Use of Robotaxi Services for Sustainable Transportation: Focusing on Their Perceived Benefits and Sacrifices as Well as Consumers’ Technology Readiness

Abstract

As a part of sustainable transportation, robotaxis have been rapidly developing around the world because of their advantages in energy saving, improving road safety, and enhancing environmental sustainability, thereby providing consumers with sustainable transportation services. In China, as the number of pilot cities increases, more people are using robotaxi services. This study investigates the factors that affect consumer satisfaction and behavioral intentions after using a robotaxi, aiming to provide data to guide market strategy decisions. To do this, the value-based adoption model was extended and modified by including the technology readiness variable to examine satisfaction, intention to reuse, and electronic word-of-mouth (e-WOM) intentions. Using 425 valid responses, structural equation modeling (SEM) and multi-group analysis were carried out with AMOS 26.0. The results indicate that perceived usefulness, enjoyment, optimism, and innovativeness positively influence service satisfaction, whereas perceived risk and discomfort have negative effects. Consumer satisfaction positively affects both intention to reuse and e-WOM intention. Additionally, uncertainty avoidance shows a moderating effect between satisfaction and intention to reuse.

1. Introduction

With the process of urbanization and the increase in car usage, environmental problems caused by automobiles in large cities are becoming more and more serious [1]. Therefore, in order to meet the growing demand for transportation and at the same time alleviate potential social, economic, and environmental negative impacts, sustainable transportation solutions are needed [2]. Sustainable transport refers to making sure that how we get around today does not mess things up for people in the future [3]. It is basically about using the ideas behind sustainability when building and designing things like roads and public transit systems [4]. However, rapid advancements in autonomous driving technology along with the integration of taxi services have recently led to the emergence of robotaxis, in which robots replace human drivers [5]. The robotaxi system integrates the convenience of taxi services with the economic effectiveness of one-way vehicle sharing based on autonomous driving [6]. Robotaxis offer a more flexible, energy-efficient, economical, and intelligent mode of transportation than traditional taxis and can meet the mobility demands of users with fewer vehicles. In this way, the robotaxi is expected to greatly improve vehicle utilization, help reduce labor costs, and provide smarter, more efficient, sustainable, and convenient services [7,8].

Autonomous vehicles (AVs) use various technologies, including computer vision, information, and communication technologies along with artificial intelligence (AI), which allows them to perceive their surrounding environment and make judgments accordingly [9]. Autonomous driving technology potentially decreases risky driver behavior, transportation costs, traffic congestion, and carbon emissions, while improving road safety, enhancing mobility independence for older adults and people with disabilities, and enabling activities such as automated parking and engaging in leisure or work during travel, thereby increasing human productivity. AVs are expected to bring about a profound transformative socio-technical transition in the automotive industry [10].

Robotaxis represent one of the most advanced and leading application cases in urban mobility services based on AVs [11]. When passengers apply for mobility services through a mobile application, the robotaxi service (RTS) dispatches an AV [12]. SAE International (2021) [13] provides a six-level classification system as the standard for assessing automation levels in AVs, ranging from Level 0 (fully manual) to 5 (fully autonomous). Robotaxis operate at Level 4 (high automation) or higher. At this level, no human intervention in vehicle operation is required, with possibly no steering wheel or pedals in the vehicle [5]. In this study, an RTS is defined as a shared mobility taxi service in which users hail an unmanned AV using a mobile application and the AV is dispatched to provide the transportation service.

The service models of several robotaxi companies in China, such as Apollo Go, Pony.ai, and WeRide, are similar to those of general mobile app-based ride-hailing services. When customers select their pick-up and drop-off locations using a mobile application, the system automatically assigns a vehicle and plans the optimal driving route. The vehicle autonomously navigates to the customer’s pickup location, where the customer can only board the vehicle through identity verification using a mobile authentication code or QR code. The entire trip is conducted under full autonomous driving, and the fare is automatically processed through the system upon arrival at the destination.

Many companies are actively investing in robotaxis, with related pilot programs and commercial services underway in the U.S., China, South Korea, Japan, Germany, and France [8,14]. The Chinese robotaxi market is expected to reach USD 200 million in 2025, with the global market expected to reach USD 300 million. By 2035, the Chinese robotaxi market will increase to USD 179.4 billion, while the global market will reach USD 352.6 billion. Accordingly, China is expected to account for more than half of global market share [15]. Therefore, continued research on robotaxis remains relevant and meaningful.

Many researchers have begun exploring users’ intention to accept RTSs [5,16,17,18] and to continue using RTSs [19] within business models based on shared mobility. But a lot of research about people’s willingness to use new tech only looks at the good side, using models like TAM or UTAUT. On the other hand, when actual users decide if a tech service is worth it, they think about both the upsides and possible downsides [20]. Moreover, studies have only paid attention to users’ initial intention to accept RTSs, whereas empirical research on satisfaction or behavioral intentions after service use (intention to reuse and word-of-mouth (WOM))—key factors for business success—is limited. Specifically, studies regarding customer satisfaction with RTSs, intention to reuse, and WOM are scarce, necessitating the conduction of further in-depth research.

This research uses an adjusted value-based adoption model (VAM) to look at how people decide whether to use a new tech. The model considers not just the good things people expect from the technology, but also what they might have to give up to use it [21]. The VAM is specifically adapted to the robotaxi context by introducing the variable of perceived risk, which is usually applied in extended VAM research [22,23,24]. Furthermore, as technology continues to drive innovative transformations in service delivery across the majority of service sectors, efforts should be made to comprehend customer responses to advanced technologies [25]. Parasuraman (2000) [26] came up with the technology readiness index (TRI) to better understand how people react to tech. This index basically checks how willing and able someone is to start using new technologies. Earlier research also shows that tech readiness plays a big role in how satisfied users feel and what they plan to do after trying out high-tech services—stuff like self-checkouts or airline kiosks [27,28]. Identifying differences in satisfaction and behavioral intentions based on consumers’ technology readiness can help companies more deeply understand consumer experiences of using robotaxis and their post-use behavioral intentions.

Based on the foregoing, this study focused on examining customer satisfaction with RTSs and behavioral intentions after service use by applying the VAM and TRI. It also explores the moderating effect of consumers’ uncertainty avoidance because RTSs involve multiple types of uncertainty such as uncertainty based on deep learning (e.g., data noise and model incompleteness) and domain-specific uncertainty unique to autonomous driving (e.g., visibility and occlusion, traffic regulations, and right-of-way issues) [29]. Based on these accounts, the study raises three research questions:

• RQ1. Which factors affect satisfaction with RTSs?
• RQ2. Does satisfaction with RTSs affect intention to reuse and eWOM?
• RQ3. Does uncertainty avoidance play a moderating role?

With these research questions being answered, this study aims to provide a theoretical foundation for the strategies necessary for the diffusion and commercialization of RTSs as well as demonstrate the validity of those strategies through empirical analyses. The findings are expected to not only offer foundational data that assist consumers in more efficiently using RTSs but also promote RTSs to be commercialized and popularized.

2. Theoretical Background and Hypothesis Development

2.1. Value-Based Adoption Model

The TAM widely serves to explain consumers’ acceptance of and intention to use information technology. However, the core constructs of the TAM do not fully reflect diverse information environments, and the model has limitations, particularly in assessing users’ cost–benefit comparisons [30]. Based on this, Kim et al. (2007) [30] developed the VAM when investigating the mobile internet [31].

The VAM was developed to explore customers’ personal technology use, with a focus on maximizing value [22]. From the perspective of the VAM, customers accept and apply technology for personal use, and voluntary adoption and usage are at one’s own risk. Therefore, the VAM distinguishes between positive factors (benefits) and negative factors (sacrifices); the subfactors of benefits include usefulness and enjoyment, while sacrifices involve both monetary and non-monetary aspects, consisting of technological characteristics and perceived costs [32].

Although the VAM was developed to explain initial adoption intentions, it can also be applied to studies on intention to reuse, as customers also consider value when making such ongoing decisions, as they do when deciding on their initial adoption. One difference, however, is that ongoing decision-making tends to be influenced by satisfaction [20]. Perceived benefits and sacrifices affect satisfaction in markedly different ways, particularly in the context of robotic services, which have fundamentally transformed the service delivery process. These technologies have not only enabled innovative service delivery through robots but also reshaped consumers’ perceptions and expectations [22]. Therefore, we must consider the roles of both perceived sacrifices and perceived benefits when studying satisfaction and intention to reuse [20].

Moreover, the VAM also enjoys a broad application in empirical research with regard to consumers’ technology adoption intentions, satisfaction, and intention to reuse, with various modifications and extensions depending on the research topic. In their study on smart farm restaurant systems, Joo and Hwang (2025) [33] replaced the technicality component of perceived sacrifices with food technophobia and added perceived risk. Lim et al. (2022) [24] incorporated perceived physical risk into the components of perceived sacrifices in their study on virtual tourism. Kim (2022) [23] conceptualized perceived sacrifices in terms of perceived bias and perceived risk in a study on a YouTube algorithmic recommendation service. Since the VAM was originally developed along with information and communication technology and includes variables that may be unsuitable for other technological contexts [33], the present study similarly modifies and extends the VAM to analyze satisfaction with RTSs.

The benefits consumers gain are not only useful but also interesting and enjoyable [32]. Here, perceived benefits consist of perceived usefulness and enjoyment. First, perceived usefulness refers to the extent to which service is perceived as beneficial in consumers’ everyday lives [34]. This key determinant can positively affect the continued use of technology-based services [35]. This study identifies perceived usefulness as the subjective perception of the extent to which an RTS improves the personal lives of consumers through its higher efficiency and convenient mobility. Second, enjoyment is the degree to which an activity is perceived as inherently enjoyable, including excitement and fun, regardless of the expected outcome [34]. Like perceived usefulness, enjoyment dramatically impacts technology satisfaction. In this study, enjoyment is the extent to which consumers experience infotainment and a pleasant environment while using an RTS (i.e., feeling positive emotions and not being bored).

Regarding the perceived benefits–satisfaction relationship, Nguyen-Phuoc et al. (2020) [36] ascertained the positive influence of the perceived usefulness of a mobility service application on satisfaction with that service. In a study by Weng et al. (2017) [37], the usefulness of a mobility service application positively affected satisfaction with that application. Similarly, Liu et al. (2018) [38] reported that the entertainment aspect of smart bicycle sharing services positively affected satisfaction with bicycle trips.

Perceived sacrifices involve monetary and non-monetary components [39], with the former referring to the cost of purchasing a product and the latter including time, effort, and dissatisfaction with the product price [30]. Service users tend to focus more on potential risks than on actual costs [32]. In particular, when using AI applications, risk factors such as data leakages, viruses, and other security and privacy issues are regarded as non-monetary sacrifices that users must additionally bear [21].

In this study, perceived sacrifices comprise perceived risk and technicality. First, when consumers decide that a certain purchasing goal cannot be achieved through behavior, they experience uncertainty and perceive risk [40]. From the consumer’s perspective, using innovative services not only involves additional costs, but also poses greater threats to personal privacy [41]. Although consumers are aware of the advantages of autonomous driving through media exposure, their psychological state remains in conflict because of their expectations and perceived risk. Fully AVs are expected to exhibit stronger safety and higher efficiency versus human-driven cars; however, consumers are often reluctant to relinquish control over the vehicle and remain concerned about software errors and safety [42]. This study considered perceived risk as the degree of perception for potential personal information leakages, service inadequacies, and unmet expectations by consumers when using an RTS. Second, technicality refers to the state in which a consumer is dissatisfied with the time and effort expended in using a technology and perceives using a new information system as difficult [30]. Using new technologies may cause psychological discomfort, such as conflict, frustration, depression, inconvenience, anxiety, tension, distress, and mental tiredness [20]. In this study, technicality is considered the degree to which users find an RTS complicated and hard to use.

In view of the relationship between perceived sacrifices and satisfaction, Seo and Lee (2021) [43] found that customers’ perceived risk toward service robots in restaurants negatively impacted their satisfaction with the restaurant. In a study on ride-sourcing services, Nguyen-Phuoc et al. (2021) [44] distinguished between perceived risks related to the ride-sourcing service application (performance risk, security risk, and dispute risk) and those related to the driver (property risk, security risk, and traffic risk), finding that both types of perceived risks negatively affected satisfaction with ride-sourcing services. Hong et al.’s (2025) [22] study on food delivery robot services confirmed the negative effect of technicality on satisfaction.

Collectively, we confirm the benefits offered to consumers (perceived usefulness and enjoyment, etc.) as pivotal factors enhancing customer satisfaction. Perceived sacrifices (perceived risk and technicality, etc.) by consumers are significant negative factors that reduce satisfaction. Hence, we propose the following hypotheses:

H1: 

The perceived benefits of RTSs have a significant effect on satisfaction.

H1-1: 

Perceived usefulness has a positive effect on satisfaction.

H1-2: 

Enjoyment has a positive effect on satisfaction.

H2: 

The perceived sacrifices of RTSs have a significant effect on satisfaction.

H2-1: 

Perceived risk has a negative effect on satisfaction.

H2-2: 

Technicality has a negative effect on satisfaction.

2.2. Technology Readiness Index

Since individuals’ responses to technology vary depending on their personality traits, Parasuraman (2000) [26] constructed the TRI to explain the intention and ability of an individual to adopt as well as use technology [45]. Technology readiness is the tendency of a person to accept and use new technologies to achieve personal or work-related goals [26]. It is a psychological factor that influences an individual’s likelihood of using new technologies, and it includes four aspects: enablers like innovativeness and optimism and inhibitors such as discomfort and insecurity [25].

Technology readiness determines whether new products and services are successful by affecting market potential and the user base [12]. Consumers’ characteristics impact their satisfaction with technology and future behavioral intentions [46,47]. Of these, their technology readiness is a pivotal characteristic explaining their attitudes toward technology and their acceptance of it [26,46]. Consumers’ level of technology readiness may make them have varying expectations, possibly impacting their service perceptions. For example, highly optimistic consumers usually emphasize the positive aspects of technology, making them more likely to be satisfied when using it, whereas highly discomfortable consumers usually feel overwhelmed during use, which may reduce their satisfaction [48].

According to research on electronic services [49], self-service [50], restaurant self-service [51], and health services [47], technology readiness remarkably predicts user satisfaction and intention to reuse. Therefore, the study explores the relationship between the above four attributes of technology readiness and satisfaction among RTS consumers.

Optimism is related to a positive perspective of technology, covering customers’ beliefs about its control, flexibility, convenience, and efficiency [26]. Optimism is the degree to which users hold positive expectations toward new technologies or innovations, and highly optimistic users tend to accept, explore, and use new technologies [52]. The study considers optimism as a positive perspective enabling consumers to control and freely use new technologies and produce efficient outcomes.

Innovativeness is the tendency to become a technology pioneer [26]. Highly innovative individuals prefer to adopt new technologies and exhibit innovativeness, driven by their high intrinsic motivation [53]. Such consumers show a larger likelihood of perceiving the functionality of new technologies more positively [54]. This study considers innovativeness as the tendency to be an early adopter of new technologies and to seek to become a technology pioneer.

Discomfort refers to the feeling of having little control and being stressed by technology [26]. People who experience high discomfort often see using technology as unpleasant and difficult [48]. In this study, discomfort is considered as how much consumers feel they cannot control new technology and feel overwhelmed by it.

Insecurity means having a lack of trust in technology and questioning whether it works correctly [55]. It is defined as arising from a lack of confidence in the technology and its capability to function properly [26]. Individuals feeling insecure about new technology tend to be more anxious and hold skeptical attitudes toward its practicality [54]. In this study, insecurity refers to a skeptical attitude toward technology and its capability to function properly.

A bunch of research has looked at how the Technology Readiness Index impacts customer satisfaction. For example, Lin and Hsieh (2007) [27] checked out self-service systems and found that all four technology readiness factors could positively affect satisfaction. Then, in 2009, Chen and Chen [56] studied self-service banking and discovered the positive effect of optimism and innovativeness on satisfaction. The same thing popped up in Hemdi et al.’s 2016 airport kiosk study [53]: optimism and innovativeness helped satisfaction, while discomfort hurt it. Later on, Chen et al.’s 2009 SST research [57] backed this up, showing that optimism and innovativeness improved satisfaction. Begum et al.’s (2023) [58] meta-analysis of technology readiness found that optimism and innovativeness positively affected satisfaction, while discomfort and insecurity exerted a negative effect. Based on all this, we are testing these ideas:

H3: 

Motivators of the Technology Readiness Index have a significant effect on satisfaction with RTSs.

H3-1: 

Optimism has a positive effect on satisfaction.

H3-2: 

Innovativeness has a positive effect on satisfaction.

H4: 

Inhibitors of the Technology Readiness Index have a significant effect on satisfaction with RTSs.

H4-1: 

Discomfort has a negative effect on satisfaction.

H4-2: 

Insecurity has a negative effect on satisfaction.

2.3. Satisfaction, Intention to Reuse, and e-WOM

Satisfaction refers to the degree of positive or negative emotions a consumer experiences after using a product or service [59]. Howard and Sheth (1969) [60] defined customer satisfaction as a cognitive evaluation of whether a consumer has received adequate compensation for the cost or sacrifice incurred. In the case of a product or service exceeding customer expectations, customer satisfaction increases; conversely, if performance falls short, customer satisfaction decreases [61]. Accordingly, our study defines satisfaction as a consumer’s evaluation of an RTS they feel and experience during use.

Intention to reuse refers to a customer’s intention to use the same service again [62]. As a form of loyalty, it can encourage consumers to use a product or service more frequently, which may grow sales [63]. Accordingly, our study defines intention to reuse as a consumer’s intention to reuse an RTS after their initial usage.

WOM is defined as any informal communication between consumers about ownership, usage, and product and service characteristics and sellers. eWOM, which includes online reviews, recommendations, and opinions, has grown in importance as new technologies are on the rise [64]. eWOM is positive or negative information pertaining to a product or company shared by potential, current, or former customers online; this includes not only consumers’ experiences with products and brands but also their experience using information systems [65]. eWOM is very different from conventional WOM and eWOM in terms of the scope of influence and speed of information diffusion [64]. eWOM spreads through platforms such as email, blogs, and online communities and primarily occurs among strangers, thereby having a much broader network and faster dissemination speed than traditional WOM [66]. The present study defines eWOM as a consumer’s intention to share positive personal experiences about an RTS with other users online after using the service.

A consumer’s intention to repurchase a product or continue to use a service mainly depends on satisfaction with past usage or purchase experiences. Higher satisfaction with the experience leads to a greater likelihood that consumers will reuse the service and recommend it to others [67,68,69]. Therefore, this study analyzes the association between satisfaction with RTSs, intention to reuse, and eWOM.

Among the various relevant studies, Wang et al. (2020) [70] ascertained the positive effect of satisfaction on intention to reuse urban rail transit services, Vu et al. (2024) [71] ascertained the positive effect of satisfaction with ride-hailing taxi services on intention to reuse, and Rosell and Allen (2020) [72] demonstrated a positive effect of satisfaction with driverless buses on intention to reuse. Among studies on the effect of satisfaction on eWOM, Hamzah et al. (2023) [73] found that satisfaction with public transport services positively affected WOM. In a study by Ruiz-Alba et al. (2022) [74], satisfaction with Uber services positively affected eWOM. Antwi et al. (2021) [75] studied airport self-service, finding that satisfaction with self-service positively affected the eWOM of the airport. Chen and Girish (2023) [76] studied airport service robots and discovered that customer satisfaction positively affected WOM. All these promote the proposal of Hypotheses 5 and 6:

H5: 

Satisfaction with RTSs has a positive effect on intention to reuse.

H6: 

Satisfaction with RTSs has a positive effect on eWOM.

2.4. Uncertainty Avoidance

Uncertainty avoidance, as a cultural dimension proposed by Hofstede (1980) [77], is defined as the degree to which members of the same culture or nation perceive risk in uncertain situations. Generally, people tend to believe that higher risk indicates lower likelihood of a successful transaction [78]. The study considers uncertainty avoidance as the tendency to perceive risk when using an RTS and avoid using such services.

Uncertainty avoidance can reflect not only cultural differences at the national level, but also individual differences within the same cultural background [79]. Since individuals within a country may exhibit varying levels of uncertainty avoidance, its effects on people’s attitudes and behaviors [80] can be examined at the individual level within geographic markets [81]. Using a cultural measurement standard to understand individuals’ responses to new technologies is thus important [82].

In particular, uncertainty avoidance regulates the relationship between attitudes toward new technological products and services (including satisfaction) and behavioral intentions, as evidenced by existing studies [83]. Individuals possessing low uncertainty avoidance tend to respond more flexibly to uncertain situations, accept diversity, and exhibit lower stress and anxiety. Conversely, those with high uncertainty avoidance adapt to respond to structured situations and experience higher stress and anxiety in confronted with unfamiliar and ambiguous circumstances [84]. In their study on cruise passengers, Sanz Blas and Carvajal-Trujillo (2014) [85] found that in collectivist cultures presenting high uncertainty avoidance, satisfaction was more strongly related to future behavioral intentions. In other words, individuals with high uncertainty avoidance tend to have a lower preference for new experiences and a stronger tendency to avoid risk; therefore, higher satisfaction leads to a stronger tendency to revisit a travel destination or recommend it to others. Moreover, eWOM about service providers is more frequently posted by individuals with low uncertainty avoidance [86]. Therefore, this study aims to elucidate how uncertainty avoidance moderates the relationships among satisfaction, eWOM, and intention to reuse.

Rohden and Matos (2022) [87] explored online shopping and found the moderating mechanism of uncertainty avoidance against the link between service recovery satisfaction and word-of-mouth (WOM). They observed that for people with high uncertainty avoidance, satisfaction with service recovery had a stronger effect on negative WOM than for those with low uncertainty avoidance. Their findings indicate that consumers with higher uncertainty avoidance tend to show a stronger relationship between post-recovery satisfaction and negative feedback than consumers with lower levels of uncertainty avoidance. Similarly, Hassna et al. (2023) [88] investigated B2C e-commerce and ascertained that the moderating mechanism of uncertainty avoidance affects the connection between customer satisfaction and the intention to repurchase. They found that higher uncertainty avoidance strengthened this relationship. Overall, previous research shows that uncertainty avoidance can change how satisfaction relates to both repurchase intention and eWOM. Based on this, the following hypotheses are suggested: Uncertainty avoidance has a moderating role in consumer behavior.

H7: 

Uncertainty avoidance moderates the relationships between satisfaction with RTSs, eWOM, and intention to reuse.

H7-1: 

Uncertainty avoidance moderates the relationship between satisfaction and intention to reuse.

H7-2: 

Uncertainty avoidance moderates the relationship between satisfaction and eWOM.

2.5. Research Model

This study focused on testing the effects of VAM-related variables and technology readiness on satisfaction with RTSs, the effect of satisfaction on intention to reuse and eWOM, and the moderating effect of uncertainty avoidance. Figure 1 explains the research model.

3. Research Methods

3.1. Survey Instrument

This study performed an online survey targeting Chinese consumers with experience using an RTS through the professional survey platform WENJUANXING (https://www.wjx.cn). The questionnaire introduced the research purpose and included textual explanations and visual materials such as images to ensure that respondents fully understood RTSs. This study collected survey data using a voluntary sampling method, which involved respondents within the target population participating voluntarily [89]. Although voluntary sampling cannot represent the entire population because of selection bias, it is a valid and suitable methodological alternative for researchers when random sampling is difficult in practice [90].

The questionnaire consisted of 41 items, divided into nine sections: (1) Perceived usefulness and enjoyment as dimensions of perceived benefits, (2) Perceived risk and technicality of perceived sacrifices, (3) Optimism and innovativeness of technology readiness index, (4) Discomfort and insecurity of technology readiness index, (5) Satisfaction, (6) Intention to reuse, (7) Electronic word-of-mouth (eWOM), (8) Uncertainty avoidance, and (9) Demographic characteristics of the respondents. Each structural variable was evaluated through multiple measurement items, all of which were modified and supplemented based on the literature to ensure scientific rigor and validity. Most existing scales primarily focus on transportation services, shared services, or unmanned services, making them not fully applicable to the context of robotaxi services. Therefore, this study made moderate adjustments to the wording of the items while preserving the core conceptual meanings from previous research in order to ensure consistency with the robotaxi service context. Furthermore, the research team requested a professional translator to review the measurement items to guarantee translation accuracy. A 5-point Likert scale was adopted for measurement (1 = strongly disagree, 5 = strongly agree).

Table 1. Questionnaire items.

Variable and Item		Reference(s)
Perceived usefulness (PU)	Using RTSs improves travel efficiency.	Adapted from [30,91,92,93]
	Using RTSs makes travel more convenient.
	Overall, RTSs are of great use in my daily life.
Enjoyment (ENJ)	Using RTSs is very enjoyable.
	I can enjoy infotainment (information + entertainment) and a pleasant environment inside a robotaxi.
	Using RTSs does not feel boring.
Perceived risk (PR)	I worry that using RTSs could put my personal information at risk.	Adapted from [30,94,95,96]
	I worry that I might not receive a proper service when I use RTSs.
	I worry that the service I get from RTSs may fall short of what I expect.
Technicality (TECH)	RTSs are difficult to use.
	It is difficult to learn how to use RTSs.
	It is difficult to use RTSs proficiently.
Optimism (OPT)	Technology allows me to better control my daily life.	Adapted from [25,45,97,98]
	New technology contributes to improving my quality of life.
	New technology enhances my productivity and creativity.
Innovativeness (INN)	I am generally one of the first among the people around me to use new technology.
	I can usually understand new high-tech products and services on my own.
	I try to keep up with the newest technology developments.
Discomfort (DIS)	I worry that technology tends to fail at the worst times.
	I tend to become flustered when I encounter problems using advanced technology products or services in front of others.
	When purchasing advanced technology products or services, I prefer basic models over those with many extra features.
Insecurity (INS)	Excessive technology is distracting enough to lower my concentration.
	Technology reduces personal interaction, which lowers the quality of human relationships.
	I believe it is difficult for technology and technology-based services to respond effectively in emergency situations.
Satisfaction (SAT)	Overall, I am very satisfied with RTSs.	Adapted from [70,76]
	RTSs meet my expectations.
	RTSs have many advantages.
Intention to reuse (IR)	I will use RTSs as much as possible.	Adapted from [70,99]
	I will choose to ride a robotaxi again in the same circumstances.
	My interest in RTSs will rise in the future.
Electronic word-of-mouth (eWOM)	I often spread positive word-of-mouth about RTSs on the internet.	Adapted from [74,76]
	I talk about the benefits of robotaxis on social media.
	I am willing to recommend RTSs to others.
	I am willing to share my positive experiences of RTSs with others.
Uncertainty avoidance (UA)	I prefer planned situations over unplanned ones.	Adapted from [100,101]
	I get stressed easily when outcomes are unpredictable.
	I prefer stability over change.

lists the questionnaire items and references.

3.2. Data Collection

To ensure the study met rigorous statistical standards, the minimum required sample size was calculated using G*Power (version 3.1.9.7) based on the method proposed by Cohen (1988) [102], resulting in a minimum of 127 responses. Chinese consumers aged 20 and over were surveyed from 9 April to 23 April 2025. Among the 500 questionnaires distributed, 425 were considered valid after excluding invalid responses; thus, the response rate reached 85%. All the respondents were in their 20s–40s because older adults are generally less familiar with hailing robotaxis using smart devices (

Table 2. Characteristics of the respondents.

Characteristic	Indicator	Frequency	%
Gender	Male	200	47.1
	Female	225	52.9
Age (years)	20–29	183	43.1
	30–39	193	45.4
	40–49	49	11.5
Education level	High school or below	9	2.1
	University or college graduate	358	84.2
	Postgraduate or above	58	13.6
Monthly income (USD)	<412	17	4.0
	412–824	56	13.2
	825–1236	112	26.4
	>1237	240	56.5

) [103]. In particular, older adults perceive the risks associated with RTSs to be relatively high [16].

3.3. Data Analysis

This study conducted statistical analyses using SPSS 26.0 and AMOS 26.0. In this study, Harman’s single-factor test was conducted to assess common method bias (CMB). The results showed that the first factor accounted for 26.60% of the variance, which is well below the threshold of 50%, indicating that common method bias is not a major concern [104]. The internal consistency and convergent validity among the measurement items were verified by conducting reliability analysis coupled with confirmatory factor analysis, and discriminant validity was verified through correlation analysis. Moreover, a two-stage approach was applied to evaluate the structural model. The examination of variable relationships was followed by a multi-group test to analyze the moderating effect of uncertainty avoidance.

4. Results

4.1. Confirmatory Factor Analysis

Cronbach’s α was analyzed for testing whether these measurement items were reliable, and all the variables showed values above 0.7, ensuring internal consistency and reliability [105] (

Table 3. Reliability and validity results.

Variable	Factor	Standardized Factor Loadings	Cronbach’s α	AVE	CR
PU	PU1	0.758	0.777	0.541	0.779
	PU2	0.688
	PU3	0.759
ENJ	ENJ1	0.705	0.749	0.502	0.751
	ENJ2	0.676
	ENJ3	0.742
PR	PR1	0.757	0.856	0.671	0.859
	PR2	0.867
	PR3	0.829
TECH	TECH1	0.712	0.775	0.536	0.776
	TECH2	0.770
	TECH3	0.713
OPT	OPT1	0.675	0.760	0.524	0.767
	OPT2	0.708
	OPT3	0.784
INN	INN1	0.814	0.746	0.502	0.749
	INN2	0.637
	INN3	0.660
DIS	DIS1	0.627	0.776	0.550	0.783
	DIS2	0.861
	DIS3	0.718
INS	INS1	0.761	0.789	0.558	0.791
	INS2	0.695
	INS3	0.783
SAT	SAT1	0.799	0.774	0.544	0.780
	SAT2	0.648
	SAT3	0.757
RI	RI1	0.814	0.766	0.530	0.770
	RI2	0.647
	RI3	0.713
eWOM	eWOM1	0.776	0.806	0.516	0.809
	eWOM2	0.753
	eWOM3	0.710
	eWOM4	0.624
UA	UA1	0.656	0.748	0.515	0.759
	UA2	0.672
	UA3	0.814
Chi-square = 763.979, df = 563, χ2/df = 1.357; p < 0.001, IFI = 0.969; TLI = 0.962; CFI = 0.968; RMSEA = 0.029

). In addition, according to the confirmatory factor analysis results, the model fit indices were χ² = 763.979, df = 563, p < 0.000, χ²/df = 1.357, IFI = 0.969, TLI = 0.962, CFI = 0.968, and RMSEA = 0.029, which indicated an overall adequate model fit [106]. Examination of standardized factor loadings assisted in evaluating the variables’ convergent validity, all of which went beyond the threshold of 0.6. The average variance extracted (AVE) values changed in the range of 0.502–0.671, all above the threshold of 0.5, and the composite reliability (CR) values changed in the range of 0.749–0.859, all exceeding the threshold of 0.7, confirming that convergent validity was maintained [107].

Discriminant validity was analyzed based on the criteria of Fornell and Larcker (1981) [108]. The AVE values had the square root > the correlation coefficients between the constructs (

Table 4. Correlation matrix of the variables.

	PU	ENJ	PR	TECH	OPT	INN	DIS	INS	SAT	RI	eWOM	UA
PU	0.736
ENJ	0.444	0.709
PR	−0.492	−0.409	0.819
TECH	−0.436	−0.353	0.550	0.732
OPT	0.366	0.669	−0.221	−0.258	0.724
INN	0.493	0.484	−0.438	−0.438	0.454	0.709
DIS	−0.435	−0.296	0.600	0.615	−0.198	−0.405	0.742
INS	−0.339	−0.329	0.615	0.558	−0.165	−0.379	0.638	0.747
SAT	0.486	0.668	−0.446	−0.313	0.706	0.480	−0.368	−0.294	0.738
RI	0.379	0.554	−0.280	−0.260	0.700	0.525	−0.323	−0.262	0.697	0.728
eWOM	0.547	0.477	−0.406	−0.350	0.451	0.688	−0.356	−0.295	0.569	0.638	0.718
UA	−0.128	0.066	0.211	0.175	−0.055	−0.245	0.365	0.315	−0.040	−0.076	−0.183	0.718

Note: Diagonal elements are the square root of the AVE.

), confirming the secured discriminant validity.

4.2. Hypothesis Testing

A structural equation model was analyzed for verifying the hypotheses (

Table 5. Results of the hypothesis testing.

Hypothesis	β	S.E.	C.R.	p-Value	Result
H1-1: PU → SAT	0.123	0.056	2.095	0.036 *	Accepted
H1-2: ENJ → SAT	0.178	0.082	2.297	0.022 *	Accepted
H2-1: PR → SAT	−0.137	0.042	−2.105	0.035 *	Accepted
H2-2: TECH → SAT	0.114	0.059	1.742	0.082	Rejected
H3-1: OPT → SAT	0.499	0.102	6.296	0.000 ***	Accepted
H3-2: INN → SAT	0.181	0.047	2.911	0.004 **	Accepted
H4-1: DIS → SAT	−0.138	0.059	−1.969	0.049 *	Accepted
H4-2: INS → SAT	0.028	0.045	0.406	0.685	Rejected
H5: SAT → RI	0.792	0.080	12.664	0.000 ***	Accepted
H6: SAT → eWOM	0.690	0.078	10.912	0.000 ***	Accepted
Chi-square = 737.973, df = 489, χ2/df = 1.509; p < 0.001, IFI = 0.958; TLI = 0.951; CFI = 0.958; RMSEA = 0.035

* p < 0.05; ** p < 0.01; *** p < 0.001.

), with the model fit indices of χ² = 737.973, df = 489, p < 0.001, χ²/df = 1.509, IFI = 0.958, TLI = 0.951, CFI = 0.958, RMSEA = 0.035, indicating an overall adequate fit.

Perceived usefulness (β = 0.123, p < 0.05) and enjoyment (β = 0.178, p < 0.05) both remarkably and positively affected satisfaction, supporting H1-1 and H1-2, indicating that higher levels of perceived usefulness and enjoyment of RTSs among consumers lead to greater user satisfaction. Perceived risk (β = −0.137, p < 0.05) greatly and negatively affected satisfaction, supporting H2-1. Technicality (β = 0.114, p > 0.05) failed to significantly affect satisfaction with RTSs, rejecting H2-2, demonstrating that when consumers evaluate satisfaction with RTSs, they respond sensitively to the burden of risk; however, the perception of technical complexity does not affect satisfaction significantly.

According to the results of testing H3, optimism (β = 0.499, p < 0.001) and innovativeness (β = 0.181, p < 0.01) both exerted an obvious positive effect on satisfaction, and thus H3-1 and H3-2 were accepted. This indicates that consumers who have positive expectations from or an open attitude toward new technology tend to have higher satisfaction with RTSs. Discomfort (β = −0.138, p < 0.05) exerted a significant negative effect on satisfaction, and thus H4-1 was accepted. Consumers with a high tendency toward technical discomfort may have lower satisfaction with RTSs. Insecurity (β = 0.028, p > 0.05) did not significantly affect satisfaction, rejecting H4-2. Anxiety about technology did not directly affect satisfaction with RTSs.

Next, according to the results of testing H5 and H6, satisfaction with RTSs (β = 0.792, p < 0.001) dramatically and positively impacted intention to reuse, supporting the acceptance of H5. Moreover, satisfaction with RTSs (β = 0.690, p < 0.001) also dramatically and positively impacted eWOM, supporting the acceptance of H6. Managing customer satisfaction is thus a key factor in an RTS being successfully diffused. Not only does it induce reuse, but it can also attract new customers through voluntary eWOM.

4.3. Moderation Analysis

This study used multi-group analysis to verify whether uncertainty avoidance moderated the relationships among satisfaction, eWOM, and intention to reuse. Following the group classification method proposed by Cheng et al. (2022) [109], the mean value of uncertainty avoidance (3.783) was taken into account to divide groups into low and high groups, followed by a chi-square difference test.

The relationship between satisfaction and intention to reuse differed significantly between the constrained and unconstrained models (

Table 6. Moderating effect of uncertainty avoidance.

Hypothesis	Δχ2, Δdf	Low (N = 193)		High (N = 232)		Result
		β	C.R.	β	C.R.
H7-1: SAT → RI	Δχ2 (df = 1) = 4.011 *	0.704 ***	7.427	0.856 ***	10.267	Accepted
H7-2: SAT → eWOM	Δχ2 (df = 1) = 3.006	0.627 ***	6.395	0.728 ***	8.582	Rejected

* p < 0.05; *** p < 0.001.

) (Δχ² = 4.011, Δdf = 1), and thus, H7-1 was accepted. Satisfaction remarkably and positively impacted intention to reuse in both groups, and such impact was stronger for those possessing higher uncertainty avoidance. Specifically, consumers possessing high uncertainty avoidance are cautious and conservative in adopting new products, and their adoption speed is also slower than those possessing low uncertainty avoidance. As RTSs are innovative services based on autonomous driving technology, a certain level of uncertainty and ambiguity is expected during their initial use. Accordingly, users with high uncertainty avoidance may be hesitant in their initial intention to use RTSs and have a wary attitude toward using such services; however, once their satisfaction with RTSs increases, their motivation to avoid uncertainty leads to a stronger tendency to continue using such services because they seek to avoid the uncertainty and risks involved in exploring or switching to new alternatives. This ultimately shows that users’ higher uncertainty avoidance triggers a stronger relationship. By contrast, users with low uncertainty avoidance are more willing to accept changes and ambiguity. They also have a more flexible attitude toward the adoption of RTSs and a stronger intention to accept new technologies and services. Therefore, even if these consumers show satisfaction with the service, they will also actively explore other alternatives that offer better features and more innovative experiences. Finally, the constrained and unconstrained models did not present an obvious difference in the relationship between satisfaction and eWOM (Δχ² = 3.006, Δdf = 1), and thus, H7-2 was rejected.

5. Discussion and Conclusions

In China, RTSs based on autonomous driving technology have recently entered the commercialization stage and are developing rapidly. More and more cities are conducting RTS pilots. Robotaxis are attracting attention as an alternative that can overcome the limitations of traditional taxi and transportation methods by alleviating traffic congestion, reducing transportation costs, and saving energy. Analyzing the relationships between user satisfaction, intention to reuse, and eWOM is thus a key task for the establishment and market expansion of RTSs to ensure their business success. Therefore, this study empirically tested our research hypotheses based on a sample dataset of 425 responses collected in China. All these provide theoretical and practical insights as follows.

5.1. Theoretical Implications

First, most studies have used the VAM to analyze perceived value and acceptance intention through perceived usefulness, enjoyment, perceived risk, and technicality. To extend the theoretical research, this study partially modified and extended the VAM by adding satisfaction to the investigation into the association between intention to reuse and eWOM. This model emphasizes the importance of satisfaction, theoretically contributing to follow-up research in related fields.

Second, individuals’ technology readiness is a key characteristic explaining their attitudes toward and acceptance of technology [26,46]. Therefore, the study analyzed the impact of the subfactors of technology readiness—optimism, innovativeness, discomfort, and insecurity—on satisfaction with RTSs, finding the obvious impact of the former three factors. These results, which suggest that those three subfactors of technology readiness are important variables for explaining customer satisfaction with RTSs, strengthen the theoretical foundation for future research on satisfaction with such services.

Third, according to our findings, satisfaction with RTSs affected both intention to reuse and eWOM, and uncertainty avoidance regulated the relevance of satisfaction to intention to reuse. This moderating effect reveals the mechanism underlying technology use behavior and theoretically benefits future research.

To address the gaps in existing research on the acceptance of robotaxi services, this study introduces the concept of perceived sacrifice to capture the potential negative aspects that users may encounter when adopting new technologies. In contrast to most prior studies that have primarily focused on the functional aspects of robotaxi services (RTSs), this research places greater emphasis on individual differences and psychological characteristics. Specifically, it examines how personal dispositions—such as technology readiness—influence service perceptions and satisfaction. Furthermore, this study proposes a theoretical framework to explain the interaction between cultural values and individual traits, highlighting how service experiences can be perceived differently depending on consumer characteristics.

Finally, this study offers important insights into user satisfaction, intention to reuse, and electronic word-of-mouth (eWOM) regarding RTSs. However, the applicability of the research model may vary across different cultural and infrastructural contexts. Specifically, cultural dimensions, such as uncertainty avoidance, individualism vs. collectivism, and attitudes toward automation, as well as infrastructural factors like transportation systems, urban density, regulatory environments, and technological maturity, may influence service perceptions. Therefore, when applying this model in regions with differing cultural and infrastructural conditions—such as Europe or North America—adaptations may be necessary.

5.2. Practical Implications

The results of the study significantly guide RTS providers. First, perceived benefits (perceived usefulness and enjoyment, etc.) remarkably and positively affected satisfaction with RTSs, conforming to the findings of the studies on mobility service applications and smart bicycle sharing services by Nguyen-Phuoc et al. (2020) [36] and Liu et al. (2018) [38], respectively. RTSs should strive to optimize the number and location of vehicles to assign rides to consumers more quickly and ensure faster arrival at pick-up and drop-off points. Moreover, the results indicate that enjoyment (β = 0.178, p < 0.05) has a stronger impact on satisfaction than perceived usefulness (β = 0.123, p < 0.05). This can be attributed to the fact that robotaxis often adopt more conservative driving strategies compared to traditional taxis, resulting in longer travel times. As a result, many consumers do not view robotaxis as a practical mode of daily transportation, but rather as an experience-oriented choice. Therefore, in addition to maximizing functional utility, such as time savings and travel efficiency, robotaxis must enhance emotional experiences during the ride, such as enjoyment. Therefore, service providers can enhance users’ interaction and emotional engagement with RTSs by implementing AI-based conversational voice interfaces—personified to communicate like humans—that are capable of real-time navigation guidance and natural language communication, along with infotainment features within the vehicle, an intuitive app, and a comfortable vehicle environment. Such approaches would promote a more enjoyable consumer experience by improving the efficiency of RTSs. Prioritizing strategies that enhance enjoyment is more effective in improving user satisfaction than solely focusing on increasing perceived usefulness.

Second, perceived risk, one of the variables under perceived sacrifices, remarkably and negatively affected satisfaction with RTSs, conforming to the findings of Nguyen-Phuoc et al. (2021) [44] on ride-sourcing services. Consumers may perceive various risks in using robotaxis, including technical errors, security issues, and unpredictability. To address these concerns, trust-based design and communication strategies, such as real-time route tracking, accident prevention and response systems, and transparent guidance on personal data protection policies, are necessary. On the contrary, technicality failed to well affect satisfaction with RTSs, which conforms to the findings of Kim and Kim (2021b) [110] on PropTech service platforms. In the current digital era, users have high familiarity with and deep acceptance of autonomous driving technology and AI-based services, and thus technical complexity tends not to function as a major dissatisfaction factor in service evaluation. In other words, users now tend to evaluate AI-based services based on the actual performance capability of the AI, rather than focusing directly on the technology itself [111]. In this context, technological complexity is no longer perceived as a differentiating factor, but rather as an inherent attribute of the technology itself, thereby diminishing its impact on user satisfaction. Instead, users tend to focus more on whether the service meets their expectations in terms of safety and responsiveness, which may limit the direct influence of technological features on satisfaction. As such, when users are already familiar with similar technologies, even a high level of complexity may have a limited effect on their overall satisfaction.

Third, among the four technology readiness factors, optimism and innovativeness positively affected customer satisfaction with RTSs. By contrast, discomfort negatively affected satisfaction, while insecurity had no significant effect. All these conform to the results of Hemdi et al. (2016) [53] on airport self-service. Notably, among all the variables influencing satisfaction in the research model, optimism (β = 0.499, p < 0.001) emerged as the most influential factor. This suggests that users’ optimistic mindset may serve as a key driver in shaping their evaluations of RTS services. This finding implies that relevant companies should actively foster user optimism in order to maximize satisfaction. Highly optimistic and innovative consumers tend to approach new technologies such as robotaxis with anticipation and curiosity and demonstrate a stronger tolerance to certain shortcomings or errors, which increases their overall satisfaction with the service. Accordingly, effective strategies include offering these consumers opportunities and beta services to experience new RTS features in advance and encouraging their active adoption of the technology, such as by providing personalized notifications through the app. Actively communicating with these consumers through the app could not only lower the inconvenience and shortcomings incurred when using RTSs but also offer practical insights to help develop new features and enhance service quality. Furthermore, such continuous interaction could strengthen relationships with consumers, increase their intention to reuse and engage in WOM, and attract new users.

Moreover, psychological discomfort and emotional resistance experienced during technology use have a more direct effect on service evaluation than technical insecurity. Therefore, strategies could aim to overcome users’ psychological barriers by providing user guides, designing intuitive interfaces, and integrating customer service channels. These strategies could improve users’ experience of using RTSs and reduce the likelihood of operational errors during usage. The introduction of legal and institutional management standards by the government in China [112] has led consumers to focus more on the functionality and convenience of RTSs than on their perceived anxiety about RTSs. As a result, the influence of insecurity on satisfaction with RTSs may be relatively limited. In other words, when institutional and regulatory safeguards are strong, the direct impact of insecurity on user satisfaction may diminish. In such contexts, insecurity is less likely to function as an independent determinant of satisfaction and more likely to influence intermediate factors such as trust and perceived service quality. Therefore, in well-regulated environments, investments aimed solely at reducing user anxiety may have limited effects on enhancing satisfaction, suggesting that companies should allocate their resources more strategically to other areas.

Fourth, satisfaction with RTSs substantially and positively affected intention to reuse and eWOM, which accords with the findings of Chen and Girish (2023) [76] on airport robot services and Wang et al. (2020) [70] on urban rail transit services. The statistical results indicate that satisfaction has a significant impact on intention to reuse (β = 0.792, p < 0.001) and electronic word-of-mouth (eWOM (β = 0.690, p < 0.001), suggesting that improvements in user satisfaction can strongly promote both customers’ intention to reuse the service and their engagement in positive eWOM. This suggests that consumers’ satisfaction with the service can lead not only to reuse but also to voluntary recommendation to other potential users. Intention to reuse is formed when consumers consistently have a positive perception of the overall service. Functional factors such as punctuality, ease of access, and price transparency increase satisfaction, and companies can encourage repeat use by considering these factors when formulating strategies. In particular, customer reviews, social media posts, and ratings directly influence new users’ evaluations in non-face-to-face services such as shared mobility [113], and thus satisfaction-based WOM can be an important strategic asset for service diffusion. Therefore, marketing strategies such as referral reward programs for satisfied customers, review-writing events, and the collection of points based on reviews can encourage customer loyalty as well as attract new customers. However, companies should be careful not to rely excessively on short-term promotions, as customers who always expect discounts or rewards may dilute the brand’s long-term value. By contrast, prompt feedback, effective conflict resolution, and responses considering customer safety as a top priority can increase satisfaction and promote positive WOM.

Finally, uncertainty avoidance could positively regulate the association between customer satisfaction with RTSs and intention to reuse. Consumers possessing a high tendency toward uncertainty avoidance showed a stronger influence of satisfaction on intention to reuse. For individuals who feel uncomfortable with uncertain situations and have a strong tendency to avoid risks, a positive experience with an RTS offers a great deal of stability and trust. In other words, when satisfaction is established, their emotional attachment and loyalty to the service are strengthened more quickly, resulting in a higher likelihood of reusing such services. On the contrary, consumers with low uncertainty avoidance tend to approach new technologies experimentally or flexibly and have other influencing factors (economic feasibility, etc.), aside from satisfaction. Developing differentiated user retention strategies based on consumers’ varying levels of uncertainty avoidance can maximize the effectiveness of satisfaction-based interventions. Therefore, for consumers possessing high uncertainty avoidance, it is necessary to implement strategies that steadily maintain and strengthen satisfaction with RTSs by providing clear information, guaranteeing service quality, and ensuring predictability and consistency throughout the service process. For consumers with low uncertainty avoidance, strategies such as differentiating RTS features, offering selectable options, designing personalized experiences, and securing price competitiveness may be more effective.

5.3. Limitations and Future Research Directions

Certain limitations of the study will be highlighted. First, the sample primarily encompassed respondents in their 20s and 30s, with those in their 40s accounting for only 11.5% of the total. With regard to education level, 97.8% of the respondents held a college degree or higher, making the study incapable of reflecting the opinions of more respondents. As a result, the study may not have sufficiently reflected differences in satisfaction and behavioral intentions toward RTSs across age groups. This study did not account for differences based on consumers’ city of residence. Therefore, future research should incorporate geographical factors to further explore potential regional variations. Future research is suggested to strengthen these findings’ external validity by securing sufficiently diverse samples that reflect a wider range of demographic characteristics.

Second, this study analyzed the determinants of satisfaction utilizing the VAM and TRI, but these theoretical models have limitations in fully explaining individuals’ attitude formation and cognitive acceptance processes. In particular, variables such as consumers’ perceptions of RTSs, usage experiences, and social influences were not considered. Future research could additionally apply cognition-based theories (the theory of planned behavior, etc.) to explore the factors affecting satisfaction with RTSs in a multidimensional manner.

Prior to conducting statistical analyses, sampling errors may exist in the raw data due to the nature of the sample selection, which could affect the accuracy and generalizability of the results. Future research could reduce such uncertainty by increasing the sample size and adopting a rigorous probability sampling design.

This study employed a cross-sectional survey design in which all variables were measured at a single point in time. While this approach is suitable for identifying associations among constructs, it has inherent limitations in establishing causal relationships—particularly when examining post-adoption behaviors, such as intention to reuse and electronic word-of-mouth (eWOM). Therefore, the relationships identified in this study should be interpreted as correlational rather than strictly causal. Future research could adopt longitudinal panel designs or field experiments to more rigorously examine the temporal dynamics and causal mechanisms underlying RTS adoption and continued use.