘you’ and should emphasize common identity, whilst the MP version should do the opposite (Bridoux & Stoelhorst, 2016).

This recommendations were implemented by adapting the following aspects of the website. Firstly the pictures were either very social (CS) or not social at all (MP) in order to (not) emphasize socialization and a community feeling. The MP landing page had pictures of a car and car keys whilst the CS landing page had pictures of a happy people with cars. Furthermore, where the MP landing page tried to persuade participants with mainly financial arguments as ‘‘Earn up to €1000 per year’’, the CS landing page presented arguments as ‘‘Car sharing is a fun way to help your neighbours and to get to know them better’’ and focused on the SnappCar community. Financial calculations were avoided in the CS version. The slogans were ‘‘Rent out your car easily and safe through SnappCar and earn up to €1000/year’’ for MP and ‘‘Make your (singular) neighbourhood your (plural)

neighbourhood again’’ for CS. The MP slogan focused on financial and utilitarian benefits, whilst the CS slogan awakened a sense of socialization and community. Lastly the CS version communicated in the you (plural) version where possible instead of using you (singular).

(34)

The difference between high and low Perceived Closeness was first of all manipulated by showing 4 mutual Facebook friends between the participant (car owner) and car user (high) or showing 0 mutual Facebook friends (low). This idea comes from the Scandinavian car sharing platform Skjutsgruppen (Skjutsgruppen, 2018). It is in line with theory since ‘shared friends’ according to Luo (2002) and Zhao et al. (2012) trigger a feeling of similarity which is a form of Perceived Closeness. Later on, this changed into showing 4 mutual Facebook friends or showing nothing at all because the former manipulation did not work. The four final landing pages or vignettes are shown in Figure 3, 4, 5 and 6.

3.4.2 Results pre-tests

Two pre-tests were needed in order to make the manipulations work as intended. A Pierson’s Correlation Matrix was formed and resulted in the following. The manipulation used in the vignettes in the first pre-test (N=73) worked for Platform Positioning, but not for Perceived Closeness, as can be seen in Table 3. In other words: the Market Pricing landing page made SnappCar being truly perceived as a platform that facilitates transactions similar to those that occur in business relationships (r=.283; α=.015) and the Communal Sharing landing page made SnappCar being truly perceived as a platform that facilitates transactions characterized by a high level of generosity, on which people have the feeling to belong to the same group (r=-.453; α=.000). However, showing either 0 mutual

Facebook friends or 4 mutual Facebook friends did not significantly correlate with the Perceived Closeness to other members of SnappCar (r=.137; α=.249).

(35)

This means the manipulation of Perceived Closeness through Facebook did not yet work. Therefore, the vignettes were adapted: the Platform Positioning was not changed since it already was significant, but the Perceived Closeness manipulation was. Instead of showing either 4 or 0 mutual Facebook friends, Perceived Closeness was manipulated by showing nothing at all (low) or showing 4 mutual Facebook friends (high) between car owner and user. Then, the second pre-test was executed (N=72). Those results are presented in Table 4.

Table 3

Pearson Correlations between variables of Pre-test 1

Variable

M

SD

1.

2.

3.

4.

5. 1. Platform Positioning

a

0.562

0.500 2. Type 1 CS

4.16

1.500 .283*

3. Type 2 EM

4.95

1.153 .102

.094

4. Type 3 MP

4.41

1.153 -.453 -.339

.226

5. Facebook

b

0.589

0.495 .104

-.057

.130

.043

6. Perceived Closeness

5.96

1.409 .215

.351**

.258*

.008

.137

Note.

a_{reference category is ‘’Low’’ (MP)}b_{reference category is ‘’0 mutual Facebook friends’’}

N= 73

* correlation is significant at the 0.05 level ** correlation is significant at the 0.01 level

(36)

Table 4

Pearson Correlations between variables of Pre-test 2

Variable

M

SD

1.

2.

3.

4.

5. 1. Platform Positioning

a

0.500

0.503 2. Type 1 CS

4.15

1.725 .251*

3. Type 2 EM

4.86

1.225 -.137

.170

4. Type 3 MP

4.26

1.768 -.293*

-.397**

5. Facebook

b

0.4583

0.501 .028

0.374**

.082

-.186

6. Perceived Closeness

5.94

1.537 .091

.449**

0.18 -.362**

.344**

Note.

a_{reference category is ‘’Low’’ (MP)}b_{reference category is ‘’nothing shown’’}

N= 72

* correlation is significant at the 0.05 level ** correlation is significant at the 0.01 level

These results show that again the Platform Positioning was valid: the Platform Positioning correlates in a significant way with the corresponding perceived characteristics of the landing pages (r=.251; α=.033 for CS and r=-.293; α=.013 for MP), but this time also the manipulation of Perceived

Closeness through (not) showing mutual Facebook friends worked. Showing either nothing (low) or 4 mutual Facebook friends (high) did significantly correlate with Perceived Closeness (r=.344; α=.003). Therefore, the vignettes were now ready for use.

(37)

Why sharing should be about caring again : a vignette study measuring the impact of platform positioning, trust, perceived closeness and familiarity on car owners’ car sharing intentions