What will be the possible barriers to consumers’ adoption of smart home services?

Telecommunications Policy xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Telecommunications Policy journal homepage: www.elsevier.com/locate/telpol

What will be the possible barriers to consumers’ adoption of smart home services? Areum Honga, Changi Nama, Seongcheol Kimb,∗ a

School of Business and Technology Management, College of Business, Korea Advanced Institute of Science and Technology, N22, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, Republic of Korea School of Media and Communication, Korea University, 145, Anam-dong, Seongbuk-gu, Seoul, 02841, Republic of Korea

b

ARTICLE INFO

ABSTRACT

Keywords: Smart home Internet of things Barriers Resistance Perceived risk Uncertainty

Recently, smart home services have come to the forefront as part of the growing market for the “Internet of Things.” Since these smart home services were introduced, they have been expected to grow rapidly. However, contrary to optimistic expectations for future market growth, the smart home market has appeared to hit a roadblock and remains at an early market stage. This study attempts to identify the possible barriers that consumers perceive when they are introduced to smart home services. Based on the resistance theory and perceived risk model, we investigate the relationship between perceived risk and resistance to smart home services, using technological uncertainty and service intangibility as the antecedents of perceived risk. Dividing perceived risk into four dimensions—performance risk, financial risk, privacy risk, and psychological risk—the empirical results show that these four risk types are affected by technology uncertainty and service intangibility, and the perceived risks, except for financial risk, have positive effects on the resistance to smart home services. When the survey respondents are divided into two types, postponers and rejecters, the result of postponers is similar with that of total sample, except that privacy risk is unimportant to postponers, and the result of rejecter cannot satisfy the recommended model fit.

1. Introduction The industry related to the Internet of Things (IoT) is expected to grow rapidly. IoT Analytics (2018) anticipates that the global IoT market will grow to USD 1.56 trillion by 2025. One representative IoT service that has received considerable attention is the smart home, namely a residence equipped with information and communication technology for interoperability of household products and services (Peine, 2008). Although still in its realization stage, the global smart home market is expected to grow from USD 76.6 billion in 2018 to USD 151.4 billion in 2024, which represents a compound annual growth rate of 12.0% (Markets and Markets, 2016). Moreover, in recent years, the smart home has become a major focus at global electronics shows such as the International Consumer Electronics Show (CES) and various global companies have introduced products related to the smart home at such shows. Major global IT companies, such as Google, Amazon, Apple, and Samsung, are active in the smart home market and are the sales leaders in smart home devices (Dean, 2017). Google acquired the company NEST to increase its development of smart home devices and DropCam to offer products that allow consumers to monitor their homes while they are away via their smartphones. Google has also developed smart speakers as part of its Google Home brand, as an intelligent home assistant. Amazon provides smart home

∗

Corresponding author. E-mail address: [email protected] (S. Kim).

https://doi.org/10.1016/j.telpol.2019.101867 Received 16 May 2019; Received in revised form 23 August 2019; Accepted 23 August 2019 0308-5961/ © 2019 Elsevier Ltd. All rights reserved.

Please cite this article as: Areum Hong, Changi Nam and Seongcheol Kim, Telecommunications Policy, https://doi.org/10.1016/j.telpol.2019.101867

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

services based on its smart home assistant Echo, and Apple announced its smart home speaker HomePod at the Worldwide Developers Conference in 2017 (Garun, 2017). Samsung also have announced that it would make smart products connected with voice-enabled computing intelligence and Wi-Fi features (Song, 2017). Other players, such as telecommunication service providers, construction companies, and security service providers, have also expressed interest in the smart home market, and have released a variety of early-stage products and smart home services. For example, in Korea, telecommunication service providers have launched the smart home services “IoT@home” and “Giga IoT Home,” which bundle various smart home devices, such as smart plugs, thermostats, door locks, energy meters, and gas locks, along with communication services for a monthly fee. Furthermore, many companies are aiming to build smart home services into new apartments in conjunction with construction companies. However, contrary to optimistic expectations of future market growth, the smart home market remains in the early stages of development and has experienced difficulty in moving to the mass-market phase of adoption (Greenough, 2016). So even though many companies provide early-stage smart home products and services, they are not being widely diffused. In the US market, in May 2015, monthly sales of smart home products decreased by 15% from the same period the year before (Argus Insight, 2015). While demand for smart home products was mainly initiated by early adopters, demand has not broadened to general consumers (Higginbotham, 2015). For instance, although recently many consumers in the US have begun to use smart speakers, they rarely use them for smart home services. According to the Adobe (2018) survey, the most common voice activities are requesting music (70% of respondents) and weather forecasting (64%), while using a voice assistant for smart home commands (31%) is still an “emerging use case.” These facts suggest that smart home products and services have not diffused broadly to general consumers to progress the industry to the mass-market phase. With investigation of the reasons behind this situation and the introduction of appropriate solutions, the smart home market may grow more rapidly and become one of the key markets within the IoT industry. Therefore, it is important to determine why the smart home market cannot move into the mass-market phase despite the optimistic expectation for future market growth. One of the many reasons why smart home services have not diffused rapidly is that the technology necessary to enable these services has not been developed adequately (Yang, Lee & Zo, 2017). Another reason is that these services are not familiar and/or not friendly to consumers. As a result, consumers have not yet embraced smart home technologies and services, and this uncertainty about these technologies and services may give rise to a number of concerns regarding cost, performance, and even lifestyle change. Moreover, according to the Acquity Group (2014) IoT report, consumers have some specific concerns regarding privacy and/or security problems. Many consumers are worried that technology and software problems could lead to a loss of control of smart homes (Dynatrace, 2018), and the expense of smart homes is another concern (PwC, 2017). Moreover, complicated installation processes and difficulties in using smart home services have created general user resistance to smart home services (Argus Insight, 2015), which may lead to a reluctance to use smart home products and services. Because the lack of demand hampers the growth of this market and the development of related technologies, it is worthwhile analyzing the kinds of concerns that lead consumers to hesitate to adopt smart home products and services. Therefore, the purpose of this study was to identify the possible barriers to consumers' adoption of smart home services by means of a comprehensive model. Identifying the barriers as consumer uncertainty, concerns, and resistance, we propose a structural equation model (SEM) in which consumer uncertainty, represented by technological uncertainty and service intangibility, affects consumer concerns, represented by perceived risk, which is divided further into performance risk, financial risk, privacy risk, and psychological risk. While previous studies on the adoption of new technologies or IT services have focused on the determinants of adoption (Davis, 1989, pp. 319–340; Ajzen, 1991; Venkatesh, Morris, Davis, & Davis, 2003; Pavlou, 2003; Yang, Liu, Li, & Yu, 2015; Hsieh, Huang, & Wu, 2017; Yang et al., 2017), we identify the factors that affect consumers’ resistance to smart homes to determine why consumers are hesitant to adopt smart home products and services. Furthermore, we consider perceived risk of smart homes as the antecedents of resistance, and separate this perceived risk into performance risk, financial risk, privacy risk, and psychological risk, which are the main barriers to adoption of smart home products and services discussed in previous studies. Consumers may be uncertain about smart home technology and services due to the ongoing development of smart home technology and their unfamiliarity with smart home services, and these uncertainties can increase their risk perception (Pavlou, Liang, & Xue, 2007). Thus, we propose two types of perceived uncertainty—technological uncertainty and service intangibility—as the antecedents of the perceived risk of smart homes. Although previous studies suggest various antecedents of resistance (Joshi, 1991; Kim & Kankanhalli, 2009; Lapointe & Rivard, 2005; Marakas & Hornik, 1996; Markus, 1983; Ram & Sheth, 1989), they do not identify specifically the types of risk that can affect resistance and the antecedents that exist based on those risks. However, it is important to determine the specific risks that affect the resistance to smart homes, and their antecedents, to identify the real barriers to smart home adoption. Therefore, we incorporate perceived risk and perceived uncertainty, which have been used in previous studies, into our model to determine the factors affecting resistance among consumers to smart home adoption. Furthermore, we conduct an online survey targeting nonadopters of smart homes to estimate consumer resistance, risk, and uncertainty, and divide the survey respondents into two groups based on actual resistance behavior—postponement and rejection—and then analyze each group. The remainder of the paper is organized as follows. Section 2 presents a review of the literature related to smart home services. Section 3 presents the theoretical background and hypotheses of our research model. In Section 4, we discuss the collection of the data, sample characteristics, and the empirical results. Section 5 highlights implications of our findings and offers concluding remarks.

2

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

2. Literature review and theory 2.1. Previous research on smart homes A smart home can be defined as a residence equipped with high-tech devices, appliances, sensors, and networks, which allow remote access, control, and monitoring to enhance the convenience of the inhabitants (Balta-Ozkan, Davidson, Bicket, & Whitmarsh, 2013a; Chan, Campo, Estève, & Fourniols, 2009; Demiris et al., 2004). Smart homes allow consumers to control and manage home appliances through smart technologies (International Telecommunication Union, 2010), including a variety of services such as ehealth, entertainment, communications, assisted living, security, energy efficiency, and convenience. Balta-Ozkan et al. (2013a) note that smart homes are characterized by four key aspects: a communications network through which different devices talk to each other; intelligent controls to manage the system; sensors that collect information; and smart features … which respond to information from sensors or user instructions as well as the system provider. As smart home services include various types of services with different instruments, research into smart homes is also diversified. However, the literature review here focuses on work related to consumer concerns such as challenges and barriers to smart home services. In their review of the leading smart home projects in several developed countries, Chan, Estève, Escriba, and Campo (2008) discuss how one future challenge involves satisfying user needs. Ehrenhard, Kijl, and Nieuwenhuis (2014) suggest four key market barriers related to meeting end-user requirements, platform management, improved value creation and capture, and the role of government. In particular, one of the key market barriers is that end users are not convinced of the value of smart home technologies because of their unfamiliarity with this complex technology, fear of losing control, and privacy issues. Balta-Ozkan et al. (2013a, 2013b) investigate smart home development in the UK through expert and consumer interviews and identified several social barriers, such as the fit with current and changing lifestyles, administration problems, interoperability, reliability, privacy and security, trust, and costs. Both experts and consumers are concerned with issues relating to the malfunctioning, privacy, and security of these services. In addition, consumers are concerned with the installation, maintenance, repair, and energy costs, and are distrustful of whether smart technologies and services really reduce their costs, whereas experts are concerned with whether the devices can communicate with each other as well as interference between devices. Demiris et al. (2004) estimate older adults' attitudes toward smart home technology by using focus group interviews and found some benefits and concerns among older adults. Examples of the benefits identified were emergency help, temperature monitoring, and automatic lighting, while the concerns involved privacy violation, lack of human responders, possible replacement of human assistance by technology, user unfriendliness of devices, and the need for training tailored to older learners. However, the findings of Demiris et al. (2004) were not directly related to the concept of consumer resistance. Luor, Lu, Yu, and Lu (2015) analyze the effect of perceived usefulness, trust, and cost on residents’ attitudes toward smart homes for the smart home functions of home entertainment, home security, and home automation. Their results show that the perceived usefulness of home entertainment and home automation positively affect attitudes toward each function, while the costs of home entertainment and home automation negatively affect attitudes toward each function. In addition, trust in home security positively affects attitudes toward home security. Examining factors affecting user acceptance of smart home services, Yang et al. (2017) find that mobility, security/privacy risk, and trust in the service provider are important factors affecting the adoption of smart home services. Shin, Park, and Lee (2018) and Baudier, Ammi, and Deboeuf-Rouchon (2018) also analyze the factors affecting adoption and diffusion of smart homes and discuss the implications for promoting the smart home market. However, although these two studies analyze the benefits and concerns/attitudes toward smart homes empirically by using consumer interviews and surveys, they do not focus on the concept of resistance. In sum, many barriers preventing mass-market smart home adoption such as high device prices, limited consumer demand, and long device replacement cycles have been identified. Thus, here, we examine why consumer demand to date has been limited. Furthermore, while some studies estimate the attitudes or barriers to smart homes, no work has empirically estimated consumer resistance to smart homes. Previous studies about smart home typically identify the barriers or challenges from the market or industrial perspective based on interviews and a literature review or only focus on consumer acceptance rather than resistance (Demaris et al., 2004; Chan et al., 2008; Balta-Ozkan et al., 2013a, 2013b; Ehrenhard et al., 2014; Luor et al., 2015; Yang et al., 2017; Shin et al., 2018; Baudier et al., 2018). Therefore, it is worthwhile empirically analyzing the resistance to and perceived risks of smart home adoption from the consumer perspective. 2.2. Technology acceptance and resistance Previous studies on new technology adoption are based on two perspectives: acceptance and resistance. When a new information system emerges, users decide to adopt or resist it based on their evaluation of the new system (Joshi, 2005). Technological acceptance studies have used various appealing theoretical approaches including the technology acceptance model (TAM), the theory of planned behavior (TPB), and the united theory of acceptance and use of technology (UTAUT) (Kim & Kankanhalli, 2009). TAM (Davis, 1989, pp. 319–340) is one of the most widely accepted frameworks for understanding an individual's adoption of new technologies. TAM suggests that the two main antecedents of technology acceptance are perceived usefulness and perceived ease of use. Perceived usefulness is “the degree to which a person believes that using a particular system would be free of effort” and the perceived ease of use is “the degree to which a person thinks that using a particular system would enhance his or her job performance” (Davis, 1989, p. 320). TPB (Ajzen, 1991) is a comprehensive model of user acceptance and suggests that human behavior is affected by three main antecedents: attitude toward the behavior, subjective norms, and perceived behavioral control. Attitude toward the behavior refers to “the degree to which a person has a favorable or unfavorable evaluation or appraisal of 3

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

the behavior in question,” subjective norms are “the perceived social pressure to perform or not to perform the behavior,” and perceived behavioral control involves “people's perception of the ease or difficulty of performing the behavior of interest” (Ajzen, 1991). UTAUT (Venkatesh et al., 2003) is a model that unifies previously identified antecedents of technology acceptance from various theories, including TAM, TPB, the theory of reasoned action, social cognitive theory, and innovation diffusion theory. UTAUT explains how performance expectancy (perceived usefulness), effort expectancy (perceived ease of use), social influence (subjective norms), and facilitating conditions (perceived behavioral control) affect behavioral intention and use behavior. Another stream of research focuses on the resistance to technology. Resistance is defined as the behavior of trying to maintain the current state and not adopt the innovation when faced with the pressure to change from the current state (Ram, 1987). While most studies focus on adoption and/or positive aspects of innovations, some researchers have studied the concept of innovation resistance. Rogers (1983) uses the term “discontinuance” to mean “a decision to reject an innovation after having previously adopted it.” Sheth and Stellner (1979) discuss the concept of “innovation resistance,” emphasizing the importance of understanding the psychology of resistance because the majority of people have no a priori desire to change and may be more typical and more rational than the minority who seek the change. Ram (1987) also argues that innovation resistance is not the opposite of adoption, but rather a process that occurs during progress toward adoption. Therefore, in analyzing how an innovative service, such as a smart home, is adopted, it might be necessary to focus on both innovative early adopters as well as the resistance of most people who are not anxious to change. Several studies propose theoretical explanations of resistance. Markus (1983) describes how user resistance comes from a loss of power. If people believe that system change affects their power negatively, they will resist using the system. Ram and Sheth (1989) identify the five barriers of usage, value, risk, tradition, and image as the reasons why consumers resist innovation. Joshi (1991) proposes the equity implementation model, in which users are resistant to change if a net benefit is perceived, which is the difference between changes in outcomes and changes in inputs associated with the new information system. Marakas and Hornik (1996) find that resistance behavior can result from the fear and stress associated with the belief that technology is intruding into the stable world of users and is not necessarily motivated by personal gain. Using a case study, Lapointe and Rivard (2005) propose a model that explains individuals’ resistance. They found that resistance behavior comes from the perceived threat associated with the interaction between the object of resistance and the initial conditions of the individual. Kim and Kankanhalli (2009) develop a comprehensive model to explain user resistance to a new information system implementation by integrating the status quo bias theory, TAM, and the equity implementation model. They tested this model in the context of a new enterprise system implementation and found that switching costs increase user resistance, whereas perceived value and organizational support for change can reduce user resistance. In sum, existing models of resistance argue that losses or threats are key issues for resistance. However, it is not clear what types of threats or losses can affect resistance and what are the antecedents of these threats or losses. Therefore, we adopt perceived risk and perceived uncertainty in our resistance model and the determinants of the concerns about and resistance to smart homes. Some studies (Ram & Sheth, 1989; Szmigin & Foxall, 1998) further divide the concept of resistance from simply “not attempting to innovate” to three distinct types of consumer behaviors: rejection, postponement, and opposition. Rejection involves the active decision not to adopt an innovation and postpone, while opposition involves an active decision to oppose the innovation. Using focus group interviews, Kleijnen, Lee, and Wetzels (2009) investigate the drivers that affect each type of resistance and find that different factors affect each type of resistance. Rejection is affected by economic risk, tradition, and norms, postponement by economic risk and functional risk, and opposition by physical risk, functional risk, social risk, tradition and norms, and perceived image. Based on these distinct types of customer behaviors, Mzoughi and M'Sallem (2013) divide customer groups into postponers, rejecters, and opponents, and identify the factors affecting resistance to Internet banking adoption using a multinomial logistic regression and estimate the differences between groups. Laukkanen (2016) classifies customer groups slightly differently into adopters, postponers, and rejecters, and analyzes the barriers to Internet banking using a binary logit model. In sum, while consumer resistance models have been used in the banking industry, their application to new technologies or services, such as smart homes, has not received much attention to date. Because we focus on customer barriers to smart home services adoption, customers are divided into two groups—rejecters and postponers—based on their actual resistance behavior. Opponents are not included in the research because smart home services are in an early stage of market phase; therefore, there are few consumers who have already used smart home services and would resist using them based on their experience. 3. Research model and hypotheses This study proposes a model based on the relationship between perceived risk and innovation resistance (Fig. 1). Perceived risk, which is positively related to technological uncertainty and service intangibility, increases consumers’ resistance level, and is divided further into performance risk, financial risk, privacy risk, and psychological risk. Potential consumer uncertainty toward smart homes is factored into the two categories of technological uncertainty and service intangibility. Gender and age are used as the control variables for resistance. 3.1. Perceived risk Perceived risk relates to the uncertainty regarding the occurrence of adverse consequences (Bauer, 1960). Featherman and Pavlou (2003) define perceived risk as the extent to which consumers feel uncertainty about the possibility of negative consequences from using a technology. Many studies adopt the perceived risk concept in analyzing consumer behavior related to information technology and show that perceived risk is negatively related to the intention to use (Featherman & Pavlou, 2003; Lu, Hsu, & Hsu, 2005; Martins, 4

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Fig. 1. Research model.

Oliveira, & Popovič, 2014; Yang et al., 2015). Although perceived risk has commonly been used in technology acceptance research, it has not been employed as a major construct. However, Kahneman and Tversky (2013) argue that people place greater emphasis on loss than they do on gain in uncertain situations. Thus, loss is a crucial factor that has a significant impact on consumer behavior in such situations. If this is the case, it is important to focus on perceived risk, which is an important attribute of loss, when analyzing smart homes in their initial and unstable stage of development from the consumer perspective. Some studies show that perceived risk affects resistance (Kang & Kim, 2009; Ram & Sheth, 1989; Sheth & Stellner, 1979). Sheth and Stellner (1979) claim that the higher the perceived risk, the greater the innovation resistance, arguing that perceived risk is the major determinant of innovation resistance. Ram and Sheth (1989) propose barriers of physical, economic, functional, and social risks as the reasons that consumers resist innovation. Kang and Kim (2009) show that resistance, as a negative attitude, increases when consumers perceive the risk. They analyze consumer resistance to participating in multihop communication using perceived risk and innovation resistance. Based on the characteristics of multihop communication, they propose four types of risk: expected network quality concerns, expected privacy concerns, expected lack of cohesion, and expected source of service performance. As these types of risk contribute to consumer decisions not to adopt innovation, resistance to smart homes could be induced by such perceived risk. In sum, findings from research agree that perceived risk is negatively related to intention to use and positively related to resistance. Therefore, we also assume that perceived risk is a critical reason for resistance to smart homes. In addition, based on the characteristics of smart home services, risk is classified into performance risk, financial risk, privacy risk, and psychological risk. 3.1.1. Perceived performance risk Performance risk means that consumers are uncertain about the performance of smart home services or products. This includes concerns about the possibility of smart homes not operating as intended or advertised, and about their quality being below consumer expectations (Featherman & Pavlou, 2003). Ram and Sheth (1989) argue that consumers wish to know whether an innovation has been fully tested or proven as they may be concerned that there is a possibility that new devices or services may not function properly or reliably. They propose functional risk as a barrier that creates resistance and define this as the uncertainty of performance. In particular, for smart home services, consumers are concerned about the reliability of smart homes and may feel some uncertainty about their performance and whether a smart home can offer the expected benefits (Balta-Ozkan, Davidson, Bicket, & Whitmarsh, 2013b). In turn, consumers’ concerns about the performance of smart homes can increase resistance. Therefore, we hypothesize the following: H1. Perceived performance risk is positively related to resistance to smart homes. 3.1.2. Perceived financial risk Financial risk represents the possibility that the product or service will not be worth its financial price or investment, or that there may be a cheaper alternative available (Lu et al., 2005). According to a number of consumer survey results about smart homes, the price of a smart home device or service is the main barrier to smart home adoption for consumers. Acquity Group (2014) notes that price is one of the barriers to consumer adoption of IoT devices. A Survey Monkey questionnaire completed by 800 respondents in the US found that 58% thought that smart devices were too expensive (Bloomberg, 2016). LG Economic Research Institute (2015) also 5

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

argues that price is the most important factor for consumers buying smart home products. Maintenance and installation costs also become a burden for consumers in using smart home devices. Balta-Ozkan et al. (2013b) show that consumers are concerned about the cost of installation, repair, maintenance, and energy usage and that this burden may elicit consumer feelings of concern and guilt about spending a considerable amount of money. Because smart homes are composed of various related electronic devices, sensors, and a network system, the end user of the service has to pay the installation costs for sensors and/or some smart home devices, as well as maintenance costs and energy usage to run the smart home, in addition to the initial purchase cost of the devices. Therefore, consumers may worry about the economic burden of such costs and these concerns could affect consumer resistance to smart homes. Ram and Sheth (1989) argue that the higher the cost of an innovation, the higher the perceived economic risk, which is one risk barrier that generates resistance. Therefore, we assume that financial risk, including concerns about the cost of purchase, installation, and maintenance, is a factor that increases resistance to smart homes. Therefore, we hypothesize the following: H2. Perceived financial risk is positively related to resistance to smart homes. 3.1.3. Perceived privacy risk Privacy risk represents the concern consumers feel about the risk of having personal data used improperly without agreement or private information disclosed to third parties (Kang & Kim, 2009). The privacy issue is a serious challenge for all high-tech products and services and represents a major barrier to smart home diffusion. Smart homes capture a large amount of data, including very private data such as the occupants’ daily routines or health status. As a result, consumers may have strong fears concerning their privacy. Acquity Group (2014) notes that privacy concerns are one of the largest barriers to smart home adoption. Icontrol Networks (2015) report that consumers were concerned about the possibility of a smart home data breach: 71% of consumers surveyed feared that their personal information could be stolen and 64% worried that their data would be collected and sold. According to BaltaOzkan et al. (2013b), many experts, consumers, and studies argue that privacy and security are the main barriers to smart home development. Consumers are particularly concerned with strangers knowing their daily routines and worry about their personal data landing in the wrong hands or their smart home system being hacked. In sum, if consumers worry about privacy infringement in relation to smart homes, this concern could increase resistance to smart home adoption. Therefore, we hypothesize the following: H3. Perceived privacy risk is positively related to resistance to smart homes. 3.1.4. Perceived psychological risk Perceived psychological risk represents the risk that smart homes would have a negative effect on consumers' peace of mind or self-perception (Featherman & Pavlou, 2003). Because smart homes are not just a service or a product, but a space that occupants live in, adoption of smart home services could affect occupants' self-image or lifestyle. Balta-Ozkan et al. (2013b) argue that the fit to occupants’ current and changing lifestyles is one of the barriers to smart home development. Consumers may be concerned that they will lose control of their smart home services, lose control of their daily household routines, and possibly, become lazy. Moreover, they may think that smart home services will make them dumb and complacent. Therefore, we hypothesize the following: H4. Perceived psychological risk is positively related to on resistance to smart homes. 3.2. Technological uncertainty Uncertainty refers to an individual's perceived inability to predict accurately the future state of their environment (Milliken, 1987; Pavlou et al., 2007). People can experience uncertainty about the future state of their environment, the impact of changes to that environment, and the consequences of their response options, because they believe that they lack sufficient information to predict accurately or because they feel unable to discriminate between relevant and irrelevant data (Ellis & Shpielberg, 2003; Milliken, 1987; Vecchiato & Roveda, 2010). This concept has been used in the literature on strategy and organizations as a key issue for business decisions (Vecchiato & Roveda, 2010), and in the literature about information systems. Furthermore, many studies show that perceived uncertainty can raise the perceived risk and decrease the purchase or utilization intentions regarding information systems (Kim & Kim, 2018; Pavlou et al., 2007; Yang et al., 2015; Yeh, Hsiao, & Yang, 2012). When consumers perceive uncertainty about their transactions, they easily tend to overestimate the probability of potential losses. Therefore, uncertainty perception makes them perceive higher risks, which reduces their intention to use information systems (Pavlou et al., 2007). Yang et al. (2015) show that perceived technological uncertainty is also related to consumers’ perceived risk of information system adoption. Technological uncertainty is a type of perceived uncertainty and has been examined in the literature on strategy and organizations (Ellis & Shpielberg, 2003; Heavey & Simsek, 2013; Ragatz, Handfield, & Petersen, 2002; Song & Montoya-Weiss, 2001; Stock & Tatikonda, 2008). Technological uncertainty refers to “the inability to completely understand or accurately predict some aspect of the technological environment” (Song & Montoya-Weiss, 2001). It relates to the unpredictability of technological development and the related technological environment, as well as to the consequences of the technology (Song & Montoya-Weiss, 2001; Yang et al., 2015). Yang et al. (2015) adopt this concept to study information system adoption, and test the effect of technological uncertainty on perceived risk in mobile payment systems. They classify perceived uncertainty into four categories—technological uncertainty, information asymmetry, regulatory uncertainty, and service intangibility—and prove empirically that these four perceived uncertainties are the antecedents of perceived risk. Therefore, based on these studies, perceived technological uncertainty is 6

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

likely to raise perceived risk. In this paper, we hypothesize that consumers’ uncertainty about smart home technology increases their perceived risk of smart homes. As the underlying technologies of smart homes, such as sensing, actuation, and network devices, are still under development, consumers are uncertain about them. Consumers may worry about the malfunctioning of smart homes, such as sensors accidently turning off or the whole system going into limbo due to a breakdown in the remote controls that operate the home (Balta-Ozkan et al., 2013b). Unstable network connections also make smart home services unreliable (National Information Society Agency, 2016). Furthermore, consumers are worried about security systems. In a consumer survey, Cisco (2017) reported that only 9% of respondents trust the security system of IoT devices. According to the global survey by intelligent software company Dynatrace (2018), many consumers worry that bugs, malfunctioning of smart home technology, or software problems may cause a loss of control of their home. These findings show that consumers perceive that smart home technology is underdeveloped, leading to uncertainty and concerns about smart home performance and privacy, and loss of control of their smart home. Therefore, we hypothesize the following: H5-1. Technological uncertainty is positively related to perceived performance risk. H5-2. Technological uncertainty is positively related to perceived financial risk. H5-3. Technological uncertainty is positively related to perceived privacy risk. H5-4. Technological uncertainty is positively related to perceived psychological risk. 3.3. Service intangibility Berry (1980) describes intangibility as a service that cannot be easily defined, formulated, or grasped mentally. Zeithaml, Parasuraman, and Berry (1985) identified the archetypal characteristics as intangibility, heterogeneity, inseparability, and perishability. Usually, these characteristics imply a definition by exclusion, and intangibility means “the lacking the palpable or tactile quality of goods” (Vargo & Lusch, 2004). Intangibility affects consumer decision-making (Laroche, McDougall, Bergeron, & Yang, 2004), and it can increase uncertainty in decision-making, thereby lowering consumer confidence and increasing perceived risk (Mitchell, 1999). In other words, as intangibility increases, so does perceived risk (Murray & Schlacter, 1990). Some studies empirically show the relationship between service intangibility and perceived risk. Laroche, Bergeron, and Goutaland (2001) propose that intangibility is composed of three distinct dimensions: physical intangibility, generality, and mental intangibility. Physical intangibility means that a good is inaccessible to the senses and lacks a physical presence, while generality means that consumers have difficulty in precisely defining or describing a particular service. Mental intangibility means that the operation of a service or product is difficult to grasp mentally. Laroche et al. (2004) show that the dimensions of intangibility, generality, and mental intangibility affect each types of perceived risks, that is, financial, time, performance, social, and psychological. Featherman and Wells (2004, 2010) show that the intangibility of information systems positively affects perceived risk. Featherman and Wells (2010) analyze the relationship between mental intangibility and the perceived risk of e-services and find that mental intangibility increases both the overall perceived risk and each type of risk (performance, financial, privacy, time, social, and psychological). Eggert (2006) also shows that intangibility positively affects the perceived financial, time, performance, social, and psychological risks, and its impact in the online context is greater than that offline. Finally, Yang et al. (2015) show that service intangibility affects the perceived risks of mobile payment system positively. Smart home services are very complex and differ from traditional services in the early stages of market introduction. Therefore, considering that consumers are unfamiliar with smart home services, and that the concept or performance properties of smart homes are very vague for consumers, service intangibility here is taken to mean that consumers have difficulty in defining, describing, or conceptualizing smart homes. As this uncertainty could increase consumers’ perceived risk of smart homes, we hypothesize the following: H6-1. Service intangibility is positively related to perceived performance risk. H6-2. Service intangibility is positively related to perceived financial risk. H6-3. Service intangibility is positively related to perceived privacy risk. H6-4. Service intangibility is positively related to perceived psychological risk. 4. Methods and results 4.1. Sample The data were collected through an online survey by Macromill Embrain, a professional research firm in Korea, from April 27 to May 4, 2017. The survey targeted consumers who do not use a smart home product and service at present. Given that the aim of this study was to determine the antecedents that make consumers hesitate to use smart home services, we focused on consumers who showed resistance behavior rather than on those who already adopted a smart home products or services. The survey included a short explanation about smart homes to help respondents understand the concept. A pilot test was conducted and the questionnaire was 7

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Table 1 Sample characteristics. Demographic category Gender Age

Monthly income (KRW)

Residence type Residence ownership

Male Female 20–29 30–39 40–49 50–59 < 1 million 1–3 million 3–5 million 5–7 million 7–9 million > 9 million Single house Apartment/multifamily house Homeowner Renter

Total sample (n = 533)

Postponers (n = 391, 73.4%)

Rejecters (n = 142, 26.6%)

265 268 131 133 132 137 31 148 183 102 47 22 53 480 322 211

198 193 104 96 92 99 22 103 139 76 35 16 37 354 233 158

67 75 27 37 40 38 9 45 44 26 12 6 16 126 89 53

49.7% 50.3% 24.6% 25.0% 24.8% 25.7% 5.8% 27.8% 34.3% 19.1% 8.8% 4.1% 9.9% 90.1% 60.4% 39.6%

50.6% 49.4% 26.6% 24.6% 23.5% 25.3% 5.6% 26.3% 35.5% 19.4% 9.0% 4.1% 9.5% 90.5% 59.6% 40.4%

47.2% 52.8% 19.0% 26.1% 28.2% 26.8% 6.3% 31.7% 31.0% 18.3% 8.5% 4.2% 11.3% 88.7% 62.7% 37.3%

slightly revised in response to comments to make it easier to understand. For responses, we used a seven-point Likert scale, from strongly disagree to strongly agree. The total number of respondents was 535; after discarding incomplete responses, 533 valid responses were included in the analysis (391 postponers and 142 rejecters). Those who responded “yes” to the question of whether they were willing to use a smart home product and service in the future were classified as postponers; those who responded “no” were classified as rejecters. Table 1 shows the sample characteristics. The gender and age of the respondents were almost uniformly distributed. There were 265 males and 268 females and 131 respondents were in their 20s, 133 in their 30s, 132 in their 40s, and 137 in their 50s. The male and female proportions in actual population in 2017 were 51.4% and 48.6%, respectively, showing almost equal distribution. The distribution of ages from twenties to fifties in actual population in 2017 was 22.0%, 24.0%, 27.3%, and 26.7%, respectively, and it is similar to the distribution of ages from twenties to fifties in the total sample, that is 24.6%, 25%, 24.8% and 25.7%, respectively. Of the 533 respondents, 53 lived in a single house and 480 in apartments or in a multifamily house; 322 were homeowners and 211 renters. 4.2. Measurement model Tables 2 and 3 shows the measurement items used in this study and their validity. The items for each construct are based on those used in previous studies; however, we modified them slightly to fit the context of this research. The research model was estimated using covariance-based structural equation modeling, and AMOS 20 was used for the analysis. The validity of the measurement model was evaluated using construct reliability, convergent validity, and discriminant validity. The reliability of each individual item was evaluated by examining each item's loading on its corresponding latent variable. All the item loadings in Table 2 are > 0.7 (Barclay, Higgins, & Thompson, 1995). To check the internal consistency, composite reliability (CR) and Cronbach's alpha were used. The CR and Cronbach's alpha in Table 2 are both > 0.7 (Hair et al., 1998; Fornell & Larcker, 1981). The average variance extracted (AVE) was used to check convergent validity. Table 3 shows that the AVE values are > 0.5, as suggested by Bagozzi and Yi (1988). Thus, the results indicate that the reliability, internal consistency, and convergent validity of the model are satisfactory. In addition, to check the discriminant validity, we compared the square roots of the AVEs with the correlations between the variables. Table 3 shows that the diagonal values, which are the square roots of the AVEs, are higher than the correlations for each construct and our measurement model achieves discriminant validity (Fornell & Larcker, 1981). However, the square root of the AVE of technological uncertainty is slightly lower than the correlation between the technological uncertainty and perceived performance risk. This is because the AVE of technological uncertainty is somewhat low, even though it is higher than the recommended value. To assess the discriminant validity between technological uncertainty and perceived risk, we constrained the correlation between technological uncertainty and perceived risk to 1 and conducted a chi-squared difference test between the constrained and unconstrained models (Anderson & Gerbing, 1988). The results show that the chi-squared difference test is statistically significant (Δχ2 (1) = 15.192, p < 0.001) and that there is discriminant validity between technological uncertainty and perceived risk. Harman's single factor test is conducted to test for common method bias. The test results showed that the total variance of the single factor was 38.4%, not exceeding 50%, which does not account for the majority. This means that the common method bias was not a threat (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Further, we found the correlation between constructs to be less than the threshold value of 0.9 (Pavlou et al., 2007). 4.3. Proposed model and hypothesis testing The results of the SEM analyses of the proposed research model are summarized in Table 4 and 5. The fit of the research model satisfied all recommended values. Recommended values of model fit are shown in Appendix C. The ratio χ2/df is 3.606, goodness-of8

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Table 2 Measurement assessment. Construct

Item

Measurement question

Loading

α

CR

AVE

Technological uncertainty

TU1 TU2 TU3 SI1 SI2 SI3 PPR1

I think that the wireless network of smart homes is unstable. The security of smart homes is questionable. I think that the technologies related to smart homes are undeveloped. It is difficult to explain the features and functions of smart homes to others. It is difficult to understand how smart homes work. It is difficult to conceptualize smart homes. If I were to use smart home products and services, I would be concerned that the smart home would not provide the level of benefits that I would expect. The performance of smart home services and products may not match their advertised level. It is uncertain whether smart homes will operate as satisfactorily as expected. I am concerned that I really would not get my money's worth from a smart home. I am concerned that it would cost a lot to purchase and install smart home products and services. I am concerned that it would cost a lot to repair and maintain a smart home. If I use the smart home products and services, private information could be misused, inappropriately shared, or sold. If I use the smart home products and services, personal information could be intercepted or accessed. If I use the smart home products and services, the chance of losing control over my private information is high. A smart home will not fit in with my self-image or self-concept. A smart home will not fit in with my lifestyle. A smart home will make me lose control of my home and become indifferent or lazy at home. I will feel uneasy if I use a smart home products and services. My current state is better than using the smart home products and services I am reluctant to use the smart home products and services. If I use the smart home products and services, I will be dissatisfied with the smart home.

0.746 0.787 0.716 0.785 0.928 0.813 0.776

0.795

0.738

0.563

0.877

0.841

0.713

0.876

0.840

0.704

0.861

0.822

0.696

0.930

0.895

0.817

0.868 0.913 0.650

0.842

0.785

0.670

0.789 0.787 0.875 0.846

0.894

0.856

0.681

Service intangibility Perceived performance risk

PPR2 PPR3 PFR1 PFR2

Perceived financial risk

PFR3 PPrR1

Perceived privacy Risk

PPrR2 PPrR3 Perceived psychological risk

PPsR1 PPsR2 PPsR3

Resistance

RES1 RES2 RES3 RES4

0.833 0.903 0.729 0.885 0.880 0.905 0.927 0.879

RES = resistance, PPR = perceived performance risk, PFR = perceived financial risk, PPrR = perceived privacy risk, PPsR = perceived psychological risk, TU = technological uncertainty, SI = service intangibility. Table 3 Discriminant validity of the construct. Construct

TU

SI

PPR

PFR

PPrR

PPsR

RES

TU SI PPR PFR PPrR PPsR RES

0.750 0.277 0.789 0.565 0.667 0.326 0.486

0.844 0.343 0.335 0.235 0.519 0.535

0.839 0.639 0.57 0.289 0.457

0.834 0.606 0.334 0.429

0.904 0.319 0.452

0.818 0.812

0.825

fit index (GFI) is 0.886, adjusted goodness-of-fit index (AGFI) is 0.854, standardized root mean square residual (SRMR) is 0.066, normed fit index (NFI) is 0.903, comparative fit index (CFI) is 0.928, Tucker–Lewis index (TLI) is 0.915, and root mean square error of approximation (RMSEA) is 0.070. The resistance variable shows high explanatory power (R2 = 0.737) and perceived performance risk, perceived financial risk, perceived privacy risk, and perceived psychological risk account for 69.0%, 46.2%, 50.7%, and 32.4% of the variances (R2 values) of technological uncertainty and service intangibility, respectively. All hypotheses, except for H2 and H6-3, are supported by the research model. As expected, the perceived performance risk, privacy risk, and psychological risk positively affect resistance to smart homes with coefficients of 0.166, 0.114, and 0.914, respectively, supporting H1, H3, and H4. However, the results show that financial risk does not affect resistance to smart homes significantly because H2 is rejected. This is because smart home products and services are available at various price points. Even though there are many expensive smart home devices and services, some Korean companies such as telecommunication service providers market early-stage smart home products and services to consumers at affordable prices. For example, the smart home service provided by LG, known as U+, provides several smart home products at the very low price of KRW 12,100 per month (USD 10.72)2. The consumers who use this service can choose three smart home products such as smart door locks, smart plugs, smart meters, security cameras, gas detectors, and smart lights. Korea Telecom also provides several simple smart home products and services at KRW 3000–11,000 per month (USD 2.66–9.74). These companies help raise the familiarity of consumers with smart homes through simple and inexpensive products and services. Thereby, consumers perceive less financial risk in smart homes, which lowers resistance. 9

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Table 4 Result of hypotheses tests (total sample).

H1 H2 H3 H4 H5-1 H5-2 H5-3 H5-4 H6-1 H6-2 H6-3 H6-4 Control variable Model fit R2

Path

Estimate

CR

p

Result

PPR → RES PFR → RES PPrR → RES PPsR → RES TU → PPR TU → PFR TU → PPrR TU → PPsR SI → PPR SI → PFR SI → PPrR SI → PPsR Age → RES Gender → RES χ2/df = 3.606 NFI = 0.903 RES 0.737

0.166*** 0.014 0.114** 0.914*** 1.065*** 0.817*** 1.002*** 0.229*** 0.136*** 0.182*** 0.060 0.411*** −0.004 0.082 GFI = 0.886 CFI = 0.928 PPR 0.690

3.732 0.361 3.087 13.910 15.089 11.925 13.236 4.534 3.314 3.991 1.287 9.154 −1.440 1.310 AGFI = 0.854 TLI = 0.915 PFR 0.462

0.000 0.718 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.198 0.000 0.150 0.190 SRMR = 0.066 RMSEA = 0.070 PPrR 0.507

Accept Reject Accept Accept Accept Accept Accept Accept Accept Accept Reject Accept – – PPsR 0.324

***p < 0.001, **p < 0.01, *p < 0.05. RES = resistance, PPR = perceived performance risk, PFR = perceived financial risk, PPrR = perceived privacy risk, PPsR = perceived psychological risk, TU = technological uncertainty, SI = service intangibility.

At the same time, our results show that other perceived risks, except for financial risk, increase resistance to smart homes. Perceived risks have a significantly positive relationship with nonadopter resistance to smart home services, supporting the belief that risk is one of the barriers that affects consumer resistance (Ram & Sheth, 1989). Although the relationship between adoption of a service and these risks has been proven empirically in many studies (Featherman & Pavlou, 2003; Lu et al., 2005; Martins et al., 2014; Yang et al., 2015), the relationship between resistance and these risks has not been elucidated. From this perspective, our study makes a unique contribution by adding empirical findings that support the positive impact of each risk on resistance to smart homes. Regarding the effects of uncertainty and intangibility, first, technological uncertainty is positively related to perceived performance risk, financial risk, privacy risk, and psychological risk with coefficients of 1.065, 0.817, 1.002, and 0.229, respectively, thereby supporting H5-1, H5-2, H5-3, and H5-4. Furthermore, service intangibility also positively affects perceived performance risk, financial risk, and psychological risk with coefficients of 0.136, 0.182, and 0.411, respectively. However, the effect on perceived privacy risk is not significant at the 5% level. Thus, H6-1, H6-2, and H6-4 are supported, while H6-3 is not. These results show that technological uncertainty and service intangibility play a role as antecedents of these risks that affect resistance to smart homes. Technological uncertainty positively affects all of the perceived risks suggested here. As smart homes are a technology-intensive product, the more consumers feel uncertain about their technology, the more they will associate risk with smart homes. Service intangibility also positively affects performance risk, financial risk, and psychological risk, although it is without a significant relationship with privacy risk. If consumers are unsure about what smart home services is, they may have concerns about their performance and cost, as well as associated changes in self-image and lifestyle. However, privacy risk is the risk of having personal data used improperly without agreement or private information disclosed to third parties, so it may be more related to the technological aspects of smart home services, rather than service aspect. The consumer can consider the privacy problem as the technological part, rather than service because especially smart home includes various technological products and sensors and the privacy leaks are mainly caused by these sensors or products such as home CCTV. Therefore, the impact of intangibility about these services on privacy risk may be low. 4.4. Analysis by group: postponers and rejecters We divided the respondents into two groups, postponers and rejecters, and analyzed the survey responses by group. Postponers have not yet adopted smart home services, but have rather postponed their decision; they were willing to adopt smart home services within on average 3.5 years. Rejecters have not adopted smart home services and are not willing to do so in the future. Table 5 shows the mean value of each latent variable and result of t-test between two groups. Mean value of all latent variables in rejecter group is higher than that in postponer group. T-test result of Table 5 shows that these mean differences are significant. This means that the rejecter group feel higher risk and resistance than the postponer group. The results for these two groups are summarized in Table 61. For the postponers, the model fit, except that for NFI, satisfies all recommended values. The ratio χ2/df is 2.732, GFI is 0.883, AGFI is 0.849, SRMR is 0.067, NFI is 0.892, CFI is 0.928, TLI is 0.915, and RMSEA is 0.067. NFI is a little lower than the recommended value but is acceptable because other indicators of model fit satisfy the recommended values. For postponers, 70.5% of the resistance to the smart home is explained by perceived risk (R2 = 0.705) and 1

At the exchange rate of KRW 1129 = USD 1. 10

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Table 5 T-test result between postponer and rejecter.

Postponer Rejecter t-value

PTU

PSI

PPR

PFR

PPrR

PPsR

RES

4.29 4.77 −4.806***

3.49 4.00 −4.493***

4.26 4.65 −3.390**

4.56 5.02 −5.011***

4.75 5.34 −4.677***

3.20 4.19 −8.352***

3.34 4.37 −10.055***

***p < 0.001, **p < 0.01, *p < 0.05. RES = resistance, PPR = perceived performance risk, PFR = perceived financial risk, PPrR = perceived privacy risk, PPsR = perceived psychological risk, TU = technological uncertainty, SI = service intangibility. Table 6 Result of hypotheses tests (postponers vs. rejecters). Path

H1 H2 H3 H4 H5-1 H5-2 H5-3 H5-4 H6-1 H6-2 H6-3 H6-4 Control variable R2

PPR → RES PFR → RES PPrR → RES PPsR → RES TU → PPR TU → PFR TU → PPrR TU → PPsR SI → PPR SI → PFR SI → PPrR SI → PPsR Age → RES Gender → RES RES PPR PFR PPrR PPsR

Postponers

Rejecters

Estimate

C.R.

p

Estimate

C.R.

p

0.153*** 0.018 0.062 0.765*** 1.078*** 0.796*** 0.947*** 0.146* 0.182*** 0.218*** 0.080 0.464*** −0.008*** 0.071 0.705 0.684 0.421 0.481 0.357

3.403 0.466 1.565 11.200 12.404 9.351 10.607 2.491 3.588 3.790 1.448 8.527 −2.695 1.103

0.000 0.641 0.118 0.000 0.000 0.000 0.000 0.013 0.000 0.000 0.148 0.000 0.007 0.270

0.226* 0.009 0.190** 0.969*** 1.039*** 0.805*** 1.059*** 0.168* 0.073 0.116 −0.021 0.238** 0.009 0.081 0.650 0.686 0.499 0.481 0.149

2.232 0.086 2.592 5.458 7.864 6.596 6.947 2.005 0.987 1.537 −0.221 3.237 1.395 0.617

0.026 0.932 0.010 0.000 0.000 0.000 0.000 0.045 0.324 0.124 0.825 0.001 0.163 0.537

***p < 0.001, **p < 0.01, *p < 0.05. RES = resistance, PPR = perceived performance risk, PFR = perceived financial risk, PPrR = perceived privacy risk, PPsR = perceived psychological risk, TU = technological uncertainty, SI = service intangibility.

the R2 of perceived performance risk, financial risk, privacy risk, and psychological risk are 0.684, 0.421, 0.481, and 0.357, respectively. Thus, all hypotheses, except for H2, H3, and H6-3, are supported. As expected, perceived performance risk and psychological risk positively affect resistance to smart homes with coefficients of 0.153 and 0.765, respectively, supporting H1 and H4. However, the results show that financial risk and privacy risk do not affect resistance to smart homes significantly. The result for financial risk is the same as that for total respondents. Furthermore, the resistance of postponers, who are the potential buyers of smart homes, is not affected by privacy risk; this shows that consumers do not appear to take privacy risks seriously when accepting new technology products, unlike the general expectation that privacy will impact acceptance or resistance. Rauschnabel, He, & Ro (2018) suggest that the privacy of individuals does not affect their purchase intentions because of four factors: flexibility and control, resignation, abstractness of consequences, and nothing to hide. In other words, the user's privacy risk does not affect their decision because they believe they can control the privacy problem, that they cannot change it, that the consequences from privacy-related problems are abstract, vague, or unrealistic, or that there is nothing to hide. These reasons explain why the privacy risk of postponers does not affect their resistance to smart homes. However, the result for the privacy risk of postponers differs from that for total respondents. While privacy risk affects the resistance to smart homes significantly in the result for total respondents, it does not affect the resistance to smart homes significantly in the result for postponers, because total respondents include rejecters. Homes are a very private space, and some residents may perceive a higher privacy risk about the smart home. Indeed, rejecters perceive a higher privacy risk than postponers according to the average value of the privacy risk measure: 5.34 for rejecters and 4.75 for postponers. Furthermore, this leads to the result that the privacy risk of rejecters affects their resistance to smart homes significantly. The result for total respondents includes this result for rejecters and differs from the result with postponers. In the results for rejecters, the model fit, except for χ2/df, does not satisfy all recommended values. The ratio χ2/df is 2.572, GFI is 0.756, AGFI is 0.687, SRMR is 0.097, NFI is 0.779, CFI is 0.851, TLI is 0.824, and RMSEA is 0.104. While the ratio χ2/df satisfies the recommended value and SRMR is < 0.1 as recommended by Kline (2015), other model fits do not satisfy the recommended values. In Table 6, we present the results for rejecters for reference only. 11

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

5. Implications and conclusions This paper examines the possible barriers to consumer adoption of smart home services, using the concepts of perceived risk, uncertainty, and resistance, and empirically finds a relationship between technological uncertainty, service intangibility, perceived risk, and resistance. We propose four types of perceived risk—performance risk, financial risk, privacy risk, and psychological risk—as the antecedents of the resistance to smart homes, and perceived uncertainty (technological uncertainty and service intangibility) as the antecedents of these risks. We test our proposed model empirically and show that technological uncertainty and service intangibility positively affect perceived risk and that the perceived risks, except for financial risk, provoke resistance. In other words, technological uncertainty and service intangibility with respect to smart homes increase each type of perceived risk and these risks increase resistance, except for the insignificant relationship between service intangibility and privacy risk and that between financial risk and resistance. This study presents academic implications as follows. First, this study focused on the resistance of new IT products and services. In the technology adoption literature, few studies have focused on the resistance to new IT services and tested the resistance model empirically. Many studies have focused on the only acceptance of information system. However, according to Ram (1987), the resistance is a process that occurs during progress toward adoption, which is not the opposite concept of adoption. It is a natural psychological state of consumers in the process of accommodating new products and services (Ram, 1987), and the higher the resistance, the slower the distribution of new products and services (Ram & Sheth, 1989). Therefore, it may be more effective to look at resistance than acceptance when looking at the slow distribution of products or services in the early stages of the market, such as smart home. Second, the researches about the acceptance based on TAM, TPB, UTAUT, etc. have been studied for a long time, and the concepts such as attitude, perceived usefulness, perceived ease of use, subjective norms, perceived behavioral control have been suggested and analyzed as the antecedent of acceptance. The constructs like perceived usefulness, perceived ease of use usually focus on the positive aspect of the new information system, but these are not appropriate to antecedent of the resistance because the resistance is not necessarily motivated by the personal gain and it usually comes from the negative aspect such as fear, stress, a loss of power, etc. (Marakas & Hornik, 1996; Markus, 1983). Therefore, this paper focused on more negative aspect of new products and service adoption, included the perceived risk in the model as the antecedent of the resistance, and proved the positive relationship between perceived risk and the resistance, regarding smart home. Third, although previous work suggests loss or threat as the antecedents of resistance, the type of threat has not been classified and the factors affecting the threat has not been identified empirically. The present study proposes a resistance model in which four types of perceived risk—performance risk, financial risk, privacy risk, and psychological risk—arising from perceived uncertainty can increase resistance and we empirically demonstrate the validity of this model in the context of smart homes. Thus, our work provides a valuable contribution to the technology adoption research field. Fourth, this paper focused on the consumer perception about the new products and services. The technology attribute such as relative advantages, compatibility, complexity, trialability, and observability have been used as the predictors for intention to adopt in the diffusion of innovations theory. These constructs reflect the nature of technology well, but it may be difficult for respondents to perceive and respond about them, in the situation where the products and services are still in early-stage market. Therefore, this study tried to use the concept that consumers can perceive better. Fifth, it is worthwhile that this research focused on the postponers, who are the potential users for smart home. We tested the model in postponer group and found that the privacy risk does not affect the resistance to smart home in postponer group in spite of general expectation that the consumer's privacy risk lead to resistance. With respect to managerial or policy implications of our results, first, the companies that market smart home services should provide useable products and services at affordable prices. Although financial risk does not affect resistance, performance risk increases resistance. So even if companies provide smart home products and services at reasonable prices, consumers will recognize that they are newly developed and will not use them if they do not perform well. Therefore, smart home companies need to focus on the quality of their products and services as much as on price. Second, smart home technology development and promoting the excellence of the technology are important. The underlying smart home technology is not yet sufficiently well developed, and consumers are uncertain about the technology. Technological uncertainty is relevant in the sense that consumers trust technology development and its performance, and this uncertainty affects perceived risks such as performance, financial, and privacy risks. According to Marakas and Hornik (1996), resistance behavior is caused by threat and is not necessarily motivated by personal gain. This means that if consumers cannot trust the underlying technology of smart homes, they will hesitate to use a smart home product and service even if the smart home companies continuously promote their benefits. Therefore, it is important that consumers have trust and confidence in smart home technology and its performance through continuous research and development of related technology and advertising. Third, smart home providers should raise consumers’ familiarity with smart homes. Many consumers are unaware of what smart homes are and their benefits. This unfamiliarity, as represented by service intangibility, can increase performance, financial, and psychological risks. Thus, if consumers are unfamiliar with smart homes and cannot conceptualize them, they will be uncertain about their performance and their worth, and how a smart home might change their lives. Fourth, even though the privacy risk affect the resistance to smart home, there is no significant relationship between resistance and privacy risk in postponer group, the potential consumers for smart home, and also service intangibility of smart home does not affect the privacy risk. This means that reducing the intangibility about smart home service and product does not decrease the privacy risk and privacy risk itself does not increase the resistance to smart home. Therefore, regarding to privacy risk, it is more efficient that 12

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

the smart home service and product providers focus on establishing safer smart home systems that prevent the leaking of private information technically, rather than raising consumer's understanding of smart home or reducing their anxiety about privacy. However, policy makers had better reinforce the regulations regarding privacy so that consumer's insensitivity to privacy problem does not lead to leaking of their private information. Our study has several limitations. First, because this study focuses on Korea, we cannot clearly generalize our findings to other countries, even though Korea has shown relatively rapid progress in the IT industry including smart home services. Second, we examine the positive relationship between perceived risk and resistance; however, there may be other factors affecting resistance that should be considered in future studies. Third, because our study targets consumers who do not presently use a smart home product and service, research on current smart home users, who can be assumed to have low resistance, would be insightful. Therefore, future studies should investigate uncertainty, risk, and the resistance of smart home adopters. Fourth, although this paper tried to find the factor affecting resistance to smart home, it was not enough to prove causal relationship between the constructs because the data gathered in this paper was based on a cross-sectional online survey, not a survey of longitudinal design. Future studies should consider the survey of longitudinal design to prove causal relationship. Acknowledgement This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2016S1A3A2924760). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.telpol.2019.101867. Appendix B Table B1

Reference of measurement item Construct

Item

References

Technological uncertainty

TU1 TU2 TU3 SI1 SI2 SI3 PPR1 PPR2 PPR3 PFR1 PFR2 PFR3 PPrR1 PPrR2 PPrR3 PPsR1 PPsR2 PPsR3 RES1 RES2 RES3 RES4

Yang et al. (2015) Yang et al. (2015) Yang et al. (2015) Laroche et al. (2004), Yang et al. (2015) Laroche et al. (2004), Yang et al. (2015) Laroche et al. (2004), Yang et al. (2015) Stone and Grønhaug (1993) Yang et al. (2015) Yang et al. (2015) Stone and Grønhaug (1993) Balta-Ozkan et al. (2013b) Balta-Ozkan et al. (2013b) Yang et al. (2015) Yang et al. (2015) Featherman and Pavlou (2003) Featherman and Pavlou (2003) Balta-Ozkan et al. (2013b) Balta-Ozkan et al. (2013b) Kang and Kim (2009) Kang and Kim (2009) Kim and Kim (2011) Kang and Kim (2009)

Service intangibility Perceived performance risk Perceived financial risk Perceived privacy Risk Perceived psychological risk Resistance

Appendix C Table C1

Recommended values of model fit Fit indices 2

χ /df GFI AGFI SRMR NFI

Recommended values

Reference

≤5 ≥0.8 ≥0.8 ≤0.08 ≥0.9

Wheaton, Muthen, Alwin, and Summers (1977) Wang and Chiu (2011) Wang and Chiu (2011) Hu and Bentler (1999) Bentler and Bonett (1980)

13

(continued on next page)

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Table C1 (continued) Fit indices

Recommended values

Reference

CFI TLI RMSEA

≥0.9 ≥0.9 ≤0.08

Bentler (1990) Rantanen, METSÄPELTO, Feldt, Pulkkinen, and Kokko (2007); Birch et al. (2001) Hair, Anderson, Tathan, and Black (1998)

References Acquity Group (2014). The internet of Things: The future of consumer adoption. Retrieved May 1, 2017, from https://www.accenture.com/t20150624T211456__w__/ us-en/_acnmedia/Accenture/Conversion-Assets/DotCom/Documents/Global/PDF/Technology_9/Accenture-Internet-Things.pdf. Adobe (2018). State of voice assistants. Retrieved March 13, 2019 from https://www.slideshare.net/adobe/adi-state-of-voice-assistants-113779956?ref=https://www. cmo.com/features/articles/2018/9/7/adobe-2018-consumer-voice-survey.html. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411. Argus, I. (2015, June 17). Consumer demand for connected home products slows dramatically in first half of 2015 and continues rapid drop off. Retrieved May 1, 2017, from http://www.argusinsights.com/connected-home-release/ 2015 . Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–94. Balta-Ozkan, N., Davidson, R., Bicket, M., & Whitmarsh, L. (2013a). The development of smart homes market in the UK. Energy, 60, 361–372. Balta-Ozkan, N., Davidson, R., Bicket, M., & Whitmarsh, L. (2013b). Social barriers to the adoption of smart homes. Energy Policy, 63, 363–374. Barclay, D., Higgins, C., & Thompson, R. (1995). The partial least squares (PLS) approach to causal modeling: Personal computer adoption and use as an illustration. Technology Studies, 2(2), 285–309. Baudier, P., Ammi, C., & Deboeuf-Rouchon, M. (2018). Smart home: Highly-educated students' acceptance. Technological Forecasting and Social Change (in press). Bauer, R. A. (1960). Consumer behavior as risk taking. Proceedings of the 43rd national conference of the American marketing association, June 15, 16, 17, chicago, Illinois, 1960. American Marketing Association. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238. Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88(3), 588. Berry, L. L. (1980). Services marketing is different. Business, 30(3), 24–29. Birch, L. L., Fisher, J. O., Grimm-Thomas, K., Markey, C. N., Sawyer, R., & Johnson, S. L. (2001). Confirmatory factor analysis of the child feeding questionnaire: A measure of parental attitudes, beliefs and practices about child feeding and obesity proneness. Appetite, 36(3), 201–210. Bloomberg (2016). Do consumers want a connected house, car? Retrieved May 1, 2017, from http://www.bloomberg.com/news/videos/b/8b545753-001f-419da2aa-dc9dc7db4628, Accessed date: 1 May 2017. Chan, M., Campo, E., Estève, D., & Fourniols, J.-Y. (2009). Smart homes—current features and future perspectives. Maturitas, 64(2), 90–97. Chan, M., Estève, D., Escriba, C., & Campo, E. (2008). A review of smart homes—present state and future challenges. Computer Methods and Programs in Biomedicine, 91(1), 55–81. Cisco (2017). The IoT value/trust paradox. Retrieved March 10, 2019 from https://www.jasper.com/resources/reports/iot-value-and-trust-survey?ecid=af_ 700000005. Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 319–340. Dean, S. (2017, January 13). Amazon, Samsung, Google, Apple: Big 4 driving smart home device sales. Tech times. Retrieved May 1, 2017, from http://www.techtimes. com/articles/192188/20170113/amazon-samsung-google-apple-big-4-driving-smart-home-device-sales.htm. Demiris, G., Rantz, M. J., Aud, M. A., Marek, K. D., Tyrer, H. W., Skubic, M., et al. (2004). Older adults' attitudes towards and perceptions of ‘smart home’ technologies: A pilot study. Medical Informatics and the Internet in Medicine, 29(2), 87–94. Dynatrace (2018). IoT consumer confidence report: Challenges for enterprise cloud monitoring on the horizon. Retrieved March 10, 2019 from https://assets.dynatrace.com/ en/docs/report/2824-iot-consumer-confidence-report-dynatrace.pdf. Eggert, A. (2006). Intangibility and perceived risk in online environments. Journal of Marketing Management, 22(5–6), 553–572. Ehrenhard, M., Kijl, B., & Nieuwenhuis, L. (2014). Market adoption barriers of multi-stakeholder technology: Smart homes for the aging population. Technological Forecasting and Social Change, 89, 306–315. Ellis, S., & Shpielberg, N. (2003). Organizational learning mechanisms and managers' perceived uncertainty. Human Relations, 56(10), 1233–1254. Featherman, M. S., & Pavlou, P. A. (2003). Predicting e-services adoption: A perceived risk facets perspective. International Journal of Human-Computer Studies, 59(4), 451–474. Featherman, M. S., & Wells, J. D. (2004). The intangibility of E-services: Effects on artificiality, perceived risk, and adoption. Proceedings of the 37th Annual Hawaii International Conference on System Sciences, 177–187. https://doi.org/10.1109/HICSS.2004.1265424. Featherman, M. S., & Wells, J. D. (2010). The intangibility of e-services: Effects on perceived risk and acceptance. ACM SIGMIS - Data Base, 41(2), 110–131. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 39–50. Garun, N. (2017, June 5). The 8 biggest announcements from Apple WWDC 2017. The Verge. Retrieved June 14, 2017, from https://www.theverge.com/2017/6/5/ 15722994/apple-wwdc-2017-news-highlights-recap. Greenough, J. (2016). The US smart home market has been struggling — here's how and why the market will take off. Business Insider. October 18 . Retrieved June 19, 2018 from http://www.businessinsider.com/the-us-smart-home-market-report-adoption-forecasts-top-products-and-the-cost-and-fragmentation-problems-that-couldhinder-growth-2015-9. Hair, J. F., Anderson, R. E., Tathan, R. L., & Black, W. C. (1998). Multivariate data analysis. New York: Prentice Hall. Heavey, C., & Simsek, Z. (2013). Top management compositional effects on corporate entrepreneurship: The moderating role of perceived technological uncertainty. Journal of Product Innovation Management, 30(5), 837–855. Higginbotham, S. (2015, June 17). Are consumers abandoning the smart home? Fortune. Retrieved May 1, 2017, from http://fortune.com/2015/06/17/consumerssmart-home/. Hsieh, J. S. C., Huang, Y. M., & Wu, W. C. V. (2017). Technological acceptance of LINE in flipped EFL oral training. Computers in Human Behavior, 70, 178–190. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. International Telecommunication Union (2010). Applications of ITUT G. 9960, ITU-T G. 9961 transceivers for Smart Grid applications: Advanced metering infrastructure, energy management in the home and electric vehicles. ITU-T Technical Paper. IoT Analytics (2018). State of the IoT 2018: Number of IoT devices now at 7B – market accelerating. Retrieved March 10, 2019 from https://iot-analytics.com/state-ofthe-iot-update-q1-q2-2018-number-of-iot-devices-now-7b/. Icontrol Networks (2015). 2015 state of the smart home report. Retrieved from Redwood City, CA https://www.icontrol.com/wp-content/uploads/2015/06/Smart_ Home_Report_2015.pdf. Joshi, K. (1991). A model of users' perspective on change: The case of information systems technology implementation. MIS Quarterly, 229–242. Joshi, K. (2005). Understanding user resistance and acceptance during the implementation of an order management system: A case study using the equity implementation model. Journal of Information Technology Case and Application Research, 7(1), 6–20. Kahneman, D., & Tversky, A. (2013). Prospect theory: An analysis of decision under risk. Handbook of the fundamentals of financial decision making: Part I (pp. 99–127). Hackensack, NJ: World Scientific (Chapter 6).

14

Telecommunications Policy xxx (xxxx) xxxx

A. Hong, et al.

Kang, Y., & Kim, S. (2009). Understanding user resistance to participation in multihop communications. Journal of Computer-Mediated Communication, 14(2), 328–351. Kim, H.-W., & Kankanhalli, A. (2009). Investigating user resistance to information systems implementation: A status quo bias perspective. MIS Quarterly, 567–582. Kim, D., & Kim, S. (2011). Factors influencing users' resistance to location based SNS application for smart phones. Korean Journal of Broadcasting and Telecommunication Studies, 25(3), 133–166. Kim, S.-H., & Kim, J. K. (2018). Determinants of the adoption of mobile cloud computing services: A principal-agent perspective. Information Development, 34(1), 44–63. Kleijnen, M., Lee, N., & Wetzels, M. (2009). An exploration of consumer resistance to innovation and its antecedents. Journal of Economic Psychology, 30(3), 344–357. Kline, R. B. (2015). Principles and practice of structural equation modeling. Guilford Publications. Lapointe, L., & Rivard, S. (2005). A multilevel model of resistance to information technology implementation. MIS Quarterly, 29(3). Laroche, M., Bergeron, J., & Goutaland, C. (2001). A three-dimensional scale of intangibility. Journal of Service Research, 4(1), 26–38. Laroche, M., McDougall, G. H., Bergeron, J., & Yang, Z. (2004). Exploring how intangibility affects perceived risk. Journal of Service Research, 6(4), 373–389. Laukkanen, T. (2016). Consumer adoption versus rejection decisions in seemingly similar service innovations: The case of the Internet and mobile banking. Journal of Business Research, 69(7), 2432–2439. LG Economic Research Institute. Gungnae smart home sijang goga jepumboda jeoga jepumbuteo (Domestic smart home market, Low-priced products than high-priced products). (2015). Retrieved from http://www.lgeri.com/report/view.do?idx=19046 [in Korean]. Lu, H.-P., Hsu, C.-L., & Hsu, H.-Y. (2005). An empirical study of the effect of perceived risk upon intention to use online applications. Information Management & Computer Security, 13(2), 106–120. Luor, T. T., Lu, H.-P., Yu, H., & Lu, Y. (2015). Exploring the critical quality attributes and models of smart homes. Maturitas, 82(4), 377–386. Marakas, G. M., & Hornik, S. (1996). Passive resistance misuse: Overt support and covert recalcitrance in IS implementation. European Journal of Information Systems, 5(3), 208–219. Markets and Markets (2016). Smart home market by product (lighting control, security and access control, HVAC, entertainment, smart speaker, home healthcare, smart kitchen, home appliances, and smart furniture), software and services, and region - global forecast to 2024. Retrieved from http://www.marketsandmarkets. com/PressReleases/global-smart-homes-market.asp. Markus, M. L. (1983). Power, politics, and MIS implementation. Communications of the ACM, 26(6), 430–444. Martins, C., Oliveira, T., & Popovič, A. (2014). Understanding the internet banking adoption: A unified theory of acceptance and use of technology and perceived risk application. International Journal of Information Management, 34(1), 1–13. Milliken, F. J. (1987). Three types of perceived uncertainty about the environment: State, effect, and response uncertainty. Academy of Management Review, 12(1), 133–143. Mitchell, V.-W. (1999). Consumer perceived risk: Conceptualisations and models. European Journal of Marketing, 33(1/2), 163–195. Murray, K. B., & Schlacter, J. L. (1990). The impact of services versus goods on consumers' assessment of perceived risk and variability. Journal of the Academy of Marketing Science, 18(1), 51–65. Mzoughi, N., & M'Sallem, W. (2013). Predictors of internet banking adoption: Profiling Tunisian postponers, opponents and rejectors. International Journal of Bank Marketing, 31(5), 388–408. National Information Society Agency (2016). Home IoT sijang bunseok mit sisajeom(Home IoT market analysis and implications). Retrieved from http://nia.or.kr/ site/nia_kor/ex/bbs/View.do?cbIdx=39485&bcIdx=18078&parentSeq=18078. Pavlou, P. A. (2003). Consumer acceptance of electronic commerce: Integrating trust and risk with the technology acceptance model. International Journal of Electronic Commerce, 7(3), 101–134. Pavlou, P. A., Liang, H., & Xue, Y. (2007). Understanding and mitigating uncertainty in online exchange relationships: A principal-agent perspective. MIS Quarterly, 105–136. Peine, A. (2008). Technological paradigms and complex technical systems—the case of smart homes. Research Policy, 37(3), 508–529. Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879. PwC (2017). Smart home, seamless life Unlocking a culture of convenience. Retrieved March 10, 2019 from https://www.pwc.fr/fr/assets/files/pdf/2017/01/pwcconsumer-intelligence-series-iot-connected-home.pdf. Ragatz, G. L., Handfield, R. B., & Petersen, K. J. (2002). Benefits associated with supplier integration into new product development under conditions of technology uncertainty. Journal of Business Research, 55(5), 389–400. Ram, S. (1987). A model of innovation resistance. NA-Advances in Consumer Research, 14, 208–212. Ram, S., & Sheth, J. N. (1989). Consumer resistance to innovations: The marketing problem and its solutions. Journal of Consumer Marketing, 6(2), 5–14. Rantanen, J., METSÄPELTO, R. L., Feldt, T., Pulkkinen, L. E. A., & Kokko, K. (2007). Long‐term stability in the Big Five personality traits in adulthood. Scandinavian Journal of Psychology, 48(6), 511–518. Rauschnabel, P. A., He, J., & Ro, Y. K. (2018). Antecedents to the adoption of augmented reality smart glasses: A closer look at privacy risks. Journal of Business Research, 92, 374–384. Rogers, E. (1983). Diffusion of innovations. New York, NY: Simon and Schuster. Sheth, J. N., & Stellner, W. H. (1979). Psychology of innovation resistance: The less developed concept (LDC) in diffusion research: College of commerce and business administration. Urbana-Champaign, IL: University of Illinois at Urbana-Champaign. Shin, J., Park, Y., & Lee, D. (2018). Who will be smart home users? An analysis of adoption and diffusion of smart homes. Technological Forecasting and Social Change, 134, 246–253. Song, S.-h. (2017). Samsung aims to connect all home appliances by 2020. The Korea Herald. August 22 . Retrieved June 19, 2018 from http://www.koreaherald.com/ view.php?ud=20170822000686. Song, M., & Montoya-Weiss, M. M. (2001). The effect of perceived technological uncertainty on Japanese new product development. Academy of Management Journal, 44(1), 61–80. Stock, G. N., & Tatikonda, M. V. (2008). The joint influence of technology uncertainty and interorganizational interaction on external technology integration success. Journal of Operations Management, 26(1), 65–80. Stone, R. N., & Grønhaug, K. (1993). Perceived risk: Further considerations for the marketing discipline. European Journal of Marketing, 27(3), 39–50. Szmigin, I., & Foxall, G. (1998). Three forms of innovation resistance: The case of retail payment methods. Technovation, 18(6), 459–468. Vargo, S. L., & Lusch, R. F. (2004). The four service marketing myths: Remnants of a goods-based, manufacturing model. Journal of Service Research, 6(4), 324–335. Vecchiato, R., & Roveda, C. (2010). Strategic foresight in corporate organizations: Handling the effect and response uncertainty of technology and social drivers of change. Technological Forecasting and Social Change, 77(9), 1527–1539. Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 425–478. Wang, H. C., & Chiu, Y. F. (2011). Assessing e-learning 2.0 system success. Computers & Education, 57(2), 1790–1800. Wheaton, B., Muthen, B., Alwin, D. F., & Summers, G. F. (1977). Assessing reliability and stability in panel models. Sociological Methodology, 8, 84–136. Yang, Y., Liu, Y., Li, H., & Yu, B. (2015). Understanding perceived risks in mobile payment acceptance. Industrial Management & Data Systems, 115(2), 253–269. Yang, H., Lee, H., & Zo, H. (2017). User acceptance of smart home services: An extension of the theory of planned behavior. Industrial Management & Data Systems, 117(1), 68–89. Yeh, J.-C., Hsiao, K.-L., & Yang, W.-N. (2012). A study of purchasing behavior in Taiwan's online auction websites: Effects of uncertainty and gender differences. Internet Research, 22(1), 98–115. Zeithaml, V. A., Parasuraman, A., & Berry, L. L. (1985). Problems and strategies in services marketing. Journal of Marketing, 49(2), 33–46.

15