Trust in the sharing economy: its antecedent and two-sidedness
of P2P platforms
Gabija Baugirdaite 11450290 Bachelor thesis
BSc Business Administration- Management in the digital age University of Amsterdam
Nicole Stofberg 29 June 2020
Statement of Originality
This document is written by Gabija Baugirdaite who declares to take full
responsibility for the contents of this document.
I declare that the text and the work presented in this document are 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.
Abstract
Peer-to-peer (P2P) sharing platforms utilize underused assets, provide instant access to those in a
need of it, and create welfare for society at large. As platforms act only as intermediaries, sharing
businesses function on created trust- trust in the platform that is enabling the service and in the
community of peers that users will have to deal with. It is verified that both types of trust lead to
a willingness to participate, but it is still not fully discovered what creates trust itself.
Additionally, research is scarce on whether trust means more to private owners who supply the
resources, versus users who need access to those resources. Whilst debate in the research
suggests that owners are more susceptible to trust as they face higher risk, no study has
compared these two types of consumers in sharing contexts so far. To address these gaps in the
literature, this research explores the degree of platform intermediation (i.e. structural assurances,
guarantees, the intensity of platform involvement), its contribution to higher levels of both types
of trust, and whether it translates into more willingness to participate. Moreover, the importance
of trust for owners versus users is analyzed herein. It is done by generating 8 vignettes with
different scenarios and manipulations as well as collecting respondents’ attitudes and opinions in
the form of a closed-ended questionnaire (N=485). Findings proved that peer and platform trusts
directly increase willingness to participate. Additionally, high intermediation has been found to
impact the likelihood of participation in two indirect ways. Firstly, by increasing platform trust
which leads to the willingness to participate. Secondly, through the trust transfer- more
specifically, by leading to platform trust, which results in peer trust, and then it all culminates in
the upsurge of willingness to participate. Lastly, it was found that even though owners are less
likely to participate in sharing, both peer and platform trusts matter equally to owners and users.
Table of Contents
1. Introduction ... 1
2. Literature review and hypotheses ... 4
2.1 Sharing economy ... 4
2.2 Trust in the sharing economy ... 6
2.2.1 Trust in the platform... 7
2.2.2 Trust in peers ... 8
2.3 Platform intermediation ... 9
2.4 Role of the consumer... 13
3. Methods... 16
3.1 Design ... 16
3.2 Vignette design ... 17
3.2.1 The impact of COVID-19 ... 17
3.2.2 Design ... 17 3.2.3 Vignette population ... 19 3.3 Survey questionnaire ... 20 3.4 Pre-test ... 22 3.5 Procedure ... 23 3.6 Sample ... 23 3.7 Analysis ... 24 4. Results ... 25 4.1 Distribution ... 25 4.1.1 Manipulation checks ... 25 4.1.2 Correlations ... 26 4.1.3 Control variables ... 26 4.2 Hypotheses testing... 28 4.2.1 Hypothesis 1 ... 29 4.2.2. Hypothesis 2 ... 31 4.2.3 Hypothesis 3 ... 32
4.2.4 Hypothesis 4 ... 33
4.2.5 Hypothesis 5 ... 34
4.2.6 Hypothesis 6 ... 34
5. Discussion... 36
5.1 Summary of the results ... 36
5.2 Discussion of the results ... 37
5.2.1 Direct link between both types of trust and willingness to participate ... 37
5.2.2 The role of platform intermediation and trust ... 38
5.2.3 Two-sided view of the platform on the willingness to participate ... 39
5.3 Theoretical implications ... 40
5.4 Managerial implications ... 41
5.5 Limitations and recommendations for future research ... 42
6. Conclusion ... 44
Reference list ... 46
List of Tables
Table 1. Correlations matrix ... 28
Table 2. Direct effect size and significance levels of hypothesis 1 ... 30
Table 3. Direct effect size and significance levels of hypothesis 2 ... 31
Table 4. Effect size and significance levels of mediation ... 32
Table 5. Effect size and significance levels of sequential mediation... 33
Table 6. Effect size and significance levels of hypothesis 5 ... 34
1
1. Introduction
Sharing economy has promised society plenty of benefits- the possibility of self-employment and
money savings, more business opportunities, instant gratification, and easier access to assets
(Crowdholding, 2017), but did not warn anyone about the possible risks and dangers. Especially
in P2P (peer-to-peer) sharing, it is difficult for platforms to gain consumers’ trust when
occasionally various scandals arise- for example, a few of the most famous ride-sharing
platforms Uber and DoorDash have faced accusations of not protecting their employees and
customers from sexual harassment cases in the workplace (Levin, 2017). As P2P firms rely on
relationships between their customers, these kinds of wrongdoings significantly undermine their
reputation, not to mention the ability to expand their customer base. This is where the concept of
trust comes in- businesses must constantly work on ensuring and maintaining trust for and
between consumers as it is the foundation of their company (Gefen, 2000; 2002).
P2P sharing platforms are known to consist of two types of trust, namely peer and
platform (Chai, Gong & Li, 2011; Chai & Kim, 2010). It means it is not enough to trust the
company that connects you with others for renting or lending out, but it is also necessary to
establish trust between consumers themselves. The importance of platform and peer trusts is
known for more than a decade, nonetheless, research into the antecedents of trust in P2P sharing
platforms is lacking. High intermediation is possibly one of the antecedents since it reduces risks
associated with sharing activity by ensuring guarantees, customer support, and insurances against
deception (Perren & Kozinets, 2018; McKnight & Chervany, 2002). The struggle of mitigating
risk and ensuring trust gets more complex with the realization that the platform consists out of
2 widely explored, but it is already known that both sides have varying intentions, face different
risks, and take distinct actions (Ertz, Durif & Arcand, 2016), thus both sides may require
attention on different aspects.
In order to investigate the presented topics further, this study focuses on both peer and
platform trusts, platform intermediation as a predictor of trust, and the two-sidedness of
platforms, namely supply of providers and demand from users. Firstly, to highlight the
importance of both types of trust and to confirm previous findings, it is tested whether they lead
to a willingness to participate (WTP) in the sharing activity. Secondly, it is tested whether high
platform intermediation affects trust in the platform, in the end leading towards a higher
likelihood of participation. In addition to that, it also analyses the trust transfer in sharing
platforms discussed in the literature (Mohlmann, 2016)- it is predicted that high platform
intermediation will affect willingness to participate through both types of trust sequentially.
Lastly, the role of the consumer is debated, more specifically, whether owners are more
susceptible to peer and platform trusts than users. It all results in the following questions:
1. What is the effect of platform trust and peer trust on willingness to participate?
2. What is the effect of platform intermediation on platform trust and willingness to participate?
3. What is the sequential mediating role of platform and peer trusts on the willingness to participate?
4. What is the effect of the role of the consumer on the relationship between both types of trust and willingness to participate?
3 These predictions are tested by simulating a P2P car-sharing platform SplitCar and changing its
factors in each vignette. In total 8 vignettes are generated in a 2 (presence vs. absence of
COVID-19) x 2 (high vs. low platform intermediation) x 2 (owner vs. user) design. After
presenting a situation, a closed-ended questionnaire is used to record respondents’ attitudes and
understanding of the manipulations.
The study adds to the existing literature by extending the understanding of trust in sharing
platforms as well as willingness to participate. It discovers high platform intermediation as being
one of the predictors of platform trust and the likelihood of participation. Additionally, as the
literature is very scarce on two-sidedness of the platform, this study boosts the knowledge about
the role of the consumer- both types of trust have been found to matter equally for owners and
users, but it is proven that significant differences between the two exist. Managers can use these
insights for generating trust in and on the platform, as well as acquiring and maintaining new
consumers.
This thesis begins with a literature review, presenting gaps in the research and
corresponding hypotheses (chapter 2), followed by a methods section where the study design is
explained together with the main variables (chapter 3), and results (chapter 4). The discussion
debates the findings, considers theoretical and managerial implications together with limitations
4
2. Literature review and hypotheses
The literature review introduces the main concepts and begins with defining and discussing the
sharing economy (paragraph 2.1), followed by trust in the sharing economy (paragraph 2.2),
platform intermediation (paragraph 2.3), and ends with the role of the consumer (paragraph 2.4).
2.1 Sharing economy
The sharing economy is actively researched in the academic literature together with its impact on
society and the environment (Frenken & Schor, 2019; Benjaafar, Kong, Li & Courcoubetis,
2019; Schor & Cansoy, 2019). Nevertheless, this concept varies by definition in many different
papers. Starting as a “collaborative consumption” in Botsman and Rogers’s (2010) work, it can be also called “access-based consumption” (Bardhi & Eckhardt, 2012), “co-production”
(Humphreys & Grayson, 2008) or simply “sharing economy” (Andersson, Hjalmarsson & Avital,
2013; Hawlitschek, Teubner & Weinhardt, 2016), all which imply unraveling the value of excess
capacity by matching supply and demand (Eisenmann, Parker & Van Alstyne, 2006). According
to Frenken and Schor (2016), confusion in definitions was caused by different pseudo forms of
sharing such as business-to-consumer (B2C) as it does not actually utilize idle capacity, thus this
research will focus on peer-to-peer (P2P) sharing. To simplify and discuss sharing economy from
the provider’s as well as user’s sides, we will be using Mittendorf, Berente & Holten’s (2019,
p.1085) definition which describes it as “a two-sided market model that allows private
individuals to share resources with potential customers through an online platform in the form of
5 famous companies include BlahBlah Car for ride-sharing, Udemy for knowledge sharing,
JustPark for parking space sharing, and Turo for car sharing.
Together with the alternating definitions, sharing economy has been a controversial topic
in the academic literature (Schor, 2016)- since P2P platforms started gaining acceptance in the
society, various benefits were discovered such as social engagement (Kim, Yoon & Zo, 2015),
increased savings and life convenience (PWC, 2015), increased sustainability (Nica &
Potcovaru, 2015) and expressed interest in globalization (Parente, Geleilate & Rong 2018). With
that, different problems arose such as lack of regulation (Koopman, Mitchell & Thierer, 2014),
monopolism (Frenken & Schor, 2019), or trust issues (Hawlitschek et al., 2016). No matter how
debatable existence of the sharing economy is, there is no doubt it is transforming the way
businesses work, compete and Uber or Airbnb are just a few of the most impactful examples
(Wallsten, 2015; Zervas, Proserpio & Byers, 2017). The sharing economy is growing quickly,
and it is predicted that revenues from sharing activity in industries such as music, temporary
housing, and car rental will hit 50% by 2025, just because they were transformed by
technological features and sharing possibilities. According to The European Commission’s
(2018) findings, nearly 25% of Europeans have used sharing platforms but only 6% have offered
their services. These results not only show increasing importance of sharing economy but also
emphasize two-sides of platforms, namely users and providers (or demand and supply). While
former participants are only using services and paying for the temporary access of the asset
(depicting demand side), latter ones provide their valuables in order to utilize excess capacity
and earn money for that (depicting supply side). In sharing platforms one can also be a
prosumer- so-called producer of supply (provider) and consumer (user) at the same time (Ritzer,
6 the continuum have contrasting functions, goals, and take different actions (Ertz et al., 2016),
thus their behaviors and intentions are also particular. To justify the importance of both sides of
the platform, Ter Huurne, Ronteltap, Corten and Buskens (2017) touch upon this topic by stating
that providers are facing serious exposure and misuse risks as much as users do, meanwhile the
focus in the literature has been only on trust from the user’s side. As two-sides of sharing
platforms have not been fully analyzed in academic literature, in this study we will differentiate
between the two in order to better predict their roles and attitudes.
2.2 Trust in the sharing economy
Even though sharing economy is very promising and significant in consumers’ lives nowadays, it
is facing severe obstacles that hinder its growth and ability to meet expectations. A major barrier
to this is stranger-danger- risk and fear induced by the fact that you are interacting with an
unknown stranger without knowing his intentions- which prevents consumers from participating.
This has been proven by PwC (2015) where 69% of US adults stated they would not trust sharing
economy firms until they are recommended by somebody they trust, thus in order to understand
how the sharing economy can grow up to its full potential, it is crucial to understand how
platforms can build trust.
Trust is being known as one of the foundations for e-commerce activities (Gefen 2000,
2002; Pavlou & Gefen, 2004). In this paper, we define trust based on Luhmann’s (1989)
definition which describes it as a mechanism to curtail the complexity of human conduct where
uncertainty is present. Trust exists only in uncertain and risky environments, where a participant
7 trustor, his experience, willingness to trust as well as others’ experience with trustworthiness (Grabner-Kraeuter, 2002). Furthermore, the online environment is risky and questionable as it is
very complex and facilitates socially distant relationships (Jarvenpaa, Tractinsky & Saarinen,
1999; Kim, Ferrin & Rao, 2008), thus it is proven that trust helps to overcome the fear of
negative consequences such as deception and misuse of information (Gefen 2000; Rousseau,
Sitkin, Burt & Camerer, 1998). Trust has also been proven to be a fundamental part of the
sharing economy- it is not only being positively associated with users’ intentions in it but also
referred to as one of the drivers for participation in P2P rental (Hawlitschek et al., 2016). When
discussing trust in P2P online-based settings, trust has been found to be a two-fold concept
consisting of trusts in the platform and peers which interrelate to each other while constituting a
sharing platform (Chai, Gong & Li, 2011; Chai & Kim, 2010).
2.2.1 Trust in the platform
Trust in the intermediary can be described as trust in the entity which owns a digital platform
(Mittendorf et al., 2019). It neither owns nor provides assets for sharing, only facilitates the
encounter of users and providers (Belk, 2010). The probability that a participant will engage in
sharing activity either as a user or a provider increases as trust in the intermediary builds up
(Mittendorf et al., 2019), thus it is already known that this type of trust is important for customer
acquisition- if participants trust the platform they are interested in, they will more likely engage
in sharing themselves, without the company needing to put enormous effort in encouraging
participants to join. It is not much known about specific effects on P2P sharing platforms, only
8 (Hong & Cho, 2011). On the other hand, in P2P example participants are more socially active in
between each other and express a strong focus on trust in peers.
2.2.2 Trust in peers
Trust in peers can be described as trust in sharing partners or simply individuals who exchange
resources with each other via sharing platforms (Mittendorf et al., 2019). Same as trust in an
intermediary, it increases the likelihood of participants engaging in sharing activity (Chen et al.,
2009; Hawlitschek et al., 2016; Jiang, Huang & Chen, 2009; Mittendorf, 2018). Logically, this
kind of trust decreases risk and uncertainty associated with participating in the exchange such as
deceit, fear of interacting with a stranger, or financial fraud. Both users and providers are facing
those risks- as a provider, you are sharing your asset and you have to rely on the user not to take
advantage of it. As a user, you rely on a provider in acquiring temporary access. Stranger-danger
phenomenon is very likely in sharing encounters as everyone included in the transaction is
unfamiliar with each other and do not know each other’s intentions, but trust in peers might help
to overcome it. Nonetheless, even though the positive effects of it are quite known, there is
ambiguity regarding how trust in peers is generated.
The lack of research on drivers of trust, especially in sharing platforms, is leaving a big
gap in the literature. Nevertheless, the trust itself turns out to be a significant factor in enhancing
the willingness to participate in the sharing economy. To highlight the importance of both types
of trust, the following hypotheses are formulated:
H1: Trust in the platform leads to more willingness to participate in sharing economy.
9
2.3 Platform intermediation
Platform intermediation in our study is a combination of platform mediation and structural
assurances. Mediation can be described as a degree of control intermediary is exerting in order to
manage processes and activity on the platform. In simple terms, it is the extent to which
intermediary is arranging everything- from connecting two sides of the platform to processing
payments- instead of users themselves. On the other side, structural assurances “like guarantees,
regulations, promises, legal resources, or other procedures are in place to promote success”
(McKnight & Chervany, 2002). Perren and Kozinets (2018) describe structural assurances as
systems constructed to provide satisfactory customer support, audit their members, provide
insurances in case of mishap or deception, all in order to alleviate risk concerns. In some cases,
institution-based trust (Pavlou & Gefen, 2004) or technology trust (Ratnasingam & Pavlou,
2003) are used to define the same concept. Pavlou and Gefen (2004) describe four ways platform
intermediation can reduce risk- by reassuring benign transaction norms, establishing clear and
fair rules and procedures, providing a secure and trustworthy environment as well as noticing
and eliminating problematic users. No matter different definitions, this construct indicates
protection for consumers, especially in the case of a loss. Airbnb is a great example where loss
can be of significant importance- if you are renting out your apartment and furniture gets broken,
you would want to be compensated for it, no matter if the platform reimburses it or the causer of
the damage. Sha (2009) has found that two most influential types of structural assurances exist,
namely perceived vendor-specific guarantees and perceived seal of approval guarantee- this
means that customers trust customer service policies, return guarantees and privacy guarantees
the most. Even though the study was conducted for vendor businesses, it emphasized the need
10 combined both features to define platform intermediation, as realistically platform cannot have
strict assurances without mediating the majority of good exchange process (and vice versa).
In high intermediation platforms, sharing medium is acting as a connector between users
and providers, resolving all the matters related to the temporal exchange of valuables, thus they
can be described as centralized systems. Those kinds of platforms actively supervise their
participants as a quality ensuring mechanism (Deng, Joshi, & Galliers, 2016; Kuhn &
Maleki,2017), have a strict screening process for users, and include structural assurances (Perren
& Kozinets, 2018). For example, direct data collection helps the platform to ensure the
credibility of a participant, liability insurance guarantees safe service in case of accidents, and
developed customer service ensures high quality as well as provides help in need. This
centralization has its own benefits- research from Lampinen and Cheshire (2016) has shown that
high platform mediation increases the convenience of connecting with participants as well as
trust in transaction on both sides- this way users and providers avoid the hassle of handling
financial transactions between them, refrain from logistical inconveniences. Furthermore, Jones
and Leonard (2008) found that third-party validation is one of two critical factors influencing
participant trust in P2P settings. On the other side, while creating favorable and seamless
connectivity solutions for both sides of the platform, an intermediary is giving up the social
aspect of people networking which is creating a feeling of mutual trust between participants.
Meanwhile, low intermediation platforms distinguish themselves by giving most of or
absolute freedom to participants to facilitate trust, sociality, and sense of community.
Participants are responsible for executing temporal exchange and financial transactions
(Sutherland & Jarrahi, 2017). In this case, an intermediary is providing the platform for
11 (2018), these platforms can be called forums that are solely connecting actors using it. They
distinguish themselves by having lots of social user experience as consumers execute all
transactions themselves and negotiations often conclude in offline in-person meetings to
exchange goods. In low intermediation environments trust is rooted in interpersonal social
interactions and it includes virtual communication as well as face-to-face contact.
Even though it is easy to define platform intermediation, it is not so straightforward in the
literature and some findings are debatable. Regarding platform mediation, Sutherland and Jarrahi
(2018) are looking at mediation from centralization point of view (as in high mediation is
centralized, low mediation is decentralized) and clarify that platform centralization should not be
viewed as a one-dimensional continuum- platforms can fully handle financial transactions but
give absolute freedom for making connections and personal interactions. It is up to the company
to choose a degree of mediation and this can cause ambiguities, but many sharing platforms are
leaning towards high or low mediation.
Additionally, research has been showing contradictory results about structural
assurances’ influence on trust. While Teo and Liu (2007) discovered that structural assurances are connected to participant trust across United States, Singapore, and China, Gefen, Karahanna
and Straub (2003) have shown that it can be the most substantial factor while perceiving online
platform’s trustworthiness. An increase in trust can be partly explained by the fact that intermediaries with structural assurances in place have to ensure the validity of exchanges
through escrow or insurance measures (Carroll and Bellotti, 2015; Weber, 2014). Also, it has
been proven to be beneficial in assuring the security of transactions (Ha & Son, 2014) and
managing trust by establishing sanctions for those who fail to fulfill obligations and by
12 the other hand, Wakefield, Stocks and Wilder (2004) found structural assurances to be the most
irrelevant factor in predicting perceived trustworthiness. These contradictory findings may be
due to the fact that the latter research has been discussed in the context of e-retail where
consumers care more about user experience than security, which totally differs from concerns
expressed in the sharing economy where goods are being exchanged between consumers. As
these studies isolate assurances and research them alone, results are difficult to apply in P2P
sharing platforms where structural assurances and platform mediation coincide.
As mentioned above, willingness to participate in the sharing activity is indirectly
influenced by trust in the platform. Platform intermediation consists out of platform mediation
and structural assurances- processes on the platform itself to ensure security, thus it is reasonable
to predict that willingness to participate will be higher in a high intermediation context, resulting
in the following hypothesis:
H3: The relationship between levels of intermediation and willingness to participate is
mediated by trust in the platform, such that high platform intermediation is associated
with higher levels of platform trust, which positively influence willingness to participate.
These structural aspects do not necessarily affect trust in peers as it is a more social construct
which does not relate to the technical aspects of the platform. Nevertheless, trust in peers is an
essential factor contributing to the willingness to participate (Mittendorf, 2018), only not being
directly affected by platform intermediation. According to Steward (2003), trust transfer is a
phenomenon when trust is being transferred from one source to another in a hierarchical order.
13 the platform and describes it as a two-fold and hierarchical structure of the trust concept. In other
words, when people trust the intermediary which is handling their transactions and information,
they ultimately trust the people they interact with as well. This is most likely occurring because
people in P2P collaborative contexts place institutional security first as they might be
considerably financially hurt and need some assurance of protection, which cannot be provided
by peers. Only after they feel safe in a sharing environment, they can engage in social
interactions and reap the benefits of it. Based on that, the fourth hypothesis is as follows:
H4: The relationship between levels of intermediation and willingness to participate is
sequentially mediated by platform trust and peer trust, such that high intermediation is
associated with higher levels of platform and peer trusts, which lead to higher willingness
to participate.
2.4 Role of the consumer
When engaging in sharing activity, both users and providers have different motivations to do so
and as discussed, vary in their intentions on the platform. It is proven that the main incentives for
providers are economic gains and additional income (Dabbish, Lee, Kusbit & Metsky, 2015;
Ikkala & Lampinen, 2015; Lampinen & Cheshire, 2016; Teodoro, Ozturk, Naaman, Mason &
Lindqvist, 2014). Nevertheless, only a very small percentage is participating as a provider
(Eurobarometer, 2016). Mostly because establishing trust is a big challenge (Luca, 2017), certain
providers are afraid of discrimination for their race or gender (Kricheli-Katz & Regev, 2016;
Edelman & Luca, 2014; Hannak et al., 2017) or the fear that low-income areas would not benefit
14 assets are being shared (such as property or vehicles), providers express a serious concern
towards if they can trust the user and if they would be compensated in case of misuse. As these
assets are high in value, the damage can considerably affect the provider’s financial situation and
be time-consuming. For service providers, it might be even more immense worry as it includes
their own time and mental state. Thus, before providing their valuables for sharing, providers
will most likely actively participate in user screening and relying on their positive intentions.
Additionally, assurance that they will not be hurt financially is needed to protect their valuables
and reduce the risk that they will be exploited.
Meanwhile, users are mostly encouraged by social interactions (Hawlitschek et al., 2016)
and interest in the sustainable activity (Plenter, Fielt, Hoffen, Chasin & Rosemann, 2017;
Kathan, Matzler & Veider, 2016). The biggest concerns for users include high perceived effort,
too much legal and economic risk as well as the fear that needed resources might not be available
at a given time (Hawlitschek et al., 2016). Furthermore, the lack of trust, efficacy, and economic
benefits were identified as deteriorating user participation in accommodation rental services
(Tussyadiah, 2015). Nonetheless, users are not facing risks as big as providers because they do
not invest substantial amounts into obtaining a good and would not financially suffer as
significantly. If a user acts according to the sharing platform’s policies, most likely they can be
compensated in monetary terms. Even if a user is misled and cannot temporarily obtain the good
as arranged, it is possible to use substitutes without vast adversity.
Sharing platforms depend on recruiting users as much as providers and one of the focal
reasons why sharing businesses fail is the lack of providers (Chasin, von Hoffen, Cramer &
Matzner, 2018). Considering that only 6% of Europeans have participated in sharing activity as a
15 providers are harder to acquire, mostly because of the above-mentioned reasons. To overcome
risk and fear barriers, trust must be established to encourage them, ensure a safe environment,
and increase willingness to participate. As we already know that trust in peers and platform
indirectly affect willingness to participate, we argue that providers rely on these types of trust
more intensively than users, thus the following hypotheses are proposed:
H5: When a consumer is a provider than a user, trust in peers leads to more willingness to
participate.
H6: When a consumer is a provider than a user, trust in the platform leads to more
16
3. Methods
This section discusses the choices related to data gathering. Firstly, the research design is
explained (paragraph 3.1), and it is followed by vignette design clarifying their content
(paragraph 3.2). Further on, a survey questionnaire is reviewed (paragraph 3.3) after which
pre-test (paragraph 3.4), procedure (paragraph 3.5), sample (paragraph 3.6), and analysis (paragraph
3.7) are justified.
3.1 Design
In order to measure variables, a mix of quantitative research approaches was chosen- a
closed-ended survey questionnaire and vignettes. Vignettes, also described as factorial surveys (Rossi &
Anderson, 1982), shortly illustrate a situation and contain references to the factors, influencing
specific decision making and can also be described as “stories which present hypothetical situations requiring action or judgment from respondents” (Wason & Cox, 1996, p.155). This was suitable for the research as it lowered the chance that ambiguous survey questions will be
misjudged because of the lack of explanatory information and allowed to manipulate few
variables simultaneously (Alexander & Becker, 1978). On the other side, the survey
questionnaire enabled to quantify the outcomes and draw conclusions regarding respondents’
answers. Thus, in order to simulate realistic situations and collect data, a combination of both
methods was used. Study design can be compared to the study of Stofberg, van Eerde and
17
3.2 Vignette design
3.2.1 The impact of COVID-19
During the time this research was being conducted, humanity was facing the biggest threat since
the second world war, known as COVID-19 (Corona Virus Disease 2019), which spread from
Asia worldwide. It restricted the movement of citizens to a minimum and affected psychological
factors- this virus had exposed changes in compulsive consumption (Baker, Farrokhnia, Meyer,
Pagel & Yannelis, 2020). Meanwhile, the fear of contagion has been studied during the earlier
epidemics and has shown to be the main causer of panic and change in consumer behavior
(Kinsman, 2012). It is unclear how the virus will affect sharing economy and whether it will
withstand, but many believe that even after the containment of COVID-19 distancing measures
will stay present, making people more resistant to engage in sharing. On the other side,
consumers may perceive car sharing as a safer option than using public transport or sharing a cab
as it involves less social interaction and the ability to disinfect cars. To take this situation into
account and its effects on positive/negative consumer intentions, this research includes the
presence of COVID-19 as an independent variable.
3.2.2 Design
In this study, vignettes were describing characteristics of a car-sharing platform and
manipulating platform intermediation together with a variable to control for COVID-19.
Independent variables can also be called features of the platform. Each feature can be classified
in levels- either if the feature is present (yes or no) or to what extent it is present (high or low),
18 peers and the platform. Some of the examples are presented in Figure 2 and explain how
variables were understood. COVID-19 situation is included as a control variable, thus every
situation is replicated in an environment where the virus is present and where it is not (Figure 3).
In each setting, variables were clearly described as a real-life car-sharing platform’s features
which intended to form respondents’ responses about dependent variables: willingness to
participate, trust in both peers and the platform (see Appendix A for an example of a vignette).
At the beginning of the study, participants were asked to indicate whether they owned a
car or not, directing them to either vignettes representative for a user or a car owner. Both types
had identical information regarding platform features and only were adjusted in wording to
separate views from both sides of the platform- more specifically, respondents who did not own
a car had to imagine a situation where they wanted to rent a car for the usage, meanwhile those
who were in a possession of a car were given situations where they had to imagine themselves
renting out their vehicle. Also, to account for the COVID-19 situation, respondents randomly
received a situation in which the virus was present and the movement was restricted, or where
the vaccine was existent and getting sick was not a concern anymore with absolute freedom to
19 Figure 2
Figure 3
3.2.3 Vignette population
The number of vignettes constructed was originally calculated by multiplying features and levels
as in 2 (presence vs. absence of COVID-19) x 2 (high vs. low platform intermediation) x 2
20
3.3 Survey questionnaire
The questionnaire was used to collect demographical information about participants and to gain
insights into dependent variables. Sections below describe questionnaire measures used for
willingness to participate from both user and owner sides, trusts in peers and platform as well as
control variables. An overview of the measurement constructs is included in Appendix B.
Willingness to participate: user
To evaluate how willing users would be to participate in the sharing activity, they were asked to
identify on a 7-point Likert scale to what extent they agreed with the provided statements. Only
participants who indicated that they do not own the car filled in their opinions. Examples
included “I would be likely to rent a car on SplitCar”, “I would put in an effort to rent a car on SplitCar” and “Splitcar would likely be my first choice to rent a car” (Lamberton & Rose, 2012; Pavlou & Gefen, 2004; White, MacDonnell & Ellard, 2012). The Cronbach’s alpha of α=.883
proves the internal consistency.
Willingness to participate: owner
In order to evaluate willingness to participate from the owner’s side, participants were asked to identify their willingness on a Likert scale from 1 to 7 (1 is completely disagree; 7 is completely
agree). Only participants who indicated that they own the car had to answer this question. Some
examples were “I would be likely to share my car on SplitCar” and “I would be willing to share my car on Splitcar” (Lamberton & Rose, 2012; Pavlou & Gefen, 2004; White et al., 2012). The scale was the same as for users, thus reliability was sufficient (α=.883).
21 Peer trust
In order to check how independent variables affect trust in peers, questions were formulated in a
way that participants had to indicate to what extent they agreed with the statements. Answers
were recorded on a Likert scale where 1 was completely disagree and 7 was completely agree.
One measure used by Pavlou and Gefen (2004) was adopted and presented as follows: “In
general, I expect the users on SplitCar to be dependable”. They were also asked to evaluate statements such as “In general, I expect the users on SplitCar to be trustworthy”. Internal validity was confirmed with a Cronbach’s alpha of α=.918.
Platform trust
Trust in the platform was also measured by asking participants to evaluate to what extent they
agreed with the presented statements. Answers were indicated on a 7-point Likert scale ranging
from ‘completely disagree’ to ‘completely agree’. For example, “I expect that SplitCar can be trusted at all times” (Pavlou & Gefen, 2004), and “I expect SplitCar to be a competent platform provider” were statements to answer. The Cronbach’s alpha resulted in α=.887 and could not be improved by dropping items.
Control variables
The results are controlled for age, gender, location, income, education level, possession of a
driver’s license, and possession of a car. Age is known to be inversely correlated with
22 Smith (2016) and Eurobarometer (2016), highly educated people are more likely to engage in
sharing activity both as a user and as a provider, therefore we incorporate education measure in
our study. Also, the COVID-19 situation is included as consumers perceive health risks
differently, depending on individual characteristics. To take this into account, we asked
respondents to indicate to what extent they believed that renting a car from SplitCar could be
risky for their health or increase their odds of being exposed to illness. Views on the safety of
other transportation options can vary as well, thus respondents had to indicate the safety benefits
they perceived from traveling by car on Splitcar as compared to traveling with public
transportation (Goodwin, Burke, Wildman & Salas, 2009). Both statements were evaluated on a
5-point scale.
Credibility
Manipulation checks were included to ensure the credibility of vignettes. They were evaluated on
a 7-point Likert scale (1=completely disagree,7=completely agree) and adapted from the research
of Stofberg and Bridoux (2019). For credibility following statement was presented “I had no
problem imagining myself in the above-mentioned situation” and realism was checked by the
sentence “I found the situation in the above-mentioned scenario realistic”.
3.4 Pre-test
To ensure that manipulations of variables in vignettes were interpreted as intended and to check
credibility, a pre-test was executed before the data collection. After the construction of vignettes,
23 feedback. Questions about each independent variable were asked as well as general ones such as “Do you understand everything written? Are there any difficult words?” and “Does the platform offer assistance beyond the booking (customer support, insurance)?”. After the pre-test, grammar mistakes were fixed as well as more difficult English words were replaced with more
understandable ones in both vignettes and scales. Also, vignettes for the owners were adjusted in
wording to make them more coherent and straightforward.
3.5 Procedure
Data collection was a part of a larger study and done by 11 Business Administration students.
Only the data relevant to this study were analyzed. Data were collected worldwide, thus included
different nationalities. Taking this into account, information was presented in basic English so
non-native language speakers could easily understand and interpret the situations and questions.
After vignettes and questionnaires were developed and adjustments after the pre-test were done,
researchers used social media for sharing the questionnaire as well as convenience sampling by
distributing it to family, friends, and acquaintances. In addition to that, the snowball effect has
been evident as all sharing happened online.
3.6 Sample
485 valid responses were left for analysis after cleaning up the data. Over 41% of them owned a
car and over 88% had a driver's license. The average age of respondents was 32 years (M=32.05,
SD=12.956), ranging from 18 to 87. Almost 56% were women, 43% were men, and less than 1%
24 had a professional one and less than 8% were low educated. In comparison, income levels are
significantly lower. The majority of respondents (41%) had a monthly income of €1,500 or lower
and around 23% had a medium income- this might have resulted from the fact that responses
were collected worldwide where income levels extremely vary- regarding location, 80%
indicated their region as Europe, 6% as America and over 10% as Asia. Overall, more than half
of the respondents were from The Netherlands (51.8%), and regarding regions, 84% were
residing in an urban area compared to 16% in a rural one.
3.7 Analysis
After the data collection, missing values were cleared out of the dataset by using person total
correlation, Mahalanobis distance tests, and removing respondents who took an unrealistic
amount of time to complete the survey. Afterwards, results were registered in SPSS 25 for
statistical analysis. In the end, 485 responses were analyzed, and categorical variables were
modified into Dummy ones. Next, final variables were constructed from the items measuring
trust in both peers and the platform, and willingness to participate as the Cronbach’s alphas were >.7. For the analysis part, SPSS extension PROCESS was used for both mediation and
moderation hypotheses to test regression, correlations, and evaluate mediation and moderation
effects. To check moderation hypotheses, centering was done first to ease the interpretability of
the results and reduce multicollinearity. The next section discusses the results of the analyses in
25
4. Results
In the results section data analyses are reviewed starting with the distribution of variables and
manipulation checks (paragraph 4.1), followed by correlation matrix, control variables, and
hypotheses testing (paragraph 4.2).
4.1 Distribution
Both trust in peers and trust in the platform as well as willingness to participate are tested for
skewness and kurtosis (see Appendix C for details). None of the variables have big skewness or
kurtosis issues, thus further normality checks are executed to ensure that. Kolmogorov-Smirnov
and Shapiro Wilk tests are significant for all three variables (see Appendix D for details), but
normality issues are not causing difficulties in big datasets (>40) (Pallant, 2007), thus no
additional checks are done.
4.1.1 Manipulation checks
One-way ANOVA test is run to test credibility and reliability of used vignettes. This results in
significant heterogeneity of variances for realism (F(7,477)=2.852, p=.006) as well as for
credibility (F(7,477)=5.397, p=.015). Welch tests are also significant for both variables
(B=2.890, p=.007 for realism; B=2.627, p=.013 for credibility), but as all vignettes’ means are over 4.5, there is no reason to do any corrections. To assure that, post hoc test is conducted- more
specifically, the Games-Howell test reveals that some vignettes are more credible than others and
26 The reason why some vignettes are more credible and realistic might result from having
to imagine situations where no COVID-19 is present as it is too different from everyday life.
Also, some countries might have less developed sharing economy and car-sharing services, thus
participants have troubles imagining unfamiliar situations (see Appendix E for correlations
between countries).
4.1.2 Correlations
Table 1 presents a correlation matrix- trust in peers, as well as trust in the platform, strongly
correlate to WTP and these findings are complementary to the findings discussed in the
theoretical framework (r=.361, p<.001, and r=.345, p<.001 respectively). Also, WTP is
significantly linked to the high intermediation of platforms (r=.103, p=.024). While high
intermediation is correlated with trust in the platform (r=.162, p<.001), it is not correlated to
trust in peers (r=.046, p=.314) and because both types of trust are significantly linked together
(r=.632, p<.001) it can be interpreted as a possible sequential mediation.
Car owners are negatively correlated with WTP (r=-.309, p<.001), meaning they are
more reluctant to engage in sharing activity and that validates our discussed theory on owners
and users.
4.1.3 Control variables
Control variables as well as demographics have some significant correlations with the main
variables in this research. The ownership of the car is positively correlated to age (r=.285,
27 (r=.0285, <.001) and low income (r=-.193, p<.001) emphasize demographic picture of the car
owners- compared to users, they are older and have a higher income. Regarding gender, females
are correlated to peer trust (r=.116, p=.011) which shows they tend to trust more often than men.
On the other side, age is negatively correlated to WTP (r=-.239, p<.001) and it implies that as
respondents get older, they are more resistant to participate in sharing activity. WTP is also
correlated to low-income respondents (r=.156, p=.001) and it implies that a car-sharing activity
is a possible way to save money, which also has been proven in theory.
Some control variables significantly correlate with each other. Low income correlates
with academic and professional education (r=.172, p<.001, and r=-.149, p=.001 respectively),
Indonesian (r=-.170, p=.000), Lithuanian (r=.120, p=.014), German (r=.093, p=.040) and Italian
28 Table 1. Correlations matrix
4.2 Hypotheses testing
H1 and H2 test platform and peer trust respectively on WTP, H3 is concerned with
intermediation and its effect on WTP while platform trust is mediating the relationship, and H4
predicts the sequential mediation which eventually proves that high intermediation affects WTP.
Further on, H5 and H6 hypothesize that peer and platform trust lead to more WTP when a
29 Figure 4
Note. *p<.05, **p<.01
4.2.1 Hypothesis 1
It is hypothesized that trust in the platform leads to more WTP in the sharing economy. This is
30 the VIF values were below .1 and none of the Tolerance values were above 10. Furthermore, the
assumption of no autocorrelation of residuals has also been met as Durbin-Watson resulted in
2.100, and assumptions of linearity and homoscedasticity have been met as well as the scatterplot
resulted in a proper bell-shaped figure. Table 2 presents the outcomes and the model turns out to
be significant (F(18, 466)=9.973, p<.001), and explaining almost 28% of the variance (R²=.278).
Regression reveals a significant direct effect (B=.497, t=8.706, p<.001) meaning with every
one-point increase in platform trust, WTP increases by .497, thus the hypothesis is supported.
Two of the control variables- age and being an owner- also predict WTP. With age
people seem to be .013 units less likely to participate (B=-.013) where t=-2.344, p=.020, and
with a 95% confidence interval from -.023 to -.002. Owners are also .808 units less likely to
participate (B=-.808) where t=-6.367, p<.001, and with a 95% confidence interval from -1.057 to
-.558. Additionally, Dutch nationality affects it as well (B=-.350) where t=-2.176, p=.030, and
with a 95% confidence interval from -.665 to -.034.
31
4.2.2. Hypothesis 2
It is hypothesized that trust in peers leads to more WTP in the sharing economy and linear
regression is used to check that. Again, all the assumptions have been met as VIF values were
below .1, none of the Tolerance values were above 10, scatterplot has shown a bell-shaped curve
and Durbin-Watson resulted in 2.016. Table 3 presents the outcomes- the model is significant
(F(18,466)=10.810, p<.001) and explains almost 30% of the variance (R²=.295). It proves that
there is a direct effect (B=.468, t=9.404, p<.001) and that with every point increase in peer trust,
WTP increases by .468, hence, the hypothesis is supported.
Similarly, as for hypothesis 1, age and car ownership also predict WTP. Owners are .749
units less likely to participate than users (B=-.749, t=-5.965, p<.001) and as participants age each
year, WTP decreases by .016 (B=-.016, t=-2.927, p=.004). Additionally, platform intermediation
also affected WTP meaning with every point increase in intermediation, WTP is increasing by
.237 (B=.237, t=2.136, p=.033).
32
4.2.3 Hypothesis 3
It is assumed that trust in the platform mediates the relationship between platform intermediation
and WTP. This is tested by running a regression analysis and Table 4 demonstrates that there is
no direct effect- the degree of platform intermediation does not impact WTP (B=.1115, t=.9801,
p=.3275). Nevertheless, platform intermediation determines trust in the platform (B=.3453,
t=3.8026, p=.002) which, in turn, as discovered in hypothesis 2, determines WTP (B=.4970,
t=8.7057, p<.001). As the confidence interval of indirect effect is totally above zero (Eff=.1716,
SE=.0483, CI: .0828 to .2744), full mediation supports hypothesis 3.
33
4.2.4 Hypothesis 4
Building on the previous hypotheses, it is assumed that the relationship between platform
intermediation and WTP is sequentially mediated by trust in platform and peers. As discussed
previously and presented in Table 5, it is proved again that degree of platform intermediation
does not directly affect WTP (Eff=.1570, t=1.4130, p=.1583) and that trust in the platform is
affected by platform intermediation (B=.3453, t=3.8026, p=.002). Furthermore, it is evident that
consecutively trust in the platform determines trust in peers (B=.7002, t=16.9559, p<.001), and
it, in turn, affects WTP (B=.3191, t=5.1160, p<.001). These results present full sequential
mediation (Eff=.0772, SE=.0258, CI: .0330 to .1364) and confirm hypothesis 4.
34
4.2.5 Hypothesis 5
The fifth hypothesis projects that the relationship between peer trust and WTP is moderated by
the role of the consumer and that the relationship should be strengthened when a consumer is a
provider than a user. After running a moderation analysis, in Table 6 it is revealed that peer trust
affects WTP just as anticipated previously (B=.5080, t=7.3169, p<.001) but the interaction effect
is not significant, meaning that consumer’s side on the platform does not play a role in
moderating the relationship (B=-.0801, t=-.8230, p=.4749). Hence, hypothesis 5 is rejected.
Table 6. Effect size and significance levels of hypothesis 5
4.2.6 Hypothesis 6
The sixth hypothesis states that the relationship between platform trust and WTP is moderated by
the role of the consumer and that the relationship should be strengthened when a consumer is a
provider than a user. Again, the previous hypotheses have proved the effect of platform trust on
35 not significant (B=-.1499, t=-1.3506, p=.1775), meaning that it does not matter whether the
consumer is a user or a provider and hence the hypothesis 6 is not supported and rejected.
36
5. Discussion
This section discusses the results in section 4. Firstly, the summary of results is presented
(paragraph 5.1), followed by its discussion (paragraph 5.2). Lastly, theoretical (paragraph 5.3)
and managerial implications (paragraph 5.4) are reviewed together with the limitations of the
study and recommendations for the future research (paragraph 5.5).
5.1 Summary of the results
As expected, both trust in peers and the platform directly influence willingness to participate.
Additionally, owners are less willing to participate in the sharing activity than users and
low-income participants are more likely to use car-sharing than high-low-income levels.
High intermediation is also positively affecting willingness to participate, although this
influence is indirect and coming from two different paths. Firstly, high platform intermediation
leads to higher platform trust which, in turn, drives the willingness to participate and it represents
full mediation. In contrast, no relationship is found between high intermediation and trust in
peers. Secondly, a sequential mediation presents another indirect effect as even though high
intermediation does not relate to trust in peers, trust in the platform has a significant impact on
trust in peers. This results in a fully mediated relationship, as high intermediation increases trust
in the platform which promotes trust in peers and this type of trust leads to the willingness to
participate.
Contrary to our expectations, being an owner instead of a user did not strengthen the
37
5.2 Discussion of the results
The discussion commences with the direct link between peer and platform trusts and willingness
to participate (paragraph 5.2.1), followed by the role of platform intermediation and trust
(paragraph 5.2.2), and ends with the two-sided view of the platform on the willingness to
participate (5.2.3).
5.2.1 Direct link between both types of trust and willingness to participate
This research hypothesized that peer and platform trusts predict willingness to participate. They
were formulated in order to prove the previous findings and to build a base for our further
hypotheses. When people trust the intermediary and its ability to perform, they are more willing
to engage in car sharing and our research has proved that. These findings support the research
from Mittendorf et al. (2019) who also argued that the probability of a participant engaging in
sharing is reinforced as trust in the platform intensifies.
Furthermore, when people trust other parties involved in sharing, they are also more
likely to participate in sharing platforms. Our research verifies that willingness to participate is
influenced by peer trust. This replicates the findings of quite a few other researches of
Hawlitschek et al. (2016), Mittendorf (2018), and Chen et al. (2009), which proves that trust in
peers is one of the ways to increase sharing.
In addition to that, few other control variables appear to directly affect willingness to
participate which is worth mentioning. Our research revealed that older people (over 32 years
old) are less likely to engage than young participants- this is most likely due to the fact that
38 as having as many disadvantages as advantages (Hawlitschek, 2016). Findings also validate the
choice of Mittendorf et al. (2019) to choose millennials as their sample for researching sharing
economy- they grew up with constant access to the Internet and they tend to care about
environmental issues, financial saving, and sharing (Lenhart, Purcell, Smith & Zickuhr, 2010;
PwC, 2015; Bardhi & Eckhardt, 2012). Also, it has been revealed that owners tend to be more
reluctant to participate and significantly older than users. This builds upon the findings of
Mittendorf et al. (2019) which demonstrate that the effect of trust in the intermediary on the
users’ willingness to participate is stronger for customers than for providers. Additionally, it is known that older people are more susceptible to age stigmas (Zebrowitz & Montepare, 2000)- a
fear of renting your valuable goods to less experienced, laid-back, and perceived as irresponsible
youth is most likely playing a role.
5.2.2 The role of platform intermediation and trust
Further hypotheses were formulated to explore the role of platform intermediation and its effect
on willingness to participate. Surprisingly, as results have shown, there is no direct effect
between those two concepts but there are a few indirect paths.
High platform intermediation increases the likelihood of participation in sharing through
trust in the platform- this full mediation indicates that high platform intermediation is one of the
factors influencing intermediary trust. This supports findings from the exploration of Lampinen
and Cheshire (2016) who debate that high platform mediation increases trust in transactions for
consumers. The reason why no direct effect was found might be the fact that mediation only
39 participate. Furthermore, high platform intermediation does not impact trust in peers which is
plausible as peer trust is more social construct than a structural one and is not related to
mediation.
Another indirect path that was discovered included a sequential effect of platform to peer
trust, demonstrating the importance of a trust transfer in sharing context. This specific trust
transfer- from platform to peers- has been discussed by Jones and Leonard (2008) as well as Vos (2018) and they all referred to P2P sharing activities. Our research adds to Mohlmann’s (2016) discovery of a two-fold hierarchical structure of the trust concept, continuing the exploration of
it, and adding platform intermediation into consideration. This implies that in order to promote
the sharing economy and boost active presence, both types of trust have to be established,
starting with the platform one. It requires an active exploration of precursors that constitute trust
in the platform, and we have discovered that high platform intermediation is one of them.
5.2.3 Two-sided view of the platform on the willingness to participate
Lastly, this research predicted that both types of trust are more important for owners than for
users as they face higher risks. Contrary to our expectations, there is no significant difference in
the role of consumer affecting the relationship between trust in platform or peers and willingness
to participate. This means that trust overall is of considerable importance for both users and
owners, thus both sides have to be taken into account equally. Nevertheless, owners were found
to be less willing to participate in sharing than users and it emphasizes notable differences
between the two. The rationale for this might be that owners are considerably older than users
40 5.2.1). This confirms Eurobarometer’s (2016) discovery that there are almost four times more
users than owners in the sharing economy, meaning they are harder to acquire and maintain. This
suggests that apart from the common factors such as trust, there are different factors for both
sides of the platform that encourage participation. As both sides have distinct intentions and risks
(Eurobarometer, 2016), it is important to treat them separately and further explore what attracts
them to share or use sharing services.
5.3 Theoretical implications
This study adds to the existing literature on sharing economy, more specifically- on P2P
platforms. It extends the understanding of overall trust that is being discussed in e-commerce and
sharing economy contexts (Gefen, 2000; Pavlou & Gefen, 2004; Hawlitschek et al., 2016) into
trust in peers and trust in the platform which directly influence willingness to participate. As the
research on those types of trust separately is quite scarce, it is unknown what antecedents create
them. Fortunately, we have been exploring platform intermediation and its influence.
Many studies have been exploring platform mediation and structural assurances
separately (Teo & Liu, 2007; Sutherland & Jarrahi, 2017), but both of them usually prevail
simultaneously, thus were combined into a definition of platform intermediation which “surfaces
from observations regarding the varying extent to which the software platform and its attendant
algorithms and tools manage and coordinate exchanges” (Perren & Kozinets, 2018, p.24). No
research up to date has analyzed platform intermediation and its impact on sharing platforms and
both types of trust. Our study adds to the current literature by revealing that high platform
41 presence of trust transfer (Mohlmann, 2016; Vos, 2018), it broadens the understanding of what
constitutes willingness to participate in P2P sharing platforms.
Lastly, this study emphasizes the importance of two sides of the platform, namely owners
and users. Even though it is known what the intentions are for both sides and that they differ
considerably, it is still unidentified what attracts them the most to engage in sharing activities.
This research adds to the knowledge of trust in both peers and the platform, and accentuates that
they both matter to owners and users uniformly. Car owners are more reluctant to participate, but
trust is not a decisive factor determining the reason for it. It is known that the main incentive for
owner participation is additional income (Dabbish, Lee, Kusbit & Metsky, 2015; Ikkala &
Lampinen, 2015; Lampinen & Cheshire, 2016; Teodoro, Ozturk, Naaman, Mason & Lindqvist,
2014), thus one of the causes of hesitation might be related to material gains rather than
intangible benefits.
5.4 Managerial implications
This research highlights the importance of both trust in peers and the platform as well as
differences between both sides of the platform. Since both types of trust are interrelated and
platform trust leads to peer trust, managers have to find a way to increase both of them, but most
importantly- platform one, as it will consequently increase trust in peers. Luckily, it was
discovered that high platform intermediation increases platform trust, thus it is important to
integrate structural assurances into processes and strengthen the control that the platform has. In
42 possibly include refunds or insurances, and make sure they are protecting both users and owners
fairly.
Additionally, as it is clear that both sides of the platform vary, it is crucial for
management to set up different incentive plans, support, or reinforcement arrangements
separately for owners and users. As the failure to recruit enough owners is one of the reasons
why sharing businesses fail (Chasin et al., 2018) and owners are overall less willing to
participate, it might be useful to give more attention to the supply side of the platform and tailor
recruitment campaigns accordingly. This could be simply done by emphasizing gains for car
owners, setting up encouraging pricing systems, or advertising positive experiences of existent
providers on the platform.
5.5 Limitations and recommendations for future research
Despite the careful conduction of the study, this study has some limitations to be discussed
together with recommendations for similar future researches.
First of all, even though it was controlled for COVID-19, the worldwide pandemic might
have had some unforeseen influence on the study. Changes in consumption and behavior during
the spread of the virus are unavoidable (Baker et al., 2020; Kinsman, 2012), thus it is impossible
to fully replicate the situation in the future and it might have had different results if done in the
context of no virus.
Secondly, sample bias might have been evident because the presence of COVID-19 has
restricted the ability to randomly approach people outside. As a consequence, respondents were
43 acquaintances. Our sample also included different nationalities, though the Netherlands
accounted for 51% of our sample and could have skewed the results, as Dutch citizens were
significantly less willing to participate in sharing than other nationalities. Additionally, the
questionnaire was considerably long to complete and as there were no incentives, plenty of
questionnaires were started and left off. In the future, we would suggest recruiting not only
familiar respondents but also include more varied nationalities, in equal proportions and outside
the inner circle. For a long survey, it is advisable to include some encouragement that is possible
with financial resources.
Lastly, as the owner side of the platform did not influence the relationship between both
types of trust and willingness to participate, future research could address unexplored both sides
44
6. Conclusion
First of all, this study analyzed the importance of platform intermediation on platform trust as
well as willingness to participate in P2P sharing context. Additionally, the mediating influence of
platform and peer trusts has been explored. Lastly, this study examined whether as claimed by
Ter Huurne et al. (2017) trust cues were more important to owners than they were to users.
Trust is often commended as the key factor that reduces uncertainty and complexity in
sharing contexts, thus it is acknowledged to attract potential users and providers to start
participating on sharing platforms (Gefen, 2000; Rousseau et al., 1998). The confirmation to
these predictions was found, that both trust in the platform and peers directly increase the
likelihood of participation. After the importance of trust was explicit, the analysis focused on the
main antecedent of trust, namely the degree of platform intermediation. This research
demonstrates that high platform intermediation indirectly increases willingness to participate in
sharing activity, and one way to do that is by increasing trust in the platform, which acts as a
mediator. Another way is through sequential mediation- high platform intermediation affects
platform trust which increases peer trust and it leads to higher levels of intended participation.
Regarding two sides of the platform, for both sides platform trust and peer trust matter
equally, even though owners are facing bigger financial risks and compromise their valuable
assets. Nevertheless, it is evident that differences between owners and users are present and have
to be taken into account- owners are substantially less likely to use car-sharing platforms than
users.
This research contributes to the literature by expanding the understanding of both types
