Factors Influencing Continuance Intention of Time-Sharing Cars

Abstract

Time-sharing cars, as a sustainable model for transport, have seen rapid developments in recent years. Prior studies on car sharing paid little attention to the continuance intention among users. But understanding ways to cultivate user habits is important to enhance car sharing’s contribution to sustainability. Based on the expectation confirmation model (ECM), this study adopts a user-centered perspective to explore factors affecting the continuance intention of time-sharing electric vehicles through personal cognitive variables (i.e., psychological ownership, familiarity, and trust) and external environmental variables (i.e., facilitating conditions and service quality). An online survey was conducted, and a total of 1072 valid samples were collected. The results of the structural equation modeling show that offline service quality and facilitating conditions had the greatest impact on the perceptions of usefulness and satisfaction users have for car sharing. Perceived usefulness and satisfaction positively predicted continuance intention, as in the ECM. Among customer variables, environmentalism, familiarity, and trust in peers positively predicted the perceived usefulness of car sharing. Psychological ownership played a subtle function by negatively affecting perceived usefulness but positively affecting satisfaction. We discuss the findings and practical implications for stakeholders and offer suggestions for future research.

1. Introduction

The continued development of information and communication technology has allowed the sharing economy to penetrate various industries, creating new business models of tremendous global value. Juniper Research, a UK-based market research firm, estimated that the sharing economy’s market size would increase to USD 40.2 billion by 2022 [1].

In addition to space sharing (e.g., Airbnb and WeWork) and consumer goods sharing (e.g., Chegg and Etsy), transportation is one of the most active segments of the sharing economy. The continuous growth of global car ownership (Car Parc) has brought with it problems such as traffic congestion, a lack of parking spaces, and environmental pollution. Car sharing is a sustainable transportation solution that can reduce the number of private cars in urban areas while allowing individuals travel flexibility [2,3,4]. Even car-sharing service membership can strongly predict car shedding, as users significantly reduce car ownership after registering for the service [5,6].

Recently, time-sharing electric vehicles have dominated the car-sharing market in mainland China. Recent research has proved that the use of electric vehicles in car-sharing systems can significantly reduce carbon dioxide emissions from 14% to 65% compared with using combustion vehicles [7]. This suggests that implementing electric vehicles in car-sharing fleets will further enhance their contribution to environmental sustainability.

The average number of active monthly users of car-sharing platforms in China reached 3.58 million in May 2020 [8]. But car sharing gradually lost its novelty following its initial popularity. To promote sustainable mobility, customers need to cultivate a habit of using time-sharing cars. Thus, understanding factors affecting their continuance intention of the service becomes critical. Extant studies on users’ intentions toward car sharing mostly focus on their acceptance or initial adoption [9,10,11,12]. Research on customers’ continued usage of car sharing is very limited. This study aimed to address this gap and believes the findings are valuable for service providers, users, and the environment. For service providers, it is more economical to retain existing customers than to recruit new ones, as acquiring new users may cost as much as five times that of keeping the current users [13]. And if providers can improve service quality to retain customers, their continued usage will ultimately benefit the environment.

Many past studies applied the expectation confirmation model (ECM) to users’ willingness to continue using information technology [14,15,16]. The time-sharing car service is a new type of information technology. But, until today, research exploring the continuance intention of car sharing among users based on the ECM is very rare. The empirical support of the ECM and its focus on the continuance intention of information technology make it a proper model for our purpose. The ECM highlights the importance of users’ confirmation of expectations, perceived usefulness, and satisfaction on their continuance intention. The original model is concise, but not complex enough to delineate users’ expectations or evaluations toward a specific information system such as car sharing.

Due to its shared nature, car sharing usually requires multi-party collaboration among service providers, sharing platforms, and consumers. Additionally, car sharing can be executed both online and offline, further increasing the number of factors to consider. Therefore, in this study, we incorporated specific external environmental variables (e.g., service quality and facilitating conditions) and personal characteristic variables (e.g., environmentalism, perception of private car ownership, and trust) into the research framework. Compared with relevant research using the theory of planned behavior (TPB), the technology acceptance model (TAM), or the unified theory of acceptance and use of technology (UTAUT) for car sharing [9,10,17], our framework was more comprehensive, as it adopted a user-centered and context-based perspective to investigate both explicit and implicit factors behind users’ continuance intention of car sharing. This approach helped us more fully understand the key drivers and provide specific suggestions for stakeholders such as car-sharing companies to improve their services and retain customers.

At the outset, we hypothesized that users’ perceived usefulness and satisfaction with car sharing would increase if car-sharing companies enhanced user experience, increased the number of deployed vehicles, and ensured that users can use the service anytime and anywhere. Good offline car service quality (e.g., clean interiors, good equipment condition, and accurate vehicle range) would also help improve users’ driving experience and impact their satisfaction with car-sharing services. When users feel satisfied with car sharing and acknowledge its usefulness, their willingness to continue using the service may increase.

The following sections will start with a literature review on car sharing, the ECM, and consumers’ internal and external environmental factors, and then propose our research model. Next, we will describe the methodology of the study and present the results of model fit and hypothesis testing. The final section will discuss the results and implications and offer suggestions for stakeholders and future research.

2. Literature Review

2.1. Car Sharing and the ECM

Car sharing has seen nearly half a century of development, with its core goal remaining the same throughout—sharing cars. Car-sharing services provide users the right to use cars temporarily. Sharing cars allows for reduced ownership costs for personal vehicles while simultaneously increasing usage rates. Car sharing has become a popular business model, thanks in part to developments in mobile networks. Different types of car-sharing services currently exist because of regional differences and target markets. Due to the late entrance of car-sharing services to China, as well as policy and social factors, time-sharing electric vehicles are currently the most popular mode of car sharing in mainland China.

According to the “Guiding Opinions on Promoting the Healthy Development of Small and Mini Bus Leasing” [18], time sharing uses the measure of a minute or an hour for its unit price and uses mobile networks, GPS positioning, and other information technologies to build a networked service platform that allows users to rent cars on their own. Time-sharing cars operate similarly to traditional car rental services. After users register as members, they follow five steps: booking cars, collecting the cars, using the cars, returning the cars, and paying for the cars.

Most studies on car-sharing users focus on the motivating and influencing factors affecting user participation. Past evidence indicated that sociodemographic variables such as gender, age, education, and income could affect car-sharing adoption [19,20,21,22]. Research based on UTAUT, TPB, and TAM also found social influence/subjective norms, performance expectancy/perceived usefulness or benefits, and personal attitude to be important factors in users’ behavioral intentions of car-sharing services [9,10,17].

This study aimed to examine consumers’ willingness to continue using the services based on the ECM. Bhattacherjee studied the continuance intention of online banking by using the ECM, which added the perceived usefulness variable in the technology acceptance model, and established a model for inspecting users’ continuance intention of an information system (IS) [13]. The ECM has four constructs: confirmation, perceived usefulness, satisfaction, and IS continuance intention. Among them, confirmation refers to the consistency between users’ expectations and the actual performance of the IS. Perceived usefulness refers to users’ expected benefits from using an IS. This construct is similar to performance expectancy in UTAUT, as both describe users’ perceived benefits from using an IS. Satisfaction refers to users’ feelings while using the IS. Finally, IS continuance intention refers to users’ intentions to continue using the IS.

The ECM features five main hypotheses. First, users’ satisfaction with the IS has a positive impact on continuance intention. Second, the more users meet their expected benefits from using the IS, the greater their satisfaction. Third, users’ perceived usefulness of the IS will positively affect satisfaction. Many studies on IT adoption have found that perceived usefulness was a key determinant of users’ adoption intentions [23,24,25]. Accordingly, the fourth hypothesis states that users’ perceived usefulness of an IS has a positive effect on their IS continuance intention. Finally, the fifth hypothesis states that the degree to which users’ expectations are met positively affects their perceived usefulness of the IS.

Ever since the ECM was created, it has been widely used in predicting consumers’ continuous usage of products or services. Car sharing is a new online-to-offline business model. While users experience the service offline, they still use IS in the form of mobile apps provided by companies. As this study was based on the ECM, it considered variables featured in the original model, including perceived usefulness, satisfaction, and IS continuance intention, and proposed the following hypotheses:

Hypothesis 1a.

Perceived usefulness of car sharing positively impacts satisfaction.

Hypothesis 1b.

Perceived usefulness of car sharing positively impacts users’ continuance intention.

Hypothesis 2.

Satisfaction of car sharing positively impacts users’ continuance intention.

2.2. Consumer Internal Factors

Electric car sharing is an emerging forward-thinking phenomenon, as evident in its social and economic characteristics [26]. Holak and Havlena pointed out that when studying consumers’ evaluations of new products, product attributes and personal consumer characteristics must be considered simultaneously [27]. In this section, we review the literature and consider car-sharing characteristics and use cases to explore factors affecting the continuous usage of car-sharing services from different perspectives.

2.2.1. Trend Orientation

Manning et al. [28] defined “novelty-seeking” in the context of motivation and processes of consumer innovation. They described it as the degree to which consumers wish to obtain the latest product information. Based on this, Moeller and Wittkowski [29] developed the concept of trend orientation, which is the main purpose for some consumers to obtain the latest products. Highly trend-oriented consumers have high demands for new product information [29] and are more sensitive to technology or IS usefulness [30]. As electric car sharing is a relatively new phenomenon, consumers with greater trend orientation may perceive greater usefulness of this type of service and have greater satisfaction from using the service. Therefore, we proposed the following hypotheses:

Hypothesis 3a.

Trend orientation positively impacts the perceived usefulness of car sharing.

Hypothesis 3b.

Trend orientation positively impacts consumer satisfaction with car sharing.

2.2.2. Environmentalism

Environmentalism is the intent and actual effort toward protecting the environment [31]. Bamberg pointed out that environmental awareness affects individuals’ behavioral intentions through norms, beliefs, and attitudes [32]. Consumers’ perceptions of a product’s environmental impact can influence their purchase decisions. That is, how much the consumers are environmentalists affects their subsequent consumption behavior [32,33].

Car sharing reduces the idle time of a single car by letting others use it. In the long run, this can effectively reduce greenhouse gas emissions [34] and, to some degree, change people’s travel habits. Past studies have confirmed the impact of environmental awareness on the use of low-carbon emission travel modes [11,17,21,22,34,35]. Thus, we speculated that users who are more environmentally conscious are more likely to perceive the ecological value of car sharing, which will positively impact its perceived usefulness and consumer satisfaction. The following hypotheses were proposed:

Hypothesis 4a.

Environmentalism positively impacts the perceived usefulness of car sharing.

Hypothesis 4b.

Environmentalism positively impacts consumer satisfaction with car sharing.

2.2.3. Psychological Ownership

The concept of psychological ownership has its roots in organizational behavior research. It is used to explain people’s feelings of ownership in informally owned companies [36]. Pierce et al. [36] defined it as the extent to which an individual feels an object belongs to them. They argued that while people can develop a strong sense of ownership over material and immaterial properties, this ownership is not completely equal to de facto ownership. For example, people may feel a sense of ownership over items that are not legally theirs.

When acquiring ownership through a purchase, the buyer obtains full rights over the property. Not only can they use the item, but they can also allow or prohibit others from using, renting, or selling it [37]. Under the sharing economy model, consumers obtain the right to use goods rather than the right to own them [38]. Car sharing allows people to use cars without buying them [6]. Research has indicated that people often hold strong emotional attachments to private vehicles [39], and consumers with higher levels of psychological ownership also possess a greater willingness to choose private cars as their mode of transport. Conversely, those with lower levels of psychological ownership consider car sharing as the closest substitute to private vehicles [40].

The relationship between levels of psychological ownership and attitudes toward car sharing has been discussed little in the past. A few studies have found that car ownership has a negative influence on users’ acceptance of car-sharing services, especially for roundtrips [9,21]. Based on related literature, we inferred that the higher the levels of psychological ownership, the greater the attention paid toward vehicle ownership, which will cause consumers to ignore the usefulness of car sharing and decrease their satisfaction with it. Therefore, we proposed the following hypotheses:

Hypothesis 5a.

High levels of psychological ownership negatively impact the perceived usefulness of car sharing.

Hypothesis 5b.

High levels of psychological ownership negatively impact satisfaction with car sharing.

2.2.4. Familiarity

Individuals conduct cost–benefit assessments when exposed to new technologies [41]. Personal value judgments directly impact the perceived difficulty of operating and learning new technologies. Some consumers may be reluctant to use shared services because they do not have any prior experience with them. A high degree of familiarity with shared services can help minimize users’ perceived transaction costs. Thus, familiarity is one factor that impacts the continued usage of shared services [42].

In this study, familiarity covers two contexts: users’ familiarity with (1) car-sharing apps and (2) electric vehicles. The former refers to a user’s ability to skillfully use car-sharing apps to reserve a vehicle, obtain its location, and return the vehicle, among other actions. The latter refers to a user’s ability to complete pre-driving inspections and other related procedures upon receiving the vehicle. According to Davis et al. [41], familiarity from learning is a type of external variable apt to influence the perceived ease of use and usefulness of a system. That is, the more familiar users are with car sharing, either through prior experiences or gathering relevant product knowledge, the more confident they are in using car-sharing services, thereby making these services easier to use. When users perceive new technologies to be easier to use, their perceived usefulness toward them will be higher [23,24,25]. Möhlmann [42] also found that familiarity had a positive impact on car2go satisfaction. Therefore, we speculated that familiarity has a positive impact on perceived usefulness and satisfaction and proposed the following hypotheses:

Hypothesis 6a.

Familiarity positively impacts the perceived usefulness of car sharing.

Hypothesis 6b.

Familiarity positively impacts satisfaction with car sharing.

2.2.5. Trust

In the sharing economy, the concept of trust includes trust in sharing economy platforms (trust in the platform) and trust in other consumers who participate in sharing economy activities (trust in peers) [43]. Trust reduces perceived risks and grants a sense of security during usage or transaction [44,45]. Consumers’ trust in car-sharing platforms and other users can give rise to the belief that car sharing is a less problematic mode of transport, which improves its perceived usefulness. The positive impact of trust on satisfaction has been supported [46]. Therefore, we speculated that trust has a positive impact on consumers’ perceived usefulness and satisfaction and proposed the following hypotheses:

Hypothesis 7a.

Trust in platform positively impacts the perceived usefulness of car sharing.

Hypothesis 7b.

Trust in peers positively impacts the perceived usefulness of car sharing.

Hypothesis 7c.

Trust in platform positively impacts satisfaction with car sharing.

Hypothesis 7d.

Trust in peers positively impacts satisfaction with car sharing.

2.3. External Environmental Variables

2.3.1. Facilitating Conditions

Facilitating conditions refer to the degree of resource support individuals believe they can obtain while using a certain information technology. If users can obtain comprehensive technical support, they will perceive the usage process as convenient [47]. Car sharing becomes more attractive to consumers when parking spaces are included in the service [48]. The success of electric car sharing also calls for a sound rental and charging network. In China, however, an insufficient number of rental stations and the small coverage of car-sharing services are obstacles to the service’s development [49].

Facilitating conditions while using car-sharing services is a type of external environmental variable that will affect user experience and their final behavioral intentions. Limiting factors, such as inconvenient charging and parking facilities, can increase user costs to varying degrees and reduce their satisfaction and perceived usefulness of car-sharing services. Therefore, we speculated that facilitating conditions will positively affect perceived usefulness and satisfaction of car sharing and proposed the following hypotheses:

Hypothesis 8a.

Facilitating conditions positively impact the perceived usefulness of car sharing.

Hypothesis 8b.

Facilitating conditions positively impact satisfaction with car sharing.

2.3.2. Service Quality

Improving service quality has always been an essential strategy in corporate competition [50]. Measuring service quality using specific attributes can help managers improve consumers’ perceived value of services. Cheng et al. [51] differentiated online and offline service quality based on the contexts in which consumers use car-hailing services, which consists of five key factors, including structural assurance, platform responsiveness, information congruity, competence, and empathy. We referenced the online and offline service quality of mobile car-hailing services in the evaluation of car-sharing service quality. Also, to align service quality measurement to consumer needs, we replaced the variable “competence” and “empathy” with “vehicle condition” based on preliminary interviews with car-sharing users. Vehicle condition refers to the appearance and maintenance status of the electric vehicle’s equipment when reserved by users.

Previous studies have found that the better the service quality provided by an IS, the more significant its positive impact on the perceived usefulness and subsequent satisfaction of the IS [51,52]. Therefore, we proposed the following hypotheses:

Hypothesis 9a.

Online service quality positively impacts the perceived usefulness of car sharing.

Hypothesis 9b.

Offline service quality positively impacts the perceived usefulness of car sharing.

Hypothesis 9c.

Online service quality positively impacts satisfaction with car sharing.

Hypothesis 9d.

Offline service quality positively impacts satisfaction with car sharing.

Based on the above, we proposed the following research framework (Figure 1):

3. Methodology

We used online surveys to collect data from adults aged 18 and over. To ensure the survey’s reliability and validity, we invited 14 car-sharing users for a semi-structured interview before conducting the online surveys. We gathered information on users’ motives, behaviors, and opinions on car sharing to assist in creating the survey questions. Additionally, we conducted a small-scale pilot study and collected 196 valid samples to test the reliability of the measurement. The reliability analysis found that the Cronbach’s alpha for each variable was >0.80, indicating that the measures are reliable.

The formal survey was conducted on an online survey platform (www.wenjuan.com, accessed on 15 March 2019). We also distributed survey links on platforms such as Weibo, Baidu Tieba, and Douban to get in touch with consumers who have prior car-sharing experiences. The online survey platform recorded the IP addresses of survey respondents to avoid respondent duplication. The formal survey of this study was available for one month, between 1 April 2019 and 1 May 2019. A total of 1072 valid samples were gathered, and having a driver’s license and having used car-sharing services were the inclusion criteria.

The survey included items on “car-sharing experiences”, “car-sharing opinions”, “evaluations after using car sharing”, “ECM variables in car sharing”, and demographic variables (e.g., gender, age, education, and private car ownership).

“Car-sharing experiences” mainly asked respondents to reply with their most frequently used car-sharing apps, the experience they had with car sharing, their average monthly use of car sharing, average travel distance per use of car sharing, and the usual scenarios where they use car sharing.

“Car-sharing opinions” mainly sought to understand respondents’ personal consumption beliefs and car-sharing perceptions. The measures of trend orientation and environmentalism were based on the indicators proposed by Moeller and Wittkowski [29] and Pavlou and Gefen [53]. The final items were revised to fit the car-sharing context. Psychological ownership was measured based on the scale of Paundra et al. [35], which was designed in reference to how ownership was defined by Van Dyne and Pierce [54]. Familiarity included two aspects: familiarity with car-sharing apps and familiarity with electric vehicles. Measurement items under familiarity were formulated based on the user interviews. Trust was measured based on the two dimensions developed by Möhlmann, including trust in the platform and trust in peers in the sharing economy [43]. To make the items more suitable for car sharing, users’ views gathered in the interview were considered while creating the measurement. This survey used a 7-point Likert scale to measure how much respondents agree with each item. The higher the score, the higher their degree of agreement, with a score of 1 representing being completely disagree and a score of 7 representing being completely agree.

“Evaluations after using car sharing” mainly asked respondents about their evaluations of car-sharing service quality and convenience. A 7-point Likert scale was also used in this section. Service quality referenced the key factors of mobile car-hailing service quality proposed by Cheng et al. [51]. Online service quality included structural assurance and platform responsiveness. Offline service quality included information congruity and vehicle condition. The measurement of facilitating conditions was based on the dimensions of convenience defined by Brown [55]: time, place, acquisition, use, and execution. The final items were revised to fit the car-sharing context.

“ECM variables in car sharing” included three ECM variables, namely, perceived usefulness, satisfaction, and continuance intention. We used a revised version of the measures of Bhattacherjee [13] and used a 7-point Likert scale for these items.

See Appendix A, Appendix B and Appendix C for a complete list of the measurement items in this study.

4. Results

4.1. Sample Composition

Of the 1072 valid samples in this survey, the majority of them were male car-sharing users; 70.7% of users were male, and 29.3% were female. The results are consistent with prior studies [20,21,22] and the findings of CBNData’s “2018 Big Data Report on Car-sharing Users”, which found that the majority of car-sharing users were male [56]. Respondents were mostly between 21 and 35 years old (80.3%). The majority of respondents hold a bachelor’s degree (67.9%). Furthermore, 54.7% of respondents did not own a private car, which was slightly higher than the number of people who purchased a private car (45.3%). This is also consistent with the findings of Martin et al., who found that the main clients of car-sharing services were highly educated people aged 20 to 40 who did not purchase a private car [6].

4.2. Car-Sharing Usage

According to the results of this study, the vast majority (91.8%) of respondents in mainland China have used car-sharing services in the past two years, and 43% of them use car sharing an average of 1–4 times a month. This shows that most users have yet to habitually use car sharing for transport. Additionally, the most popular scenario where respondents use car sharing was in “suburban areas with inconvenient public transportation services”, which accounted for 40.2% of respondents. This was followed by use in “tourist spots”, accounting for 17.4% of respondents. This suggests that car-sharing operators can focus on developing rental stations as a supplement to other public transportation services.

4.3. Model Fit and Hypothesis Testing

This study used AMOS 21.0 for confirmatory factor analysis and structural equation modeling (SEM). First, confirmatory factor analysis shows that factor loading (λ), composite reliability (C.R.), and Cronbach’s α coefficient for each construct are within the acceptable range. This indicates that each construct has good reliability and explanatory power. Next, demographic background and car-sharing experience-related variables were used as control variables in the SEM analysis. The results of the goodness of fit test show that the research model of this study meets the fitness standards; that is, there is a good model fit (CMIN = 5187.11; df = 2606; CMIN/df = 1.99; p = 0.00; CFI = 0.94; RMSEA = 0.03), confirming the suitability of the research model.

As to the hypothesis testing results (see Figure 2 for the SEM path analysis results), the original assumptions in the ECM are all established. The relationships between perceived usefulness and satisfaction (β = 0.19, p < 0.001) and continuance intention (β = 0.28, p < 0.001) both show significant positive effects. Thus, H1a and H1b hold true. This indicates that the greater the perceived usefulness of car sharing among users, the greater their satisfaction and willingness to continue using the service. Additionally, satisfaction (β = 0.55, p < 0.001) also has a significant positive impact on customers’ willingness to continue using the service. Thus, H2 also holds true.

Next, seven sets of hypotheses predicted the effects of trend orientation, environmentalism, psychological ownership, familiarity, trust, facilitating conditions, and service quality on perceived usefulness and satisfaction, respectively. The results show that 10 of the 18 hypotheses are supported. Among them, environmentalism (β = 0.08, p < 0.01), familiarity (β = 0.07, p < 0.001), trust in peers (β = 0.06, p < 0.01), facilitating conditions (β = 0.28, p < 0.001), and offline service quality (β = 0.63, p < 0.001) all have significant positive effects on perceived usefulness, whereas psychological ownership negatively impacts perceived usefulness (β = −0.05, p < 0.01). These findings are consistent with the proposed hypotheses. Thus, H4a, H5a, H6a, H7b, H8a, and H9b hold true. Moreover, trust in peers (β = 0.04, p < 0.05), facilitating conditions (β = 0.65, p < 0.001), online service quality (β = 0.06, p < 0.05), and offline service quality (β = 0.26, p < 0.001) have positive effects on satisfaction. Thus, H7d, H8b, H9c, and H9d also hold true. However, trend orientation and trust in the platform have no significant positive impact on perceived usefulness and satisfaction. Therefore, H3a, H3b, H7a, and H7c are not supported. The impact of environmentalism, psychological ownership, and familiarity on satisfaction, as well as that of online service quality on perceived usefulness are not as expected, so H4b, H5b, H6b, and H9a cannot be supported either (see

Table 1. Results of hypotheses testing.

Hypothesis and Structural Path	β	p-Value	Result
H1a: PU → S	0.19	<0.001	Accepted
H1b: PU → CI	0.28	<0.001	Accepted
H2: S → CI	0.55	<0.001	Accepted
H3a: TO → PU	0.02	>0.05	Rejected
H3b: TO → S	0.00	>0.05	Rejected
H4a: EN → PU	0.08	<0.01	Accepted
H4b: EN → S	−0.00	>0.05	Rejected
H5a: PO → PU	−0.05	<0.01	Accepted
H5b: PO → S	0.05	<0.01	Rejected
H6a: F → PU	0.07	<0.001	Accepted
H6b: F → S	−0.04	<0.01	Rejected
H7a: TPL → PU	−0.02	>0.05	Rejected
H7b: TPE → PU	0.06	<0.01	Accepted
H7c: TPL → S	0.03	>0.05	Rejected
H7d: TPE → S	0.04	<0.05	Accepted
H8a: FC → PU	0.28	<0.001	Accepted
H8b: FC → S	0.65	<0.001	Accepted
H9a: ONSQ → PU	0.04	>0.05	Rejected
H9b: OFFSQ → PU	0.63	<0.001	Accepted
H9c: ONSQ → S	0.06	<0.05	Accepted
H9d: OFFSQ → S	0.26	<0.001	Accepted

Notes: The significance threshold was set at 0.05. TO = Trend Orientation; EN = Environmentalism; PO = Psychological Ownership; F = Familiarity; TPL = Trust in Platform; TPE = Trust in Peers; FC = Facilitating Conditions; ONSQ = Online Service Quality; OFFSQ = Offline Service Quality; PU = Perceived Usefulness; S = Satisfaction; CI = Continuance Intention.

for the summary of hypothesis testing results).

5. Discussion

According to the SEM path analysis results, customers’ perceived usefulness of and satisfaction with car sharing continue to play key roles in the model, as both positively predict their continuance intention of car sharing. The findings are consistent with the ECM [13]. The more users perceive car sharing as fulfilling their needs and giving them satisfactory experiences, the more likely they will be to re-use the services.

Among customers’ internal and external variables, car sharing’s offline service quality and facilitating conditions both show strong positive effects on their perceived usefulness and satisfaction of the services. This indicates that the information quality on car-sharing apps, the condition of vehicles, and the convenience of accessing the services all significantly affect users’ cognitive and affective responses toward car-sharing services. The findings are relatively novel with practical implications. Particularly, the detailed items of offline service quality can give car-sharing operators specific guidance to improve their services. In order to retain their customers, operators need to ensure information accuracy on the apps and keep their vehicles well-maintained and available. Especially during the COVID-19 pandemic, people have higher standards for hygiene and may become more cautious of car sharing. In response to the pandemic, a recent study showed that people do have greater trust in car-sharing operators complying with cleaning requirements [9]. As the tendency may still exist in the post-pandemic era, car-sharing providers should continue to take heed of the sanitation and condition of their vehicles.

Car-sharing companies also need to solve the issues of access and parking to ensure user convenience anytime and anywhere they wish to use the services. Respondents interviewed for this study mentioned that a lack of pick-up stations nearby or bookable cars on the apps are problems often encountered while using car-sharing services. Providers need to ensure their deployment of stations and the vehicle fleets they have built can meet the users’ needs.

On the customer side, results indicate that environmentalism and familiarity positively affect the perceived usefulness of car sharing. These findings resonate with prior studies in which users’ environmental awareness and perceived ease of use lead to greater perceived benefits from car sharing [11,17,21,22,42]. Also, trust in peers positively affects both perceived usefulness and satisfaction. This evidence is novel, as extant studies on car sharing have not used a two-fold trust construct (trust in the platform and in peers) to examine its differential effects. Trust in peers appears to have a greater impact compared with trust in the platform in influencing customer responses. The sharing economy is based on collaborative consumption. Users share the commodities, so they will expect others to maintain the good condition of the commodities after usage. Other than the sanitation and general safety concerns of shared cars, shared electric vehicles require additional attention to the charging conditions. This may increase the weight of trust in peers in users’ evaluations.

On the other hand, trust in the platform has no significant positive impact on perceived usefulness and satisfaction. A possible reason could be the proliferation and development of mobile services in mainland China. Consumers frequently make daily transactions online and have become accustomed to using online platforms. Security is not their main concern, as they are less sensitive to it. Hence, when using car-sharing services, people are less likely to take platform security into consideration.

Trend orientation has no significant positive impact on perceived usefulness and satisfaction either, which is not consistent with the predictions of this study. Car-sharing companies in mainland China mostly use low- and medium-end vehicles due to cost considerations. So even though users might recognize car sharing as a forward-thinking mobility trend, they may not find the vehicles to be trendy in actual use. Thus, perceived usefulness and satisfaction cannot be significantly improved.

Additionally, few studies have explored the impact of psychological ownership on car-sharing behavioral intentions. The core of sharing economy activities is consumption in the form of non-ownership. In the case of car sharing, it is a service that replaces private cars with rental cars [6]. Users of sharing economy services may be more willing to share things with others when they are less concerned with owning the shared goods. In this study, psychological ownership does have a significant negative impact on perceived usefulness. That is, consumers with higher levels of psychological ownership have a lower perceived usefulness of car sharing, which is consistent with our prediction.

However, psychological ownership is found to have a significant positive impact on satisfaction, contrary to our prediction. More users in the sample did not own a private car, which may explain this finding. Although car ownership can be important for people without cars, when car sharing is able to meet similar needs and provide additional perks, such as using electric vehicle license plates (bypassing many cities’ license plate restriction policies in mainland China), having dedicated charging stations, and removing the need to find parking spaces, user satisfaction can increase. The role of psychological ownership revealed here in effect echoed a recent study which found a negative indirect effect of car ownership on user acceptance of electric car-sharing services [9].

Overall, this study verifies the roles of perceived usefulness and satisfaction in the ECM as key antecedents of users’ continuance intention of time-sharing cars. It contributes to current literature by identifying significant internal as well as external variables affecting customers’ continuance intention of car sharing. For external factors, it highlights the importance of offline service quality and facilitating conditions for service providers to attend to. For internal factors, it suggests raising users’ environmental awareness and familiarity of car sharing and increasing trust in peers to help elevate the service’s perceived usefulness. It also discovers the subtle impact of psychological ownership on continuance intention, although, in total, the construct’s negative influence on continuance intention is greater than its positive one. This suggests that lower psychological ownership of private cars can still increase customers’ continued usage of car sharing. These findings, if utilized properly, may help consumers build a habit of using car sharing as a mode of transport.

6. Suggestions

In sum, based on the findings and user needs, we suggest the following recommendations for policy makers and car-sharing operators: (1) the government and car-sharing companies should improve urban planning to enable reasonable access to car-sharing services, including a more efficient distribution of car-sharing stations and, where costs permit, an increase in infrastructure (e.g., charging stations) and the number of deployed vehicles, to meet customer needs; (2) car-sharing companies can improve operational efficiency through technical solutions, such as analyzing big data to create a smart scheduling system for vehicles to increase each car’s utilization rate; and (3) car-sharing companies can establish a joint smart driving platform to address the last mile problem between users and car pick-up stations. For example, GoFun, a car-sharing app, features automatic parking and sets boarding locations within certain areas. With these features, users do not need to pick up and return vehicles at the station in person. When they reserve a vehicle through the app, the rental car will drive itself to users with the aid of driverless technology. After users arrive at their destination, the car can automatically park itself.

Furthermore, offline service quality, which includes information congruity and vehicle condition, has a very significant positive impact on perceived usefulness and satisfaction. Hence, in addition to considering the online service quality of car-sharing apps, companies should improve the various offline service quality factors measured in this study to retain users, including information accuracy and consistency on the apps, as well as good vehicle functions. Given this, we further recommend the following: (1) companies should collaborate with automobile manufacturers to customize electric vehicles and launch car-sharing-specific cars with greater mileage and charger efficiency to meet consumer demands, and (2) companies should improve vehicle maintenance or updates to ensure their deployed cars are in good condition, especially the sanitation and equipment condition of the cars. This will improve the driving experience and satisfaction among users and consequently increase their willingness to continue using car-sharing services.

Also, as users’ personal cognitive variables, including environmentalism, familiarity, trust in peers, and psychological ownership, can influence their perceived usefulness and continuance intention of car sharing, government and service providers should market the ecological and economic benefits of car sharing and provide incentives such as free trials, mileage promotion, or membership rewards to attract and retain users.

Finally, this study used car-sharing users as its research object but did not consider the impact of residence among demographic variables. Traffic congestion, private car purchase, and usage restrictions in different cities may affect users’ perception of car-sharing usefulness. Becker et al. [57] pointed out that urban areas with good public transportation are more conducive to promoting car-sharing services. Prieto et al. [20] found that residents living in city centers were more likely to use car-sharing services for travel. Thus, future studies may explore the similarities and differences in car-sharing continuance intention among residents in different regions to enrich findings in the field.

Another aspect to look into is customers’ attitudes toward electric cars. In the present study, we only focused on the willingness to continue using car-sharing services without further exploring its impact on attitudes toward electric vehicles. Most vehicles used in car-sharing services in mainland China are electric. Qiao and Ji [58] believe that car sharing is key to promoting electric vehicle adoption in China. Consumers can become more aware of electric vehicles and reduce their concerns by using car-sharing services, thereby building a habit of using electric vehicles and indirectly improving their opinions toward electric vehicles. This will eventually help to mitigate environmental pollution. Also, while we used experience as a control variable in the analysis, we found that it has a significant positive impact on the continuance intention of car-sharing services. This is in line with extant research, which indicated that prior experience increases users’ behavioral intentions of electric car sharing [9]. Future studies may explore whether, compared with non-users, consumers with different car-sharing experiences or usage frequencies have different attitudes toward electric vehicles. This will aid in our understanding of the impact of electric car-sharing usage on consumers.