Gaining Traction in the Carsharing Economy
The role of platform characteristics and personal motivations on participation intent
in the carsharing economy
Lucca van Holten Charria – 11749202 Bachelor Thesis
BSc Business Administration - Management in the Digital Age University of Amsterdam
Nicole Stofberg 10 July 2020
1
Statement of originality
This document is written by Lucca van Holten Charria 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.
2
Table of content
1. Introduction 6
2. Literature review 7
2.1 Defining the sharing economy 7
2.2 The sharing economy has not lived up to its potential 8
2.3 Platform mediation and platform sociality 9
2.4 How platforms realize convenience and sociality 10
2.5 How sociality and convenience translate to willingness to participate H1a, H1b, H2 11 2.6 The moderating influence of social and functional motivations H3, H4 12
3. Methodology 16
3.1 Research design 16
3.2 Vignette design 16
3.3 Survey instruments 18
3.4 Data collection and procedure 20
3.5 Analytical plan 20
4. Results 20
4.1 Realism and credibility 21
4.2 Distribution 21
4.3 Descriptive statistics and correlations 21
4.4 Regression analyses 23
4.4.1 Hypothesis 1a 23
4.4.2 Hypothesis 1b 23
4.4.3 Hypothesis 2 23
4.5 Moderation analyses 25
4.5.1 Hypothesis 3.1a and 4.1a 25
4.5.2 Hypothesis 3.1b and 4.1b 26
4.5.3 Hypothesis 3.2 and 4.2 27
4.6 Additional analysis 30
5. Discussion 34
5.1 Summary of the results 34
5.2 Discussion of the results 34
5.2.1 Platform characteristics 34
5.2.2 Personal motivations 35
5.2.3 Additional analysis 36
5.3 Theoretical implications 37
5.4 Managerial implications 38
5.5 Limitations and suggestions 39
6. Conclusion 40
References 42
3
Appendices
Appendix A. Vignette designs used in study 47
Appendix B. Pre-test questionnaire 48
Appendix C. Example of vignettes 49
Appendix D. Survey transcript 51
Appendix E. Overview of measures and Cronbach’s alpha 57
Appendix F. Dummy variables overview 58
Appendix G. Realism and credibility tests 59
Appendix H. Distribution 60
Appendix I. Regression analyses 61
Appendix J. Moderation tables for insignificant effects 62
List of tables
1. Overview of hypotheses 15
2. Overview of dimensions and their levels 16
3. Operationalization of the dimensions and levels as described in the vignettes 17
4. Descriptive statistics of demographics 21
5. Descriptive statistics and correlations 22
6. Regression table with control variables and LH>HL on willingness to participate 24 7. Overview of regression analyses for all vignettes on willingness to participate 25 8. Linear predictors of willingness to participate (H4.1b) 26 9. Linear model of predictors of willingness to participate (H4.2) 28 10. Effect sizes of regression and moderation analyses of vignettes 29
11. Outcome of the hypotheses 30
12. Regression table with control variables and business model on willingness to participate 31 13. Linear model of predictors for willingness to participate (Ad.) 32 14. Linear model of predictors for willingness to participate (Ad.) 32
List of figures
1. Lateral Exchange Market Types 11
2. Conceptual model 15
3. Effect of social motivations on HH, LH and willingness to participate 27 4. Effect of social motivations on LH, HL and willingness to participate 28 5. Significant effects confirmed in HH>LH and LH>HL vignettes 29 6. Effect of motivations on business model and willingness to participate 33
4
Acknowledgements
First and foremost, I would like to express my gratitude to my supervisor Nicole Stofberg. Her enthusiasm, energy and knowledge have not only made the process of writing this thesis an educational journey, but a very enjoyable experience as well. Her guidance and valuable advice have truly helped me along the way.
I would also like to thank my fellow students. Together we formed an energetic group that completed the data collection successfully and supported each other when needed.
5
Abstract
Collaborative consumption, especially carsharing, allows individuals to utilize idle resources in a sustainable manner. Despite the environmental, social, and economic benefits carsharing provides, platforms struggle to find the most suitable platform architecture to appeal to more users and increase participation. Focusing on platform mediation and functional dimensions has been applauded by scholars, yet the sociality dimension remains relatively unaccounted for. This study distinguishes between three types of carsharing platforms (Matchmakers, Hubs, and Forums), which vary in their degrees of platform mediation and sociality. The aim of this study is to investigate how configurations of these dimensions influence willingness to participate directly, and the moderating (self-selecting) role social and functional motivations of prospective users play herein. Results of an online vignette study (N = 1112) indicate that platform mediation is the principal value driver when no motivations are taken into account. However, in the presence of social motivations, platforms focusing on sociality features score higher (when combined with mediation characteristics) or experience a more significant increase (when isolated) on willingness to participate than platforms excluding sociality. No effect of functional motivations is concluded. Additional analyses illustrate that business-to-consumer models are preferred over peer-to-peer models, implying access-based consumption is favored as opposed to true collaborative consumption. Social and functional motivations do, however, show a more positive effect on peer-to-peer models. Although Matchmaker constructs do not result in higher participation intent than Hub constructs, the degree of sociality employed in platform architectures should be carefully considered. Carsharing platforms may indirectly benefit from incorporating sociality features, as users driven by social motivations are ultimately more altruistic. Alternatively, managers can opt for a B2C business model – increasing participation but not qualifying as collaborative consumption.
Keywords: sharing economy, collaborative consumption, access-based economy, carsharing, platform mediation, platform sociality, functional motivations, social motivations, platform architecture
6
1. Introduction
In an era of resource efficiency and sustainable consumption, sharing networks are gaining traction on a large scale. Online platforms that facilitate sharing tangible products are becoming more prevalent, as individuals increasingly value the efficient use of resources. Besides price advantages and environmental benefits, sharing contributes to the surge of sustainable ownership alternatives, aiming to reduce overconsumption through the clever use of technological advances (Wilhelms et al., 2017). Also known as collaborative consumption, the sharing economy has redefined traditional business models by providing structures for sharing, lending, renting, or trading underutilized assets amongst peers (Botsman & Rogers, 2010). In particular, carsharing allows consumers to benefit from others' resources without bearing the costs of acquisition and ownership (Philip et al., 2015). As carsharing is becoming increasingly conventional, platforms attempt to offer large varieties of characteristics in the pursuit of competitive advantages.
The environmental benefits effectuated by carsharing are substantial. By providing an eco-efficient alternative to car ownership, platforms like ZippCar, Car2Go, and SnappCar ultimately reduce CO2 emissions of their users (Metz, 2013). One carsharing car replaces nine to thirteen personally owned cars (De Luca & Di Pace, 2015), and car ownership has reduced by 30%. In addition, about one-fifth fewer kilometers are driven, and 13% to 18% less CO2 related to car ownership and use are emitted compared to non-carsharing users (Nijland & van Meerkerk, 2017). Considering that one-third of greenhouse gas emissions are generated by passenger cars (European Environment Agency, 2019), the wide scaled adoption of carsharing results in substantial environmental and societal gains.
Despite these advantages, carsharing has not lived up to its potential. Companies have yet to develop the most suitable architectures to cater to their audiences (Demil & Lecocq, 2010). In order to increase participation rates, the functional and economic benefits of sharing have consistently been emphasized. Numerous scholars advocate for this approach (Hamari et al., 2016; Bardhi & Eckhardt, 2012), disregarding the interpersonal and communal benefits sharing platforms could leverage. Although sociality is often not seen as the primary dimension providing value to the platform (Collier & Sherrel, 2010; Zhang et al., 2016), the interpersonal aspect of sharing may decrease negative externalities resulting from information asymmetry and opportunistic behavior (Ert et al., 2016). Focusing solely on convenience and functionality might therefore limit the potential of carsharing platforms, though this can be counteracted by employing sociality aspects in the platform architecture.
The role of personal motivations on platform selection is highlighted by Hellwig et al. (2015), stating that participants do not uniformly search for conventional benefits associated with sharing, such as pro-sociality and altruism. Instead, participants can be clustered into specific segments differing in the value sought from sharing transactions. Whilst “sharing idealists” fit the classical description of being driven by
7 intrinsic motivations, “such as being part of a community or prosocial ideals linked to helping each other" (Hellwig et al., 2015, p.904), “sharing pragmatists” are driven by practical reasoning. They search for convenience and functionality in sharing transactions. Participants guided by social motivations are more attracted to sharing platforms emulating sociality characteristics, being deterred from sharing driven by convenience and utilitarianism. Conversely, those participants guided by functional motivations seek utilitarian value and are more likely to self-select into models emphasizing convenience. This differentiation provides a critical insight in the platform selection by participants, based on the interests they seek from sharing.
This study aims to explore the relationship between platform characteristics and intention to participate by replicating carsharing platforms differing in levels of sociality and mediation. Additionally, social and functional motivations are examined to investigate the drivers that lead individuals to partake in different platform types. Previous literature has examined the drivers (Botsman & Rogers, 2010) and the characteristics (Perren & Kozinets, 2017) of carsharing disconnectedly, providing stand-alone insights that may result in uninformed undertakings by platforms. By combining these factors, a more coherent outline of effective platform architectures, paired with specific recommendations, can be constructed. The overarching research question guiding this study is the following:
"To what extent do sociality and mediation platform characteristics influence willingness to participate, and how is this relationship affected by personal motivations?"
To answer this question, this study will begin by providing a theoretical framework through a literature review, leading to the conceptual model and hypotheses. Next, the methodology is outlined, and results are reported. These results are debated in the discussion section, providing theoretical and managerial implications, limitations, and recommendations for future research. Lastly, all findings are summarized in the conclusion.
2. Literature review
2.1 Defining the sharing economy
Being a relatively new concept, the ambiguity of the sharing economy definition allows for many misconceptions and various interpretations. To ensure coherence throughout this study, the following insights will be used to provide a clear definition. Frenken (2015) defines the sharing economy as "consumers granting each other temporary access to under-utilized physical assets, possibly for money" (pp. 4-5). From this definition, four key components can be derived, distinguishing the sharing economy
8 from any other market type. Temporality implies there is no transfer of ownership over the shared good. The goods transferred must be physical, and the sharing must occur between consumers, implying a peer-to-peer relationship. Lastly, the physical good must be under-utilized.
In essence, sharing economies enable exchanges. Hamari et al. (2016) note the importance of the mode of exchange. Donating, purchasing, and swapping fall under transfer of ownership, whilst renting and lending are a mode of access. Access over ownership implies the handover of goods without permanent transfer of ownership. In addition to non-ownership, Belk (2014) emphasizes the importance of two-directional interactions through an internet platform. These interactions allow consumers to engage with each other and with the platform.
Within these exchanges, there is an essential differentiation between peer-to-peer (P2P) and business-to-consumer (B2C) sharing. B2C sharing platforms own the under-utilized assets, allowing for a more traditional business model, whilst in P2P settings, the platforms more often earn money by commission (Schor, 2016). Shaheen et al. (2018) elaborate on this, stating that in P2P sharing, the platform facilitates the exchange whilst the consumer provides the asset, whereas in B2C sharing the platform owns and provides the exchange. This is an access-based rather than a collaborative consumption model (Belk, 2014), implying that in P2P sharing platforms, users are in contact with each other. In B2C models, the platform merely facilitates this exchange. Frenken and Schor (2017) describe B2C sharing as pseudo-sharing, as it involves a platform simply lending to customers.
2.2 The sharing economy has not lived up to its potential
Sharing economy platforms rely on their transactional networks' robustness to facilitate successful interactions (Schor et al., 2016). Economic reasoning seems to overshadow altruistic motives to share in such networks. Functionality is considered more important than sociality in providing value to the consumer (Richardson, 2015). Belk (2010) affirms that carsharing is often not driven by altruistic motivations. Although a sense of community between users should be facilitated in sharing contexts where collaboration is desired, in many carsharing contexts the platform is merely a service provider, and social connections are not prized. Therefore, carsharing platforms have increasingly altered their platform architectures to be as convenient and least social as possible, based on the reasoning that users are only after utilitarian value, not social value.
However, most research supporting the pursual of convenience has been confined to B2C models, where the service provider owns the shared asset, and interpersonal connections are irrelevant. Social interactions are more critical when sharing between peers. Stofberg and Bridoux (2019) note how removing inter-personal interactions for the sake of convenience has worked reversely for carsharing platform Turo: a key-less, impersonal approach was not appreciated by its users, hence the platform reverted to employing
9 sociality. Additional research provides evidence that sociality is a key driver in the effectiveness in sharing platforms. For instance, Van der Glind (2013) outlines how intrinsic motives for sociality and community-building are important drivers for willingness to participate in P2P technologies. Likewise, Fitzmaurice et al. (2016) explain how platforms have enabled individuals to build social networks and connect to other individuals. They identify a desire for markets organized around social connections. Perren and Kozinets (2018) further describe how sharing platforms can enable social exchanges and provide high platform mediation simultaneously. The value ascribed to sharing platforms does therefore not solely rely on the strengths of the intermediation offered by the platform, but also on the obtainable sociality between actors.
This poses the question of whether the steps managers are taking to reduce platform sociality for the sake of utilitarianism and convenience are the right ones. To investigate this question, this paper will outline how platforms can realize convenience and sociality in varying degrees, depending on platform design choices. The relationship between platform mediation and platform sociality on willingness to participate will be examined, as well as the self-selecting effect of personal motivations. By acquiring these insights, carsharing companies can adjust their platform characteristics to cater to their target audience better. In addition, by crossing the dimensions, this study generates novel insights on how value is created on platforms with divergent characteristics.
2.3 Platform mediation and platform sociality
The degree to which carsharing platforms differ in support and involvement with the user can be defined as 'platform mediation'. On one side of the spectrum, these software-based platforms function as algorithmic matchers, providing a space for users to exchange. Alternatively, sharing platforms with high mediation are active participants in the search, exchange, and completion stages of exchanges (Perren & Kozinets, 2018). Examples of mediation include user verification, customer support, and insurance. De Rivera et al. (2017) note that low mediation platforms provide a basic and passive technological structure, whilst high mediation platforms facilitate and mediate exchanges more actively. By increasing the efficiency and extensiveness of platform features, effective mediation can be ensured. The higher contribution of online platforms diminishes the search and decision costs for users (Anderson & Srinivasan, 2003), translating directly to higher perceived convenience (Berry et al., 2002). As each interaction is unique, the management and coordination of exchanges are essential in providing value.
Platform sociality is defined by Perren and Kozinets (2018) as "the physical and/or virtual copresence of social actors in a network, which provides an opportunity for social interaction between them" (p. 23), fostering communication, coordination, and negotiation on the platform. Wittel (2001) refers to sociality as tendencies to build associative and cooperative relationships between users. Platforms with high sociality allow for free-flowing interactions between their users, resulting in social benefits. On the
10 other hand, low platform sociality is more embedded in the anonymity of its users (Perren & Kozinets, 2018). Bleier et al. (2019) identify social presence as a determinant of purchase intentions in online interactions. The relationship between social presence and purchase intentions is higher for experience products than for search products. Gefen and Straub (2003) further recognize exchanges that are rich in sociality have positive relations on intentions to purchase, due to increased social value. Similarly, Möhlmann (2015) states that the probability of repeat sharing increases when such feelings of sociality are facilitated.
2.4 How platforms realize convenience and sociality
Perren and Kozinets (2018) propose four distinct Lateral Exchange Market Types, all varying in degree of platform mediation and platform sociality (Figure 1). Forums connect actors, Hubs centralize exchanges, Matchmakers pair actors, and Enablers merely equip actors. These ecosystems act as intermediaries in value creation through mediation and sociality dimensions. As platforms lacking sociality and mediation characteristics inhibit the connectivity sought in the carsharing economy (Frenken & Schor, 2017), Enablers are not investigated in this study.
The exchange on Forums is facilitated directly between actors, placing the responsibilities of coordination, payment, and location with the individuals. Convenience plays no role in these interactions, which becomes the participants’ responsibility. The result is a social exchange market where trust between peers has a central role, as there is little facilitation by the service platform. Forums ensure value through social connections and interpersonal communication. Sharing is grounded in altruism and seen as a communal act to bond with other community members (Belk, 2010). Such community-driven business models are often not driven by profit motives, which are thought to damage social cohesion (Bardhi & Eckhardt, 2012).
Hubs' primary function is to provide exchange platforms. The platforms facilitate the supply and demand side of the exchange by removing interactions between its users. The search, transaction, and exchange costs are lowered, resulting in higher convenience for both the supply and demand side. Platforms like these engage in pseudo-sharing (Belk, 2014), are driven by profit motives, and provide little or no communal feelings.
Matchmakers combine platform sociality and platform mediation. These platforms facilitate direct interactions between participants and ensure the management and coordination of transactions. In addition to providing convenience dimensions, the high degree of sociality emphasizes the interpersonal benefits P2P platforms can provide.
11 This study further builds on this theoretical construct and focuses on three platform types.
Forums provide high platform sociality but low platform mediation (HL), Hubs provide low platform sociality but high platform mediation (LH), and Matchmakers provide high levels of both platform mediation and platform sociality (HH).
Figure 1. Lateral Exchange Market Types
Adapted from: Perren, R., & Kozinets, R. V. (2018)
2.5 How sociality and convenience translate to willingness to participate
Participation is both essential and a significant obstacle for platforms in online environments (Ardichvili, 2008). Chiu et al. (2006) postulate that shared goals, mutual accountability, social ties, and community sense are important factors for willingness to participate in virtual communities. Furthermore, Zhang et al. (2014) propose that interactivity and sociability features positively influence virtual experiences, and Botsman (2015) states that the desires of individuals to connect and belong to a community directly are reflected through the sharing economy.
Numerous scholars oppose this view; affirming utility and convenience are the central drivers of participation intent. The convenience a platform provides through platform mediation is seen as a major driver of behavior in the sharing economy (Nielsen et al., 2015). The lowered costs related to access, decision making, and transactions strongly influence behavioral intentions (Collier & Sherrell, 2010). Zhang et al. (2016) further suggest that increased convenience enlarges intentions to participate in the sharing economy. The degree of mediation provided on the platform's interface is hereby a precursor to the user's convenience.
12 Kumar et al. (2018) describe convenience and cost as the primary values for choosing a service platform. Likewise, a study by Möhlmann (2015) reveals a significant positive effect of platform utility, reflecting the convenience dimension, on satisfaction, and the likelihood of choosing the platform again. Community belonging, reflecting the desire to connect with others, only affects the latter.
The combination of sociality and convenience in sharing platforms increases the value for its users. This is because the primary economic dimension is extended with rich experiences to connect socially (Schor & Fitzmaurice, 2015). It is therefore hypothesized that sharing platforms with high sociality and high mediation features (HH; Matchmakers) will score highest on willingness to participate. This is because both dimensions, reflecting convenience or utility and need for community, are important participation drivers in the sharing economy. Hypothesis 1a isolates mediation, whereas Hypothesis 1b isolates sociality.
H1a. HH platform characteristics (Matchmakers) are more positively related to willingness to participate than HL platform characteristics (Forums).
H1b. HH platform characteristics (Matchmakers) are more positively related to willingness to participate than LH platform characteristics (Hubs).
Convenience is considered the primary driver of participating in carsharing platforms. Schor (2015) denotes that the sharing economy is more about utilitarian values than "feel-good” values of sociality. A study by De Rivera et al. (2017) demonstrates that network-oriented platforms focusing on efficient and sophisticated infrastructures perform better than socially-oriented platforms focusing on building communities. It is therefore hypothesized that sharing platforms with low sociality and high mediation (LH; Hubs) score higher on willingness to participate than platforms with high sociality and low mediation (HL; Forums). This is because convenience, and therefore utility, remains unchanged when platform mediation is high regardless of platform sociality. As convenience or utility are seen as the predominant reasons for participating in the sharing economy, platform mediation is expected to have a stronger relationship with willingness to participate than platform sociality.
H2. LH platform characteristics (Hubs) are more positively related to willingness to participate than HL platform characteristics (Forums).
2.6 The moderating influence of social and functional motivations
The motivations of individuals partaking in collaborative consumption range from altruistic and sociality-seeking to utilitarian and gain-sociality-seeking (Hamari et al., 2016). Stofberg and Bridoux (2019) propose a
13 distinction between social and economic benefits a platform can deliver. The hedonic value and social reinforcement sharing platforms offer relate to their social and altruistic motivations, whilst utility and savings are more closely related to functional and economic motivations. Hellwig et al. (2015) distinguish between “sharing idealists” and “sharing pragmatics”: idealists are driven by social and emotional value searching for social relationships, whilst pragmatics are driven by the functional value of sharing, whose use is appealed by the convenience and efficiency of sharing platforms. The nature of participants’ motivations may determine their platform type preference, resulting in self-selection based on platform characteristics. In this study, a distinction between social and functional motivations is made, as these can be seen as counterparts.
When functional motivations are prevalent, the platform is seen as a utilitarian mean to the user's goal. Functional motivations tend to underline collaborative consumption models that are heavily reliant on market mediation (Bardhi & Eckhardt, 2012). Hamari et al. (2016) indicate that functional motivations result in positive behavioral intentions towards collaborative consumption, and various scholars identify functional motivations as the main drivers in collaborative consumption models as well (Botsman & Rogers, 2010; Edbring et al., 2016). When users are driven solely by such motivations, sociality aspects provided by platforms are futile and might decrease willingness to participate. Platforms delivering high sociality and high mediation features (HH; Matchmakers) are therefore expected to score lower on participation intent than platforms with low sociality and high mediation features (LH; Hubs), but higher than platforms providing high sociality and low mediation (HL; Forums). Motivations have a moderating effect on the relationship between platform characteristics and willingness to participate, such that the degree of platform sociality reduces willingness to participate for users with strict functional motivations.
H3.1a. The relationship between HH (Matchmakers) and willingness to participate is moderated by functional motivations, such that HH (Matchmakers) are more positively related to willingness to participate than HL (Forums) when functional motivations are high compared to when they are low H3.1b. The relationship between HH (Matchmakers) and willingness to participate is moderated by functional motivations, such that HH (Matchmakers) are less positively related to willingness to participate than LH (Hubs) when functional motivations are high compared to when they are low H3.2. The relationship between LH (Hubs) and willingness to participate is moderated by functional motivations, such that LH (Hubs) are more positively related to willingness to participate than HL (Forums) when functional motivations are high compared to when they are low
14 Alternatively, Lamberton and Rose (2012) indicate social motivations drive collaborative consumption due to the social utility, instead of economic utility, derived from sharing. Likewise, Bucher et al. (2016), indicate that social motivations have the most considerable impact on sharing intentions. Furthermore, van de Glind (2013) provides evidence that intrinsic social motivations, such as the need for social cohesion, are strong determinants for partaking in sharing platforms. The desire to interact with others is thereby a critical factor. Bleier et al. (2019) state that social presence enhances the intention to purchase in experience products rather than in search products. Although carsharing platforms do not offer physical products for sale, implications regarding motivations can be derived. Users guided by social motivations are expected to view the carsharing platform as an experience, or a way to connect with others. This implies that the positive relationship of social presence on willingness to participate increases when personal motivations are social compared to functional. The desire for social cues and communicative means lead the assumption that the positive relation between platform mediation and willingness to participate becomes less strong when participants are guided by strictly social motivations. Platforms providing high sociality and high mediation features (HH; Matchmakers) are expected to score lower on participation intent than platforms with high sociality and low mediation features (HL; Forums), but higher than platforms with low sociality and high mediation features (LH; Hubs). Motivations have a moderating effect on the relationship between platform characteristics and willingness to participate, such that the degree of platform mediation reduces willingness to participate for users with strict social motivations.
H4.1a. The relationship between HH (Matchmakers) and willingness to participate is moderated by social motivations, such that HH (Matchmakers) are less positively related to willingness to participate than HL (Forums) when social motivations are high compared to when they are low
H4.1b. The relationship between HH (Matchmakers) and willingness to participate is moderated by social motivations, such that HH (Matchmakers) are more positively related to willingness to participate than LH (Hubs) when social motivations are high compared to when they are low
H4.2. The relationship between LH (Hubs) and willingness to participate is moderated by social motivations, such that LH (Hubs) are less positively related to willingness to participate than HL (Forums) when social motivations are high compared to when they are low
Table 1 illustrates the hypotheses examined in this study. The conceptual model (Figure 2) illustrates the relationships examined in this study.
15 Table 1. Overview of hypotheses
16
3. Methodology
3.1 Research design
This study utilizes a full factorial vignette design followed by a closed-ended survey questionnaire to collect data. This approach allows the combination of several scenarios for balanced results, and the manipulation of each variable to determine whether participants react differently (Dülmer, 2007). By combining survey techniques with multivariate design attributes, identifying factors that affect behavioral intentions is simplified (Porter, 2001; Rossi & Anderson, 1982). The advantages of qualitative and survey methods are combined in vignette studies, making them superior to utilizing solely one (Finch, 1987).
Participants are randomly allocated to a 2x2x2x2 between-subject design to account for all possible platform characteristic combinations (Table 2). Nearly identical scenarios are presented, and respondents are asked to appraise their willingness to participate and personal motivations. Additionally, demographic characteristics are measured. The vignettes are elaborate descriptions of a fictitious carsharing platform where the independent variables are manipulated (Table 3), resembling realistic situations and ensuring internal validity (Steiner et al., 2017).
3.2 Vignette design
Sixteen vignettes are created, each depicting a similar but different scenario regarding the features of the fictious carsharing platform “SplitCar”. The independent variables (platform sociality and platform mediation) and control variables (Covid-19 and business model) utilized in the vignettes are known as dimensions, each consisting of levels (Table 2). The dimensions are manipulated to identify shifts in participants’ behavior, influencing the outcome variable willingness to participate (Porter, 2001).
Vignettes 11, 12, 15 and 16 depict low mediation and low sociality features, contradicting the carsharing objective of connectivity (Frenken & Schor, 2017). Vignettes 7 and 8 depict B2C platforms with low mediation, which is improbable as the business owns the vehicles. Hence, these designs are omitted from the vignette population used in this study, resulting in ten final vignettes (Appendix A).
17 To ensure the vignettes are interpreted correctly, a qualitative, exploratory pre-test is performed (Appendix B). Respondents are allocated randomly to the vignettes and the open-ended questionnaire examining the various levels to ensure reliability. Examples are "Is it easy to meet new people on the platform?" and "How well does the platform facilitate your booking?" to ensure correct dimensions are communicated. For instance, the phrase “1.5 meters apart” is adjusted to “social distancing regulations”. After accounting for misinterpretations, the vignettes are deemed clear and fit to use as illustrations of the variables.
Table 3 presents the finalized description of the platform sociality, platform mediation, business model and Covid-19 dimensions, as described in the vignettes. These features (except Covid-19) are based on the conduct of real carsharing platforms such as Turo, Getround, and Snappcar to enhance realism. An example of a complete vignette is presented in Appendix C.
18 3.3 Survey instruments
Following the vignettes, respondents are asked to complete a closed-answer survey questionnaire. This allows for precise measurements regarding the moderating variables (functional and social motivations) and the dependent variable (willingness to participate), as well as the control variables. The use of a survey questionnaire permits a straightforward interpretation of the collected data. All responses are measured on a 7-point Likert scale, facilitating data interpretation and manipulation. In addition, data is collected more cost- and time-efficiently compared to face-to-face collection (Wright 2005).
As the sample is international, the survey is in English. This ensures coherence in understanding the questions and overcomes potential linguistic misinterpretations that could be caused by translation. The complete survey questionnaire is presented in Appendix D.
The measures used in this study are adopted from existing constructs and adapted for the fictitious carsharing platform SplitCar, ensuring reliability and internal validity. All items have a Cronbach's alpha > .8, indicating good reliability (Gliem & Gliem, 2003). Control variables are included for demographic differences and attitudes towards the Covid-19 virus, as well as measures to ensure realism and credibility (Appendix E).
3.3.1 Functional motivations.
Functional motivations are measured on a 7-point Likert scale ranging from “completely disagree” to “completely agree”. The measurement is adopted from Hamari et al. (2005) and Stofberg and Bridoux (2019). Items include “I see myself participating because it will improve my economic situation” and have a Cronbach’s alpha of .878.
3.3.2 Social motivations
Social motivations are measured on a 7-point Likert scale ranging from “completely disagree” to “completely agree”. The measurement is adopted from Paul et al. (2009) and Stofberg and Bridoux (2019). Items include “I see myself participating because it creates a feeling of attachment with other members on the platform” and have a Cronbach’s alpha of .888.
3.3.3 Willingness to participate
For willingness to participate, the outcome variable of this study, three scales have been combined from Lamberton and Rose (2012), Pavlou and Gefen (2004) and White et al. (2012). These scales have been manipulated to account for users and car owners. All are measured on a 7-point Likert scale ranging from “completely disagree” to “completely agree”. The measure from White et al. (2012), “I would be likely to
19 rent a car on [SplitCar]” is adapted as follows to account for car owners: “I would be likely to share my care on [SplitCar]”. Both scales have a Cronbach’s alpha of .863.
3.3.4 Control variables
To ensure fair results and control for unrepresentative outcomes, several variables are controlled for. The variables age, gender, and nationality have proven to have significant influences on participation in the sharing economy. Millennials, for instance, engage in sharing activities more often than older age groups (Kats, 2017), and elderly are less socially embedded (Cornwell et al., 2008). Furthermore, women are more likely to experience intrinsic motivations to share (Hellwig et al., 2015). Regarding cultural background, Hofstede (2001) notes that non-Western cultures are often more collectivist, which may result in higher willingness to engage in the sharing economy. Business model is controlled for as well. Further control variables include education, income, and living area and Covid-19.
3.3.5 Covid-19 measures
During this study, the Covid-19 or Corona virus spread across 176 countries, significantly affecting the attitudes and behavior of most consumers. This is primarily due to fear of contagion, which has been studied academically after the MERS epidemic (Jung et al., 2016). Due to social distancing, consumers are less inclined to participate in social activities, potentially influencing this study negatively. Conversely, carsharing might be considered a safer option for public transportation or taxi services. In order to account for potential negative intentions or biased responses, the vignettes account for two levels: Threat of Corona and Neutralized threat of Corona (Table 3). In addition, 7-point Likert scale Corona measurements are included in the questionnaires. Measures include "Social distancing measures are effective in protecting me from Corona" to control for perceived health risks, and "To what extent is your willingness to try carsharing changed by Corona" to control for carsharing safety perceptions.
3.3.6 Realism and credibility
To ensure realism and credibility, two manipulation checks are included in the survey. For both a 7-point Likert scale ranging from “completely disagree” to “completely agree” is adopted from Stofberg and Bridoux (2019). The single-item scales are “I found the situation in the above-mentioned scenario realistic” measuring realism, and “I had no problem imagining myself in the above-mentioned situation” measuring credibility.
20 3.4 Data collection procedure
The data collection is carried out internationally utilizing online distribution. The online links leading to randomized vignettes are dispersed on social media. Friends, family, and colleagues are approached as well, allowing for convenience and snowball sampling. Due to the Covid-19 regulations, physical data collection is limited. As this research is part of a larger study, more variables were measured than accounted for in this study. A total of 2012 out of 3200 respondents completed the questionnaire, resulting in a response rate of 62.9%.
3.5 Analytical plan
After the data is downloaded in SPSS 26, it is cleaned, and missing values are deleted. Categorial variables as platform characteristics and gender are transformed into dummy variables (Appendix F), after which scaled items (platform sociality; platform mediation; functional benefits; social benefits; willingness to participate) are combined into individual variables with Cronbach’s alpha of > .8. No items have been removed to improve internal validity.
To test Hypotheses 1 and 2, the relationship between platform characteristics and willingness to participate, linear regression is performed in SPSS. To test Hypotheses 3 and 4, the interaction effect between platform characteristics and individual motivations on willingness to participate, the PROCESS macro of Hayes (2018) Model 1 is utilized. Individual motivations (functional and social) act as the moderators in this model.
4. Results
Of the 2012 respondents that completed the questionnaire, 1264 completed vignette questionnaires with relevant dimensions to this study (Appendix A). To identify and remove outliers, person-total correlations (PTC) and Mahalanobis distance (MD) tests are performed on all responses (cut off = .001). Both measures detect invalid and fraudulent data in online studies: Mahalanobis distance identifies the distance between responses, and person-total correlation measures the extent to which responses fit general tendencies (Maesschalck et al., 2000; Dupuis et al., 2019). In addition, respondents taking less than 6 minutes to complete the survey are deleted. In total, 152 of the data is removed, resulting in a final sample size of 1112 respondents.
21 4.1 Realism and credibility
One-way ANOVA test is performed to examine the realism and credibility of the vignettes (Appendix G). There is no homogeneity of variances for realism (F = 2.949, p < .001), nor for credibility (F = 3.038, p < .001). This indicates that the vignettes are not equally realistic nor credible. Tukey HSD post-hoc test reveals the mean scores of credibility and realism for all vignettes are above 4.5, on a scale of 1 to 7. The mean score for realism is 5.16, and for credibility 5.03. The vignettes are therefore considered sufficiently realistic and credible.
4.2 Distribution
The dependent variable willingness to participate and both moderating variables social motivations and functional motivations are tested on skewness and kurtosis to ensure the data is normally distributed (Appendix H). None of the variables have skewness issues: the values are between -1 and 1. Kurtosis values are near 0, indicating a near-normal distribution. Kolmogorov-Smirnov and Shapiro-Wilk tests show significant p-values (< .05), indicating a violation of normality. However, due to the large sample size (>50), the violation of normality is not expected to cause issues in the analysis (Pallant, 2007).
4.3 Descriptive statistics and correlations
Descriptive statistics of the sample, consisting of 1112 participants, are presented in Table 4. Of the respondents, 57% are female and 43% are male. The average age is 30 (SD = 11.8), reflecting the Millennial users of the carsharing economy (Kats, 2017). The majority of respondents live in urban areas (84.9%), reflecting the metropolitan occurrence of carsharing initiatives (Kim, 2015). Most respondents are pursuing or have completed an academic education (73.7%), and earn an income below €1500 (45.3%), which is categorized as low. This could be attributed to the large number of students participating in this study, due to convenience sampling. The two most frequent nationalities are Dutch (52%), and Italian (15%).
22 Correlations and descriptive statistics are presented in Table 5. The outcome variable willingness to participate correlates with the HH>HL (r = .204, p = < .01) and the LH>HL vignettes (r = .236, p = < .01). This is in line with Hypothesis 2, suggesting mediation is a more important predictor of participation than sociality. HH>LH isolates platform sociality but does not correlate with either the outcome variable or the moderating variables. It is therefore not expected that sociality influences willingness to participate. Social and functional motivations correlate with willingness to participate (r = .339; r = .312, p = < .01), but not with the vignettes. As these are moderating variables, their interaction effects could be significant. B2C correlates positively with all vignettes and willingness to participate (r = .190, p = < .01), indicating a B2C rather than P2P platform might be preferred amongst participants.Owners, compared to users, correlate negatively with the vignettes and willingness to participate (r = -.292, p = < .01), but positively with functional motivations (r = .070, p = < .01). This could indicate owners are more reluctant to participate in carsharing and are less driven by functional motivations.ddddddddddddddddddddddddddd
Moreover, females are more likely to participate than males (r = .068, p = < .01), reflecting their intrinsic motivation to share (Hellwig et al., 2015). Participants with an academic education correlate negatively with HH>LH vignettes (r = -.074, p = < .05), indicating high mediation is valued more than high sociality. Those with a professional education are less inclined to participate (r = -.084, p = < .01), and are possibly less driven by social motivations (r = -.056, p = < .05). Corona correlates negatively with functional motivations (r = -.063, p = < .01), indicating individuals are driven less by utilitarian motives in presence of the Covid-19 virus.
23 4.4 Regression analysesdddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd To examine which platform characteristics are positively related to willingness to participate, linear regression analyses are performed. The regression tables for insignificant relations, and therefore rejected hypotheses, can be found in Appendix I.
4.4.1 Hypothesis 1a
To test Hypothesis 1a, stating HH platform characteristics are more positively related to willingness to participate than HL platform characteristics, linear regression is performed. In model 1, the control variables owner/user, B2C/P2P, Corona, age, urban/rural, education, gender, income, and nationality are used. In model 2, the variable HH>HL is added to check whether the variable adds statistically significant value to the model. The total variance explained by model 1, indicated by R2, is 20.3% and model 2 is 20.8%. The adjusted-R2, which corrects for the loss of degrees of freedom by the addition of HH>HL increases from 17.8% to 18.2%. There is an increase in R2 of .5%, which is insignificant (p = .076) when adding HH>HL, suggesting model 2 does not add significant explanatory value over model 1. Hypothesis 1a is therefore not supported. There is no statistically significant difference between HH and HL platform characteristics on willingness to participate.
4.4.2 Hypothesis 1b
To isolate the effect of sociality in Hypothesis 1b, investigating whether HH platform characteristics are more positively related to willingness to participate than LH characteristics, a regression analysis is performed. The total variance explained in model 2, indicated by R2, is 17.4%. This is 1% higher than model 1, but the adjusted-R2 indicates a decrease of 1%. The addition of HH>LH does not add explanatory value over the control variables (p = .339). Hypothesis 1b is therefore not supported. There is no statistically significant difference between HH and LH platform characteristics on willingness to participate.
4.4.3 Hypothesis 2dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd To test Hypothesis 2, stating LH platform characteristics are more positively related to willingness to participate than HL platform characteristics, linear regression is performed with the same control variables (Table 6). The total variance explained by model 2, indicated by R2, is 17.5%, compared to 15.5% in model 1. The adjusted R2, correcting for the loss of degrees of freedom by the addition of LH>HL is 14.8%. There is an increase in R2 of 2% (p < .001), suggesting model 2 adds significant explanatory value over model 1. LH>HL has a β of .449 (t = 3.584, p < .001). Participants on LH platforms relative to HL platforms are more willing to participate by 1 x .449 units. This is a significant positive relationship thus Hypothesis 2 is supported.
24 An overview of the regression analyses outcomes for all vignettes is presented in Table 7. Hypotheses 1a and 1b are rejected, whereas Hypothesis 2 is supported.
25 Table 7. Overview of regression analyses for all vignettes on willingness to participate
4.5 Moderation analysis
To test Hypotheses 3 and 4, whether functional and social motivations moderate the relationship between platform characteristics and willingness to participate, six moderation tests are performed (Appendix J). Three determine whether the relationship between platform characteristics and willingness to participate is influenced by functional motivations (3.1a/b, 3.2), and three determine whether those relationships are influenced by social motivations (4.1a/b, 4.2). To isolate the effect of personal motivations, PROCESS Macro model 1 (simple moderation) by Hayes (2018) is performed for each motivation.
4.5.1 Hypothesis 3.1a and 4.1a
To test Hypothesis 3.1a, whether the relationship between HH compared to HL platform characteristics is more positive when functional motivations are high, PROCESS Macro model 1 by Hayes (2018) is used. The results indicate there is no significant direct effect between HH>HL and willingness to participate (β = .592, SE = .482, t = 1.228, p = .220). A significant direct effect of functional motivations on willingness to participate is found (β = .359, SE = .071, t = 5.084, p < .001). However, no interaction effect is concluded (β =-.074, SE = .093, t = -.798, p = .425), hence Hypothesis 3.1a is not supported. Functional motivations do not moderate the relation between HH versus HL platform characteristics and willingness to participate.
To test Hypothesis 4.1a, whether the relationship between HH compared to HL platform characteristics is more negative when social motivations are high, the same analysis is performed. The results indicate there an insignificant direct effect between HH>HL and willingness to participate (β = .255, SE = .129, t = 1.892 p = .059), confirming the findings of Hypothesis 1a. A significant direct effect of social motivations on willingness to participate is found (β = .410, SE = .042, t = 9.749, p < .001). However, there is no interaction effect between the variables (β = -.005, SE = .085, t = -.062, p = .951), hence Hypothesis 4.1 is not supported.
26 4.5.2 Hypotheses 3.1b and 4.1b
To isolate the effect of sociality, the interaction between HH versus LH characteristics and functional motivations on willingness to participate is examined. There is no direct effect of HH>LH (β = -.061, SE = .089, t = -.685, p = .494) on willingness to participate and no interaction effect between HH>LH and functional motivations (β = -.030, SE = .076, t = -.389, p = .697). Functional motivations do have a direct effect on willingness to participate (β = .289, SE = .038, t = 7.563, p < .001), but Hypothesis 3.1b is not supported.
The same analysis is performed to investigate an interaction with social motivations. Results indicate a significant direct effect of social motivations (β = .297, SE = .034, t = 8.617, p < .001), but no significant direct effect of HH>LH on willingness to participate (β = -.108, SE = .088, t = -1.221, p = .222). There is a significant interaction effect between HH>LH and social motivations on willingness to participate (β = .145, SE = .066, t = 2.173, p = .030), meaning Hypothesis 3.1a is supported. These results are illustrated in Table 8 and Figure 3. When social motivations are present, HH compared to LH platform characteristics are more positively related to willingness to participate than when social motivations are not present. This effect is significant at low levels of social motivations (-1SD; β = -.299, SE = .126, t = -2.375, p = .018), but not at high levels of social motivations (+1SD; β = .084, SE = .123, t = .680, p = .497). When social motivations are not present, there is no significant difference in the effect of characteristics on willingness to participate.
Table 8. Linear model of predictors of willingness to participate
Social motivations demonstrate a more positive effect on the relation between HH platform characteristics and willingness to participate than for LH platform characteristics (Figure 3). At high levels of social motivations, platforms employing both sociality and mediation characteristics score higher on willingness to participate than platforms employing solely mediation characteristics.
27 Figure 3. Effect of social motivations on the relation between HH, LH and willingness to participate
4.5.3 Hypotheses 3.2 and 4.2
To test Hypothesis 3.2, whether the relationship between LH compared to HL platform characteristics is more positive when functional motivations are high, PROCESS Macro model 1 by Hayes (2018) is used. The results indicate there is a significant direct effect between LH>HL and willingness to participate (β = .405, SE = .119, t = 3.392, p = .001), confirming Hypothesis 2. In addition, a significant direct effect of functional motivations on willingness to participate is found (β = .323, SE = .043, t = 7.553, p < .001). However, no interaction effect is found (β = -.020, SE = .087, t = -.230, p = .818), hence Hypothesis 3.2 is not supported.
To test Hypothesis 4.2, whether social motivations influence the relationship between LH compared to HL characteristics and willingness to participate, the same analysis is performed. The results indicate a significant relation between LH>HL and willingness to participate (β = .502, SE = .119, t = 4.210, p < .001), confirming Hypothesis 2, and a significant relation between social motivations and willingness to participate (β = .301, SE = .040, t = 7.522, p < .001). There is a significant interaction effect between LH>HL and social motivations on willingness to participate (β = -.174, SE = .081, t = -2.150, p = .032), such that LH characteristics are less positively related to willingness to participate than HL characteristics when social motivations are high (+1SD; β = .296, SE = .159, t = 1.688, p = .092), than when social motivations are low (-1SD; β = .735, SE = .163, t = 4.516, p < .001). These results are presented in Table 9. Social motivations have a greater effect on willingness to participate on HL platforms, thus Hypothesis 4.2 is supported.
28 Table 9. Linear model of predictors of willingness to participate
Social motivations show a more positive effect on the relation between HL platform characteristics and willingness to participate than for LH platform characteristics (Figure 4). At high levels of social motivations, the difference in participation intent between both platform types becomes less large than at low levels of social motivations.
Figure 4. Effect of social motivations on the relation between LH, HL and willingness to participate
The relevant findings of this study are presented below (Figure 5, Table 10). These will be elaborated upon in the discussion section. The hypotheses and their significance are illustrated in Table 11.
29 Figure 5. Significant direct effects and moderation effects confirmed in HH>LH and LH>HL vignettes
30 Table 11. Outcome of the hypotheses
4.6 Additional analysis
Due to the rejection of Hypotheses 1a/b, 3.1a/b, 3.2, and 4.1a, an additional analysis is performed to explore whether B2C, relative to P2P, business models influence willingness to participate. As can be seen in the correlation matrix (Table 2), B2C correlates strongly with willingness to participate.
To test whether there is a significant relationship between business model (B2C vs. P2P) and willingness to participate, a regression analysis is performed (Table 12). It is important to note that the business models are from the user's perspective. The total variance explained in model 2, with the addition of the B2C variable, is indicated by R2, 16.8%. The adjusted-R2, which corrects for the loss of degrees of freedom by the addition of B2C, is 15.3% in model 2. There is an increase in R2 of 1.8% (p < .001), suggesting model 2 adds explanatory value over model 1. This indicates a positive direct effect of B2C,
31 compared to P2P, on willingness to participate (β = -.546, t = -5.072, p < .001). Participants on B2C platforms, relative to P2P platforms, are more willing to participate by 1 x .417 units.
32 To examine whether the positive relationship between B2C and willingness to participate is subject to an interaction effect by functional motivations, PROCESS Macro model 1 by Hayes (2018) is used. Results (Table 13) indicate a significant interaction effect between B2C and functional motivations on willingness to participate (β = -.229, SE = .069, t = -3.319, p = .001), such that B2C is more negatively related to willingness to participate when functional motivations are high (+1SD; β = .127, SE = .126, t = 1.010, p = .313) than when functional motivations are low (-1SD; β = .668, SE = .118, t =5.652, p < .001).
Table 13. Linear model of predictors for willingness to participate
The same analysis is performed for social motivations (Table 14). There is a significant interaction effect between B2C and social motivations on willingness to participate (β = .153, SE = .065, t = 2.332, p = .020), such that B2C is more negatively related to willingness to participate when social motivations are high (+1SD; β = .189, SE = .121, t = 1.562, p = .119), than when social motivations are low (-1SD; β = .602, SE = .118, t = 5.088, p < .001).
33 Figure 6 illustrates that B2C platforms generally score higher than P2P platforms on willingness to participate. Social and functional motivations have a more positive effect on P2P platforms than on B2C platforms. The difference in participation intent between B2C and P2P platforms is less large at high levels of social and functional motivations than it is at low levels of social and functional motivations. Figure 7 illustrates the overall significance of the additional analyses.
Figure 6. Effect of motivations on the relationship between business model and willingness to participate
34
5. Discussion
5.1 Summary of the results
Contrary to expectations, there is no significant direct relationship found between HH>HL nor HH>LH and willingness to participate (H1a, H1b). However, a significant direct relationship between LH>HL and willingness to participate (H2) is concluded, meaning participants confronted with high mediation characteristics are more willing to participate in the carsharing platform SplitCar than users confronted with high sociality characteristics.
Both social and functional motivations have significant direct relationships with willingness to participate. Also, social motivations act as moderators in two of the three examined models. Functional motivations do not act as moderators, as their effects are insignificant. Although there is no significant difference between HH and LH characteristics without the influence of motivations, when social motivations are present, HH characteristics are more positively related to willingness to participate than LH characteristics (H4.1b). A similar effect is found in Hypothesis 4.2, where the presence of social motivations results in HL platform characteristics being more positively related to willingness to participate than LH characteristics. Without the addition of social motivations, there is an opposite relationship where LH rather than HL characteristics are more positively related to willingness to participate.
The additional analysis demonstrates that users are more willing to participate on B2C platforms than on P2P platforms. Social and functional motivations significantly influence this relationship, such that P2P platforms show a greater increase in willingness to participate than B2C platforms. Nevertheless, B2C remains the preferred choice for participants.
5.2 Discussion of the results 5.2.1 Platform characteristics
Platforms employing high levels of sociality and mediation do not lead to increased participation intent compared to different characteristic combinations, indicating Matchmakers are not superior to Forums or Hubs. This is not in line with previous research by Schor and Fitzmaurice (2015), indicating the combination of both sociality and mediation increases platform value for participants. However, the same study proposes that the primary dimension for value creation is the economic one, followed by sociality. This could provide an alternative explanation to the insignificance of combining both characteristics compared to only mediation: the addition of sociality does not add value to the platform. Fear of contagion poses an alternative explanation as to why Matchmakers are not superior to Hubs. Due to the Covid-19 pandemic, social distancing is strongly encouraged in nearly every country, resulting in a natural aversion to social interactions (Flaxman et al., 2020). Hence, consumers may prefer platforms where social
35 interactions are minimized. Alternatively, fear of exploitation or opportunism by other users may impede the attractiveness of social platforms (Bardhi & Eckhardt, 2012).
Furthermore, the results indicate that the presence of mediation characteristics is more important than sociality characteristics in participants’ decisions to participate. This is consistent with the findings of numerous scholars arguing features as enhanced utility, transactions, and convenience increase the value and behavioral intentions of participants in the sharing economy (Collier & Sherrel, 2010; Zhang et al., 2016). Platforms that focus on facilitating exchanges are superior to those focusing merely on interactions and relationships between users, indicating Hubs are superior to Forums. However, willingness to participate is not lowered when combining sociality with mediation characteristics, suggesting there is room for business models as Matchmakers, seeking both convenience and sociality.
5.2.2 Personal motivations
Although functional motivations are identified as the main drivers in collaborative consumption (Botsman & Rogers, 2010; Edbring et al., 2016), in no instance do they influence the relation between platform characteristics and participation intent. These results are inconsistent with previous findings by Bardhi and Eckhardt (2012), suggesting individuals with strict functional motivations self-select into carsharing platforms that are anonymous, and market mediated. This is possibly due to the fact that Hubs do not cater sufficiently to the extrinsic values sought by participants driven by strictly functional motivations (Ryan & Deci, 2000).
Although in absence of motivations there is no difference between platforms providing only mediation and platforms providing both sociality and mediation, the presence of social motivations results in higher willingness to participate for the latter. This demonstrates that one size does not fit all: there is room for a variety of business models, as participants do not possess the same levels of social motivations. Individuals driven by social motivations are therefore more likely to participate in Matchmaker structures than Hubs. This is consistent with Lamberton and Rose's (2012) findings, suggesting social motivations drive sharing due to the social utility (the increased approval by the community) derived from such actions, as opposed to economic utility. Bucher et al. (2016) support this stance by noting that social motivations have the most significant impact on intention to share. Hence, carsharing is perceived as a voluntary and social rather than transactional and monetary activity when social motivations are present.
A similar effect is found when comparing mediation-focused platforms to sociality-focused platforms. Social motivations result in a greater increase in willingness to participate for the latter. When social motivations are present, participants increasingly self-select into Forums rather than Hubs, due to the presence of platform sociality. Although platform mediation is the most vital element, notable as Hubs remain the preferred platform choice, social motivations significantly increase the desire for social
36 interactions on sharing platforms. This is in line with Belk (2014), stating convenience and social motives may co-exist in sharing, but convenience takes overhand to sociality. However, due to previously mentioned Covid-19 implications, the valuation of sociality may be distorted.
Matchmakers and Forums do not experience a difference in willingness to participate in the presence nor absence of social motivations. This could be due to the presence of sociality in both platform architectures; hence participants driven by social motivations are indifferent to the presence of mediation characteristics when social value is conveyed on the platform.
5.2.3 Additional analysis
In this study, participants do not prefer a sharing economy where exchanges are facilitated between peers. B2C business models, often referred to as 'pseudo-sharing' (Belk, 2014), are more positively related to willingness to participate. These business models are consistent with Perren and Kozinets' (2018) description of Hub platforms, which provide only platform mediation and remove social interactions. This relates to previous findings (H2), indicating users are more willing to participate in high mediation platforms rather than high sociality platforms in the absence of motivations. The lower willingness to participate in P2P business models could be due to risk aversion of users resulting from interactions with strangers, and asymmetric information in transactions (Ert et al., 2016). Participants may perceive B2C platforms as secure and reliable due to associations and interactions with a parent company, compared to self-service exchanges on P2P platforms (Philip et al., 2015). This notion calls for further attention and should be examined more elaborately to identify specific drivers of business model preference.
Despite the more positive relationship between B2C and willingness to participate, functional and social motivations increase willingness to participate in P2P platforms more than they do for B2C platforms. Although access-based platforms score higher on participation intent, this indicates that in presence of such motivations a mutual exchange platform becomes more preferred.
The active participation and co-creation of value in P2P business models can be a deterrent factor for users lacking social motivations. Philip et al. (2015) provide evidence for the preference of P2P platforms when functional and social motivations are of importance, supporting the results of this study. Interpersonal connections provided by P2P platforms are appreciated through the desire for communities and interactions, hence participants driven by social motivations are more willing to participate in P2P platforms than those who do not possess social motivations. P2P platforms become more attractive due to their ability to foster interpersonal relations.
Alternatively, findings by Bardhi and Eckhardt (2012) demonstrate users resist building communities around carsharing initiatives, valuing anonymity. However, these results are rooted in commercial platforms. This clarifies why, regardless of motivations, B2C business models score higher on
