Intimacy, familiarity and continuance intention: An extended expectation–confirmation model in web-based services

Electronic Commerce Research and Applications 10 (2011) 342–357

Contents lists available at ScienceDirect

Electronic Commerce Research and Applications journal homepage: www.elsevier.com/locate/ecra

Intimacy, familiarity and continuance intention: An extended expectation–confirmation model in web-based services Yonnim Lee a,1, Ohbyung Kwon b,⇑ a

School of International Management, Kyung Hee University, Yongin, Republic of Korea College of Management, Kyung Hee University, Seoul, Korea and Visiting Scholar Department of Information and Decision Systems, San Diego State University, San Diego, United States b

a r t i c l e

i n f o

Article history: Received 1 March 2010 Received in revised form 19 November 2010 Accepted 19 November 2010 Available online 25 November 2010 Keywords: Affective factors Behavioral research Continuance intention Expectation–confirmation model Familiarity Intimacy Web-based services

a b s t r a c t To date, plenty of theories, such as the expectation–confirmation model (ECM), have been proposed to explain why and how consumers are motivated to continue to use web-based services. In particular, various affective factors have been proposed to explain user satisfaction and continued use of web-based services recently in the IS community. In IS continuance research, several affective factors, such as perceived playfulness, perceived enjoyment and pleasure, have been examined. Affective factors discussed in the existing continuance intention-related studies are mostly short-term emotional factors like this. However, if a user’s continued usage of a web-based service can be interpreted as a long-term relationship between a user and the service, then the factors such as familiarity and intimacy which are the emotions created accumulatively over time based on an established relationship with the user can be helpful for better explaining the user’s continuance intention. Also, if relationships between consumers and web-based services have been built up due to repetitive usage, then we can assume that both affective and cognitive factors may explain consumers’ continuance intention. Hence, the purpose of this paper is to propose an extended ECM. We focus on two new constructs, familiarity and intimacy, as persistent affective factors. To investigate how cognitive and affective factors are interrelated in continuance intention, we conducted surveys focusing on users’ continued intention to use web-based services. The results indicate that continuance intention is affected conjointly by cognitive factors, such as perceived usefulness, and affective factors, such as familiarity and intimacy. However, the effects of affective factors such as intimacy were larger than those of cognitive factors such as perceived usefulness. In addition, the results indicate that intimacy, a purer affective concept than familiarity, affects users’ continuance intention more than familiarity. Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction Web-based services have been used by a growing number of consumers for more than a decade. Web services in virtually every area of business—banking, travel, social activities, and more—have continued their rapid spread, gaining increasing importance in the lives of a wider range of the population. However, many companies are facing fierce competition due to the evolution and proliferation of web-based services; they are struggling to find strategies to retain their existing customers to ensure the company’s success and overall sustainability. Also, web-based services have low entry barriers by its nature, if one service is created, then a number of comparable alternative web-based services follow, resulting in a high switching rate between those services by users (Vatanasombut et al. 2008). In fact, to date, a tremendous number of web-based services have appeared and disappeared. Thus, web-based service ⇑ Corresponding author. Tel.: +82 2 961 2148. 1

E-mail addresses: [email protected] (Y. Lee), [email protected] (O. Kwon). Tel.: +82 31 201 2309.

1567-4223/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.elerap.2010.11.005

providers are very eager to identify the conditions that lead to long-running web-based services. Accordingly, continuance intention has become an important subject of study in the information system (IS) research area. IS continuance intention describes the user’s decision to continue to use a specific IS that he or she has already been using. This is different from the user’s first-time usage of the IS. IS continuance is crucial because the long-term viability of an IS and its eventual success depend on its continued use rather than first-time use (Karahanna et al. 1999, Bhattacherjee 2001a). IS continuance at an individual level has been regarded as crucial for sustainable web-based services such as online retailers, banks, and travel agencies. Continuance intention has been understood based on the expectancy-confirmation model (ECM). In this model, users’ continuance intention is decided by their satisfaction with IS usage and perceived usefulness of continued IS usage. Other researchers have extended the ECM by incorporating new factors. In the existing ECM-related studies, factors influencing users’ continuance intention have been suggested from two perspectives. First, we see several cognitive factors dominating the studies: perceived

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

usefulness, perceived quality, perceived usability, subjective norm, subjective task value, loyalty incentives, and others. Second, a few studies have supported the affective perspective relating to continuance intention in the ECM (Lin et al. 2005). Even though this perspective has received less attention than the cognitive factors, some affective factors such as perceived enjoyment, pleasure, and arousal have recently been proposed as major predictors of corresponding research models such as the ECM (Lin et al. 2005, Thong et al. 2006, Min and Shenghua 2007). However, these factors are relatively temporary and change easily according to the context; they may not be adequate to explain continuance intention behavior fully, especially behavior that lasts for a longer time. Hence, there must be other, relatively more persistent affective factors at work. Recently, researchers in psychology have confirmed that familiarity and intimacy are emotions that develop cumulatively over time, formed quite persistently and differently from short-term affective factors. For example, Gobbini et al. (2004) argued that familiarity accrues naturally with years of social interaction, and James (1992) showed that familiarity has to have time to grow. In addition, Cordova and Scott (2001) insisted that intimacy is not based on a single event, but instead on an accumulation of events over time. Bagarozzi (1997) explained that intimacy evolves over time. These findings are compatible with the notion that familiarity and intimacy are potentially related to continuance intention. However, in the IS community, this theory has not been empirically tested yet. Affective factors such as familiarity and intimacy have been regarded as essential for strengthening human relationships in consumer behavior. If relationships between consumers and web-based services develop due to repetitive usage, then we can assume that both affective and cognitive factors may contribute to explain consumers’ continued use of web-based services. We intend to examine whether affective factors, which are created accumulatively over time, based on a prior relationship with the user and related to social bonds, could influence their continuance intention. Moreover, comparing perceived usefulness as a cognitive factor and familiarity and intimacy as affective factors in terms of their explanatory power for continuance intention is an interesting challenge, since perceived usefulness has long been regarded as a representative determinant of user satisfaction and continuance intention. Hence, the purpose of this paper is to propose an extended ECM including two affective factors—familiarity and intimacy—to inform our understanding of continuance intention as it is related to web-based services. We focus on the fact that the continued use of such services is based on the continuous relationship between service or service provider and user. We introduce human relationship studies suggesting the affective factors related to relationship reinforcement among people, and extend Bhattacherjee’s (2001b) ECM model, the most representative research model on continuance intention. In the extended ECM, perceived usefulness is also considered as a competing factor, allowing us to analyze the explanatory power of familiarity and intimacy. The remainder of the paper proceeds as follows. The next section presents the theoretical background. The detailed research model and hypotheses are then described in Section 3. In Section 4, we present the research methodology of our empirical study. The results are then described in Section 5. Finally, in Sections 6 and 7, we provide discussion and then conclude the paper with a summary of future research plans.

2. Research background 2.1. Expectation–confirmation theory Introduced by Bhattacherjee in 2001, the ECM is an IS-related early research model that has its origin in expectation–confirma-

343

tion theory (ECT). Originally this theory mainly was used to investigate consumer relationship satisfaction and repeat decisions in the consumer behavior literature (Oliver 1980, 1993). Bhattacherjee, however, focused on the congruence between individuals’ continued IS usage decisions and consumers’ repeat purchase decisions, and suggested the ECM in the IT literature to explain the idea of customers’ continuance intention. Bhattacherjee carried out an empirical verification of the ECM using a survey of online banking users to explain the process through which users develop continuance intention. The empirical analytical results were as follows: users’ continuance intention was decided by their satisfaction with IS usage and their perceived usefulness of continued IS usage. Bhattacherjee’s is a landmark study in that it brought attention to the differences between the behavior of a user accepting an IS, versus the behavior of trying to continue using it. It was an early theoretical study on the continued usage of IS, and brought to light new factors such as expectation fulfillment and user satisfaction. We chose the ECM as the theoretical apparatus in our research for the following reasons. First, Bhattacherjee’s was a representative study whose results were verified by several other studies (Lin et al. 2005, Thong et al. 2006, Min and Shenghua 2007, Limayem and Cheung 2008, Vatanasombut et al. 2008). Second, the ECM has been widely used to study post-adoption behavior in various web-based service contexts, such as online banking and e-learning (Roca et al. 2006, Min and Shenghua 2007, Chiu and Wang 2008, Limayem and Cheung 2008). In particular, applying the ECM framework was found to be appropriate even within the e-commerce context that is the empirical analytical context in this paper (Atchariyachanvanich et al. 2006, Liao et al. 2007). 2.2. Continuance intention The concept of continuance intention has been variously described in the literature as ‘‘incorporation’’ (Kwon and Zmud 1987), ‘‘routinization’’ (Cooper and Zmud 1990), and ‘‘confirmation’’ (Rogers 1995). Despite these variations in terminology, however, studies agree that continuance behavior assumes institutionalizing IS use as normal, ongoing activity. Hence, continuance intention may be defined as continued usage of IS by adopters, where a continuance decision follows an initial acceptance decision (Bhattacherjee 2001b). Post-ECM, numerous studies followed to explain users’ continuance behavior in web-based, service domains. They are classified into two types: extended studies based on the ECM, but adding a new construct, and integrated studies that combine the ECM and another theory or model. For example, Lin et al. (2005) extended the ECM model by adding an additional relationship between perceived playfulness and satisfaction. Thong et al. (2006) expanded the ECM’s set of post-adoption beliefs to extend the model’s application beyond an instrumental focus. Limayem and Cheung (2008) extended Bhattacherjee’s model by considering a moderating effect (IS habit) on continuance intention. Atchariyachanvanich et al. (2006) added a new factor, customer loyalty. Min and Shenghua (2007) combined the ECM with theory about perceived enjoyment of individual users. Other studies attempted to integrate theories or models to explain users’ IT continuance intention. Vatanasombut et al. (2008) combined Commitment-Trust Theory, the ECM, and the Technology Acceptance Model (TAM) to develop a model of continuance intention. They showed that relationship commitment and trust are stronger predictors of continuance intention. Roca et al. (2006) decomposed the TAM based on Expectancy Disconfirmation Theory (EDT). In the model, the perceived performance component is divided into perceived quality and perceived usability. Liao et al. (2007) developed an integrated model designed to predict and explain an individual’s continued use of online services based on

344

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

the concepts of EDT and the Theory of Planning Behavior (TPB). They suggested that a customer’s behavioral intention towards e-service continuance is mainly determined by his or her satisfaction, and additionally is affected by perceived usefulness and the subjective norm. Above and beyond these studies, Chiu and Wang (2008) adapted the United Theory of Acceptance and Use of Technology (Venkatesh et al. 2003). Table 1 summarizes the previous research models on continuance intention with their research domains. The factors suggested in the continuance intention studies so far can be largely classified into two categories: cognitive and affective. Cognitive factors are those related to the mental process of knowing, including aspects such as perception, reasoning and judgment. Cognitive factors command a majority of those factors found to affect IT service continued usage in prior studies. Representative cognitive factors suggested are: perceived usefulness, perceived ease of use, perceived value, security, confirmation, disconfirmation, perceived behavioral control, perceived risk, perceived switching cost, and perceived usability. In contrast, affective factors are related to specific emotions or states of feeling. We see that, more recently, continuance intention research has shifted its focus from cognition-oriented factors to affective factors. In fact, several affective factors have been examined in recent continuance intention studies. For instance, during the last five years, as shown in Table 2, perceived playfulness, perceived enjoyment, pleasure, and arousal have been proposed as major affective constructs of the continuance intention model. However, those affective factors are relatively short-term and temporary feelings. Therefore, to explain users’ continuance intention after the initial period of use is over, which is the purpose of this study, we need to find new affective factors created accumulatively over time based on a prior relationship with the user. This study identifies two main factors to explain continuance intention behavior better: intimacy and familiarity. 2.3. Affective factors: intimacy and familiarity The difficulty in attempting to construct definitions is that although ‘‘everybody knows’’ what an intimate relationship means, Table 1 Summary of previous research on continuance intention. Study

Characteristics of the research model

Research domain

Lin et al. (2005)

Based on ECM; introduces perceived playfulness as a new factor Based on ECM; introduces postadoption beliefs and perceived enjoyment as new factors Based on ECM; adds IS habit as a new factor Based on ECM; adds customer loyalty as a new factor Based on ECM; adds perceived enjoyment as a new factor An integrated study that combines Expectancy Disconfirmation Theory and the Technology Acceptance Model; adds perceived quality and perceived usability as new factors An integrated study that combines Expectancy Disconfirmation Theory and the Theory Of Planning Behavior; adds subjective norm as a new factor An integrated study that combines United Theory Of Acceptance And Use Of Technology; adds subjective task value as a new factor

Web-portal

Thong et al. (2006)

Limayem and Cheung (2008) Atchariyachanvanich et al. (2006) Min and Shenghua (2007) Roca et al. (2006)

Liao et al. (2007)

Chiu and Wang (2008)

Internet service Internet-based learning E-commerce E-learning E-learning

E-service

Web-based learning

Table 2 Summary of previous affective constructs. No.

Affective construct

Definition

Major references

1

Perceived playfulness (enjoyment)

The extent to which the activity of using the IT is perceived to be enjoyable in its own right, apart from any performance consequences that may be anticipated

2

Pleasure

3

Arousal

The degree to which a user feels good or happy with the target object The degree to which the user feels excited, stimulated or active

Hsu and Chiu (2004) Lin et al. (2005) Thong et al. (2006) Kim et al. (2007) Kim et al. (2007)

there is little agreement as to what ‘‘it’’ is (Chelune and Waring 1984). The word ‘‘intimacy’’ is used to describe a specific type of feeling (Cordova and Scott 2001). The nature of intimacy requires a sense of closeness beyond ‘‘embodiment of commodities,’’ something emotional that is potentially profound, something that is ‘‘real’’ rather than superficial, requiring enduring involvement rather than purely situational involvement, and a commitment to wanting to identify with the other (Trauer and Ryan 2005). Tolstedt and Stokes (1983) defined intimacy as feelings of closeness and emotional bonding, involving intense liking, moral support, and the ability to tolerate flaws in the significant other. In this study, we adopt Tolstedt and Stokes’s (1983) definition, which emphasizes the aspect of feelings of emotional closeness in its definition of intimacy. Ecker et al. (2007) defined familiarity as a general feeling of having encountered a person or specific object before, without conscious access to contextual details, such as the time or place of the encounter. At this point, familiarity can be regarded as an emotional term because using the term ‘‘feeling’’ suggests a relation to emotion (Kelley and Jacoby 1998). Ramachandran (1998) and Ratcliffe (2002) use the phrase ‘‘feeling of familiarity’’’ to describe familiarity as an affective concept. They explain it as an emotional sense associated with a known subject. Familiarity has been defined as both a cognitive concept and an emotional (affective) concept at the same time. However, we adopted Ecker et al.’s (2007) definition, which considers familiarity as an affective concept, and proposed familiarity as a new affective factor influencing users’ continuance intention. Additionally, there is a relationship between familiarity and intimacy; the nature of intimacy involves an individual’s deepening familiarity with a subject (Jourard 1971, Altman and Taylor 1973). In fact, familiarity is one of the prerequisites to intimacy (Williams 2001). That is, there are precedence relations between familiarity and intimacy; however, these two concepts are clearly different from each other and have been used concurrently but distinctly in many theses. For instance, Lam and Mizerski (2005) measured degrees of familiarity and intimacy conjointly to characterize in-group relationships, while Zhao (2002) used familiarity and intimacy simultaneously as indicators to measure tie strength. Oatley and Jenkins (1996) conceptualized the relationships among different aspects of emotion on a continuum. The organizing features of this continuum are (a) whether the emotion is elicited by a specific event and (b) the length of time that the emotional experience lasts. Applying the standards they proposed, we see that affective factors such as perceived playfulness, perceived enjoyment, pleasure, and arousal in the existing continuance intention model correspond with feelings created through the single use of a web service. These factors bring about short-term emotions that do not last for long. On the other hand, many references confirm that familiarity and intimacy are emotions created accumulatively over time, not from a single specific event, and that they

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

are different from short-term emotions. For example, Gobbini et al. (2004) argued that familiarity accrues naturally with years of social interaction, and James (1992) showed that familiarity has to have time to grow. Cordova and Scott (2001) insisted that intimacy is not based on a single event, but instead on an accumulation of events over time, and Bagarozzi (1997) explained that intimacy evolves over time. Most social psychological researchers also define intimacy as a quality of interaction and relationships between people (Acitelli and Duck 1987, Laurenceau et al. 1998, Timmerman 1991). In addition, familiarity is affected by the quantity of prior interactions and occurs through repeated interaction (Rindfleisch and Inman 1998). In particular, intimacy has been regarded as an essential aspect of interpersonal relationships, and of maintaining relationships. Lowenthal and Haven (1968) introduced intimacy as a critical variable for interaction and adaptation. Moss and Schwebel (1993) reported that intimacy is of central importance in enduring relationships. Schaefer and Olson (1981) also argued that intimacy plays an integral role in the solidification of relationships. Similar to intimacy, a positive relationship between familiarity and long-term relationships has been identified in many prior studies. Several researchers of consumer behavior have noted positive associations between affective familiarity and customers’ reuse behavior through good feelings such as liking (Colombo and Morrison 1989, Raj 1985). Zhao (2002), based on human-relationship theory in contexts such as social networks, proposed familiarity as an indicator to measure tie strength, arguing that familiarity influences the formation of long-term relationships and showing that strong ties existed among the high-familiarity group. The factors familiarity and intimacy are helpful for explaining the user’s intention to continue using web-based services, since they are the emotions created accumulatively over time based on an established relationship with the user. Unlike the existing continuance intention-related studies, which discussed only shortterm emotional factors during use of the service, this study extends the legacy of the ECM using intimacy and familiarity as factors that have proved to be particularly useful in strengthening long-term relationships in human relationship-related theories. For our purposes, a user’s continued usage of a web-based service can be interpreted as a long-term relationship with that service.

3. Hypotheses and research model We propose an extended ECM that contains two new affective factors—intimacy and familiarity—to explain why consumers continue to use a specific web-based service. The factors are created accumulatively over time, based on a prior relationship with the user, and related to maintaining a relationship between consumers and service providers. Fig. 1 shows an overview of the research model proposed in this paper. We examine the model within the context of online shopping services, one of the most representative web-based services. 3.1. Intimacy Intimacy is defined as feelings of closeness and emotional bonding, involving intense liking, moral support, and the ability to tolerate flaws in the significant other (Tolstedt and Stokes 1983). It has been described as a critical variable for interaction and adaptation (Lowenthal and Haven 1968). We took a particular interest in the role that intimacy plays in continuance intention, comparing it to the way that intimacy plays an integral role in solidifying human relationships (Schaefer and Olson 1981). It is considered to be an essential aspect of establishing and maintaining interpersonal relationships. We first examine the relationship between intimacy and confirmation. At the adoption stage, where the relationship between the user and the web-based service begins to build, self-disclosure begins in order to solicit the hoped-for ‘‘supportive response,’’ which entails a certain amount of risk on the part of the user (Duck 1988). After using the web-based service, if the outcome is uncertain and perhaps does not mesh with the user’s expectations, the user’s tension level increases and he or she becomes reluctant to engage in any kind of self-disclosure again. On the other hand, if the interaction is perceived as confirming, the anticipation that positive interactions will be repeated in the future encourages the building of an intimate relationship. Such causality between cognitive beliefs and affect is well established in the social psychology literature (Ajzen 1991), and has been validated across a wide range of IS use behaviors (Davis et al. 1989, Mathieson 1991, Taylor and Todd 1995). Based on these arguments, we anticipate that confirmation in

Perceived Usefulness H10 H8

H9

H7

H6

Satisfaction

Confirmation

345

H5 H3 H1

Familiarity H2 H4

Intimacy

Fig. 1. Research model.

Continuance Intention

346

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

web-based services has a positive influence on users’ intimacy with those services. Hypothesis 1. Intimacy with web-based services is influenced by confirmation during use of those services. Intimacy has been regarded not only as an important measure of tie strength (Granovetter 1982, Marsden and Campbell 1984), but also as an essential part of tie formation (Wellman 2001). Although we found a variety of conceptualizations of intimacy in the literature, it has been commonly treated as a tool for assessing closeness of relationships or as an essential factor for maintaining relationships. As the importance of relationships in marketing has received considerable research attention, several studies have focused on intimacy between service provider and consumer (Reis and Shaver 1988, Stern 1997). These studies argue that intimacy can be successful in strengthening and improving relationships between parties. If service providers can invoke feelings of connection and intimacy—where consumers feel understood, cared for, and validated—a stronger bond and sense of loyalty is likely to result from marketing activities such as advertising. Several studies have reported the positive effect of intimacy on relationships. Grayson and Ambler (1999) showed that perceived intimacy or quality of interaction with an advertising agency affected clients’ use of services in the long term. Jap and Ganesan (2000) found that intimacy has a positive effect on relationship continuance in marketing channels. Verhoef et al. (2002) examined the effect of commitment-based intimacy on consumer referrals and the number of services purchased. Finally, in IS research, Bickmore and Picard (2005) studied the establishment and maintenance of long-term human–computer relationships, and argued that the relationships could be better maintained when emotional support in the form of intimacy is fostered. Rau et al.’s (2008) study took a close look at how the intimacy level in online social network services (SNS) affected the frequency of posting, and how lurkers and posters in SNS differ in intimacy levels. Their result shows that intimacy has a positive influence on posting frequency, and that social–emotional factors have a significant impact on people’s continued posting behavior on SNS. Based on these arguments, we set up a hypothesis as follows: Hypothesis 2. Intimacy with web-based services has a positive effect on users’ intention to continue using those services. 3.2. Familiarity Familiarity is usually defined as ‘‘one’s understanding of an entity, often based on previous interactions, experience, and learning of the what, who, how, and when of what is happening’’ (Gefen 2000, Komiak and Benbasat 2006). However, this study aims at observing familiarity as an affective factor, as mentioned above, defining it as a feeling, that is, how much a user feels familiar with web-based services, and the feelings pertain to personal experience. Confirmation, a cognitive belief representing the extent to which consumers’ ex ante expectations of service use were met in reality, refers to an evaluation process. It is the outcome of a rational process of comparing initial expectations with actual experience. We can see here that confirmation is a variable that inherently captures actual experience. Thus, considering that familiarity is basically a feeling created through previous experience, the relationship between familiarity and confirmation is expected to be positive. Sørebø and Eikebrokk (2008) argued, based on cognitive dissonance theory (Festinger 1957), that the degree to which users experience confirmation is positively associated with their perceived ease of use. Here ‘‘perceived ease of use’’ is a measurement of how easy the user feels it is to use the service; the concept is

directly related to users’ degree of understanding of the IS in question. So, the relationship between familiarity, meaning understanding for each specific individual, and confirmation, is again expected to be positive. Yet here familiarity can be considered to be a cognitive factor rather than an affective factor, and recalling that there is a positive relationship between cognitive familiarity and affective familiarity, there is again expected to be a positive relationship between familiarity and confirmation. Such causality between cognitive beliefs and affect is well established in the social psychology literature (Ajzen 1991), and has been validated across a wide range of IS use behaviors (Davis et al. 1989, Mathieson 1991, Taylor and Todd 1995). Based on these arguments, we anticipate that confirmation within web-based services has a positive influence on users’ familiarity with those services. Hypothesis 3. Familiarity with web-based services is influenced by the confirmation users experience during use of those services. Psychologists have long observed that repeated exposure to a stimulus results in an increase in positive affect (Bornstein 1989, Maslow 1937, Matlin 1971, Zajonc 1980, Zajonc and Helen 1982). A review by Bornstein (1989) documented mere exposure effects on a range of affective responses (liking, attractiveness, pleasantness, goodness) to a range of stimuli (nonsense words, polygons, line drawings, abstract paintings), including social stimuli. Many social psychologists have long recognized the strong link between familiarity and liking (Sherif et al. 1961, Aronson 1978). The observation that well-known brands appear to be better liked than less familiar brands has been documented in the marketing literature since the late 1960s (Shuchman 1968). Therefore, as intimacy indicates a deeper emotional state, a deeper relationship than familiarity, we expect that familiarity with a web-based service will have a positive effect on users’ intimacy with that service. Hypothesis 4. Familiarity with a web-based service will have a positive effect on users’ intimacy with the web-based service. A positive relationship between familiarity and long-term relationships has been observed in many studies. Several researchers of consumer behavior noted positive associations between affective familiarity and customers’ reuse behavior in the form of good feelings such as liking (Colombo and Morrison 1989, Raj 1985). Zhao (2002) developed a human-relationship theory in the social network context, proposing familiarity as an indicator to measure tie strength and arguing that it influenced the formation of longterm relationships by showing that ties were stronger among the high-familiarity group. Moreover, in Söderlund’s (2002) study, it was also found that a high level of pre-purchase familiarity was associated with more extreme post-purchase responses in terms of repurchase intention compared to a low level of pre-purchase familiarity. Based on these arguments, we anticipate that affective familiarity with web-based services has a positive influence on customers’ web-based services continuance intention. Hypothesis 5. Familiarity with a web-based service has a positive effect on users’ intention to continue using that service. 3.3. Hypotheses related to ECM Bhattacherjee (2001a) employed ECM to examine online banking service continuance, and showed that the ECM was also applicable in a web context. Bhattacherjee empirically tested his model via a field survey of online brokerage users. Online brokerage is a B2C e-commerce service that allows individual investors to buy or sell securities over the Internet at commissions substantially lower than those of full service brokerages. Several other studies

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

have extended the ECM to explain customers’ intention to continue using web-based services such as e-commerce web sites. For example, Lin et al. (2005) investigated the value of including playfulness in the ECM when studying continued use of a web site. Atchariyachanvanich et al. (2006) provided a theoretical model addressing technical, business, and consumer issues involved in online repurchasing, and expanding the existing customer satisfaction/continuance model based on the ECM. Because the ECM is an already verified model in various web-based service areas, the hypotheses included in the ECM will be treated identically in this study, with e-commerce as the subject domain. Satisfaction is generally recognized as ‘‘a positive affective state resulting from a global evaluation of performance based on past purchasing and consumption experience’’ (Anderson and Fornell 1994, Lam et al. 2004, Szymanski and Henard 2001). According to the ECM, confirmation of initial expectations of a web-based service leads to subsequent satisfaction, while the reverse leads to dissatisfaction and discontinuance intention. Hence, we expect that the extent to which users experience confirmation will have a positive influence on their satisfaction with a web-based service.

347

service use has positive influences on their satisfaction with the web-based service use. Hypothesis 9. The perceived usefulness of a web-based service has a positive effect on users’ satisfaction with the web-based service. In addition, perceived usefulness has been regarded as positively related to usage (Venkatesh and Davis 2000, Venkatesh 2000). IT adoption studies have consistently found that perceived usefulness is one of the most important determinants of users’ adoption intention (Davis 1989, Taylor and Todd 1995, Venkatesh and Davis 2000). The ECM studies also demonstrate that users’ perceived usefulness of IT has a positive effect on their intention to continue to use the IT. Hence, we expect that there is a positive causal relationship between users’ web-based service continuance intention and the perceived usefulness of the web-based service. Hypothesis 10. The perceived usefulness of a web-based service has a positive effect on users’ intention to continue using that service.

4. Survey Hypothesis 6. The extent to which users experience confirmation has a positive effect on their satisfaction with a web-based service. Studies using the ECM have also suggested that users’ satisfaction with IT has a positive effect on their continuance intention. Moreover, satisfaction is one of the most important concepts in the marketing of both services and products. Bolton and Lemon (1999) empirically demonstrated that customers with higher levels of satisfaction have higher usage than customers with lower levels of satisfaction. Danaher and Rust (1996) found empirical evidence that a customer who is satisfied with a service will have higher subsequent use. Szymanski and Henard (2001) also suggested that the overall satisfaction experienced by online customers reduced the perceived benefits of switching to other e-retailers. By building upon this line of argument, we expect that users’ level of satisfaction with their initial period of use of a web-based service has a positive influence on their web-based service continuance intention. Hypothesis 7. Users’ level of satisfaction with the initial period of use of a web-based service has a positive effect on their web-based service continuance intention. Perceived usefulness (PU) is defined as ‘‘the degree to which a user believes that use of the system will result in benefits being accrued to the user or the user’s organization, and often includes increases to job performance and productivity’’ (Davis 1989). Bhattacherjee’s ECM posits that confirmation of users’ expectations will have a positive effect on their perceived usefulness of IT. Perceived usefulness of IT could be affected by the confirmation experience, particularly when the users’ initial perceived usefulness is not concrete due to uncertainty over what to expect from IT usage. As noted previously, Bhattacherjee’s results can be applied to a web context. Hence, the extent to which users experience confirmation might have a positive influence on how useful they perceive a web-based service to be. Hypothesis 8. The extent to which users experience confirmation has a positive effect on the perceived usefulness of a web-based service. According to the ECM, users’ satisfaction with IT is determined by its perceived usefulness. The ECM posits that users’ perceived usefulness of IT has a positive effect on their satisfaction with it. PU works as a baseline for reference as users look for confirmation. Hence, we assume that users’ perceived usefulness of a web-based

4.1. Data collection and samples Our study was conducted within the context of an online shopping service. We chose the online store service over other web-based services for two main reasons. First, online store usage experiences tend to be longer-term than other web-based services such as social networking or blogging; hence continuance intention behavior can be better observed. We judged the online shopping service to be appropriate for verifying user-dependent factors, and more concentrated since there are no major differences in function and quality among online shopping services provided by reputable vendors. Findings from online shopping cases are also likely to provide insight into the continuance intention of users of other web-based services. We used a survey to collect data. The survey was conducted by commission through a specialized survey organization. With this professional organization, experienced in conducting many other surveys and controlling a survey pool of subjects, it was convenient to contact the target survey subjects, worth the extra expense. However, for fear of obscuring the intent, purpose, etc. of the questionnaire through a third-party survey, we communicated the following instructions to the survey organization. First, as regards population, we asked for a random sampling with a similar ratio between males and females. Second, after explaining that the main purpose of this study is research on the relationship between affective factors and continuance intention, efforts were made to prevent confusion and misunderstanding by expounding on the meanings of familiarity and intimacy, the core constructs of this study, as emotional factors. We especially requested that an explanation be provided to the subjects of the meanings of familiarity and intimacy as emotional factors in conducting the actual survey; these definitions were also on the first page of the questionnaire so that the subjects might read them before responding to the questions. Additionally, the survey was implemented offline through questionnaires, and all respondents voluntarily participated in the survey. As a result, we received 473 completed surveys out of 1000 subjects who were approached (return rate = 47.3%). Among the completed surveys, 53 were excluded due to missing values. The 420 remaining questionnaires were used as the subject of analysis. As for the composition of the sample, the gender distribution was fairly equal with 50.4% males and 49.6% females. In terms of age, 73.8% were in their 20s and 30s (31.3% 20s; 42.5% 30s). In

348

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

total, 75% of respondents were employed, typically indicating some level of disposable income, with the distribution as follows: clerical Table 3 Sample profile. Gender

Male Female

50.4% 49.6%

Age

Under 20 20–30 30–40 50–60 Over 60

1.3% 31.3% 42.5% 22.5% 2.4%

Marital status

Married Not married

53.8% 46.2%

Occupation

Student Clerical workers Housewife Professional Etc.

7.1% 65.0% 12.1% 10.0% 5.8%

Comfortable with shopping online

Extremely uncomfortable Very uncomfortable Somewhat uncomfortable Neither comfortable nor uncomfortable Somewhat comfortable Very comfortable Extremely comfortable

0.0% 2.6% 2.9% 22.1% 33.3% 28.3% 10.8%

workers 65.0%, housewives 12.1%, and professionals 10.0%. Most (72.4%) of the respondents were comfortable with shopping online. Hence, the responses seem qualified for analysis of the factors affecting continuance intention in the context of a web-based shopping service. Table 3 offers more detail about the demographics of the respondents. 4.2. Operational definitions and measurement We define web-based services as those services based on the web, such as online banking, e-learning, and online shopping. The operational definition, measurement items, and related literature for each variable in this study are shown in Table 4. The items were slightly modified to fit the context of an online shopping service. More specifically, the measurement items continuance intention, confirmation, perceived usefulness and satisfaction were brought almost intact from previous ECM research such as Bhattacherjee (2001a) and Tsai and Huang (2007), which had focused on other IT services. The measurement items were adjusted by specifying the online shopping mall service that is the focus of this study instead of other IT services. We also added the factors familiarity and intimacy to the ECM. A measurement of familiarity was developed by modifying questions from previous studies so that they would be suitable to the online shopping service context and

Table 4 Measurement items of research variables. Construct

Measurement Items

Continued intention to use

Continued usage intention of IS by adopters, where a continuance decision follows an initial acceptance decisiona CU1. I will continue to use the relevant online shopping mall in the future CU2. I have an intention to continue to use the relevant online shopping mall rather than another alternative shopping mall CU3. I will consider the relevant online shopping mall preferentially when purchasing goods online

Confirmation

The extent to which users perceive their initial expectations of a service to be confirmed or disconfirmed during actual usea CONF1. My experience of the relevant online shopping mall was better than I’d expected

References Bhattacherjee (2001a) Tsai and Huang (2007) Bhattacherjee (2001a)

CONF2. The service level provided by the relevant online shopping mall was better than I’d expected CONF3. My expectation for using the relevant online shopping mall was satisfied as a whole CONF4. The shopping experience in the relevant online shopping mall satisfied my expectation as a whole Satisfaction

A positive affective state resulting from a global evaluation of performance based on past purchasing and consumption experiencea SAT1. I was satisfied with my decision to use the relevant online shopping mall SAT2. It was good on the whole to go shopping at the relevant online shopping mall

Bhattacherjee (2001a) Tsai and Huang (2007)

SAT3. My decision to purchase goods at the relevant online shopping mall was wise SAT4. In general, I am satisfied with the service and the product provided by the relevant online shopping mall Perceived usefulness

The degree to which a user believes that use of the system will result in benefits being accrued to the user or the user’s organization, and often includes increases in job performance and productivitya PU1. The relevant online shopping mall enables me to purchase the product I need in a shorter time PU2. The relevant online shopping mall provides me with useful information in purchasing the product I need PU3. The relevant online shopping mall is useful on the whole in helping me purchase the product I need

Familiarity

Intimacy

Feeling of the understanding of an entity, often based on previous interactions, experience, and learning of the what, who, how, and when of what is happeninga FAM1. I feel familiar with purchasing goods at the relevant online shopping mall FAM2. I feel familiar with the interface of the relevant online shopping mall FAM3. I feel familiar with the terms used (e.g. menu name, service name, etc.) in the relevant online shopping mall FAM4. I feel familiar with the relevant online shopping mall Feelings of closeness and emotional bonding, involving intense liking, moral support, and the ability to tolerate flaws in the servicea INT1. I enjoy my time at the relevant online shopping mall and feel well at ease INT2. INT3. INT4. INT5.

a

Operational definition.

I I I I

think of the relevant online shopping mall as a friend of mine choose the relevant online shopping mall without any hesitation when purchasing goods on the Internet feel a sense of intimacy with the relevant online shopping mall feel purchasing goods at the relevant online shopping mall is a very important part of my consumption life

Bhattacherjee (2001a) Lin et al. (2005)

Gefen (2000) Gefen et al. (2003)

Chelune and Waring (1984) Tomasi (2007)

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

the emotional aspect of familiarity would be emphasized, based on Gefen et al.’s (2003) study. In addition, an intimacy instrument was developed by adding new questions asking about users’ degree of intimate feeling. These were based on the operational definition of intimacy defined in this study; the questions were based on the suggested measurement questions in Tomasi’s study and modified in accordance with the context of this study (Tomasi 2007). Based on the initial measurement indices, we conducted a focus group interview of IS usage-related researchers, where we discussed online shopping users and collected additional opinions about important variables, composition of questions, and measurement methods. After improper or ambiguous questions were eliminated or modified through this prior investigation process, the final questionnaire was completed. All items were measured on a 7-point Likert scale, with anchors from ‘‘Strongly disagree’’ to ‘‘Strongly agree.’’ Table 4 outlines the measurement items of the research variables used in this study. ‘‘Relevant online shopping mall’’ means the online shopping mall which the user has begun to use most recently. This definition was especially intended to minimize the bias caused by the positive effect that can occur from using an online shopping mall over the long term. Respondents were also instructed to choose a general shopping mall that deals in diverse items, not a particular-theme shopping mall, in order to minimize the income, gender and occupation bias that might result. For reference, online shopping malls can be categorized into overall and specialty types according to the kinds of products they carry. The former refers to a business organization that sells diverse products online through an Internet shopping mall, while the latter means a company that specifically sells a particular product such as books, computer goods, and so on. In the actual questionnaire, the first question asked, ‘‘What is the online shopping mall and portal type which you have begun to use most recently?’’ and then stated that ‘‘The ‘relevant online shopping mall’ in the next question refers to the online shopping mall in your answer to question 1.’’

5. Results We first consider sample size before testing the reliability and validity of our research model constructs. Sample size should be greater than five times and less than ten times the parameters (Bentler and Chou 1987), and a sample size of at least 200 is necessary for critical model testing (Hoetler 1983). Since we gathered 420 valid samples, these constraints were satisfied. Also, regarding analysis of the collected data, we used SPSS for descriptive statistics, exploratory factor analysis, and reliability analysis, including sample statistics, and AMOS 18, the structural equation modeling program, in the confirmatory factor analysis and covariance structure analysis of each factor. 5.1. Validity and reliability of measurement items This study used various measurement items to verify the empirical model, and executed exploratory factor analysis and confirmatory factor analysis together to evaluate the validity and reliability of the measurement items. We used exploratory factor analysis to determine whether the initially suggested items were derived as a single factor or separately from those measuring other concepts. Based on this factor analysis, reliability was evaluated by Cronbach’s alpha, measuring internal consistency among items constituting the same factors. Finally, the measurement model was set, with the result that exploratory factor analysis and confirmatory factor analysis were implemented on all the research variables.

349

We conducted the exploratory factor analysis using the varimax method, an orthogonal rotation method. Following Hair et al.’s (1998) recommendation, factor loadings greater than 0.50 were considered to be very significant. At first, the factor loading of the INT1 variable was lower than 0.50. Therefore we removed that item and repeated the analysis. As a result, all the factor loadings were actually greater than 0.50, with most of them above 0.60. The six variables formed each single factor clearly, and the dispersion explained by these factors accounted for 79.27% of the whole. The measurement variables are reliable according to Cronbach’s alpha values, which meet the general allowance: all the values were above 0.60 (Hair et al. 1998). Descriptive statistics are presented in Appendix A, and the analytical results of the validity and reliability testing of the measurement items by exploratory factor analysis are presented in Appendix B. 5.2. Confirmatory factor and correlation analysis The results of confirmatory factor analysis of all research variables, verifying the theoretical measurement model, are shown in Table 5. As the overall fitness meets the generally recommended standards, this measurement model is appropriate. The convergent validity and unidimensionality of each research variable were secured, as the standard loadings for each research variable are significant (t > 1.96) (Anderson and Gerbing 1988). In addition, the conceptual reliability values for all the research variables exceeded 0.7 (0.939–0.976), which is the general recommended standard of conceptual reliability. The average variance extracted (AVE) also exceeded 0.5 (0.783–0.928). The necessary standards having been met for each concept according to the results of the conceptual reliability analysis and the AVE, the measurement items used in this study accurately represent each proposed research variable (Hair et al. 1998). In addition, Table 6 shows that the research model is valid. Finally, we examined whether the square root of the AVE exceeds the correlation coefficients or not, as shown in Table 7 (Fornell and Larcker 1981). The discriminant validity was examined as a 95% confidence interval (ر2  standard error) of the correlation coefficients, showing that the square root of AVE is larger than the correlation coefficient (Fornell and Larcker 1981). The averages and standard deviations of each variable are also shown in Table 6 together with the correlation coefficients among the variables. In addition, in this study, to investigate the correlation between independent variables, a multicollinearity test was conducted using the SPSS statistical package. Generally, in this process, if the VIF (Variance Inflation Factor) value is found to be over 10 it is seen as a multicollinearity problem, but VIF values for the variables in this study were all under 10, as shown in Table 8, revealing that there is no issue with multicollinearity. 5.3. Verification In order to verify the research model and the hypotheses suggested in this study, we chose the structural equation method that evaluates the validity of whole research model synthetically instead of the method that evaluates each individual hypothesis. The parameter estimation method used in covariance structure analysis is maximum likelihood. Evaluation of the fitness of the model and the relations among variables should be conducted to verify the set of hypotheses through the path coefficient obtained by covariance structure analysis. The overall number of theoretical variables in this study is six. The model has one external variable, confirmation, and five internal variables: perceived usefulness, familiarity, intimacy, satisfaction, and continuance intention. Fig. 2 is a path diagram showing the results of AMOS analysis of the suggested research model.

350

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357 Table 5 Results of confirmatory factor analysis. Research variable

Measurement

Standard loading

Measurement error

t-value

Conceptual reliability

AVE

Confirmation

CONF1 CONF2 CONF3 CONF4

0.621 0.646 0.792 0.868

0.53 0.503 0.293 0.186

12.855 12.796 10.448 7.797

0.945

0.832

Satisfaction

SAT1 SAT2 SAT3 SAT4

0.734 0.763 0.837 0.718

0.385 0.34 0.248 0.411

12.035 11.643 9.923 12.227

0.941

0.783

Perceived usefulness

PU1 PU2 PU3

0.366 0.624 0.796

1.147 0.529 0.238

13.095 7.627 4.285

0.945

0.866

Familiarity

FAM1 FAM2 FAM3 FAM4

0.74 0.778 0.761 0.905

0.426 0.369 0.426 0.154

12.357 11.767 12.026 7.426

0.939

0.811

Intimacy

INT2 INT3 INT4 INT5

0.417 0.767 0.851 0.543

1.448 0.541 0.3 1.196

13.16 9.634 6.771 12.958

0.976

0.928

Continuance intention

CU1 CU2 CU3

0.796 0.588 0.554

0.367 0.782 0.96

8.093 12.071 12.381

0.959

0.899

The fitness indices of the research structure model are presented in Table 9. The model suggested in the study is judged to be appropriate to estimate the relations among the variables, as the results show satisfactory levels such as Chi-square = 433.316 (p = 0.00, degree of freedom = 190), Chi-square/degrees of freedom = 2.28, GFI = 0.92, AGFI = 0.89, NFI = 0.95, NNFI = 0.96, CFI = 0.97, RMSEA = 0.05 and so on.

Table 6 Results of goodness-of-fit measures. Goodness-of-fit measure

Recommended valuea

Structural model

Chi-square/degree of freedom Goodness-of-fit (GFI) Adjusted goodness-of-fit (AGFI) Normalized fit index (NFI) Non-normalized fit index (NNFI) Comparative fit index (CFI) Root mean square error of approximation (RMSEA)

63.00 P0.90 P0.80 P0.90 P0.90 P0.90 60.10

2.00 0.93 0.90 0.96 0.97 0.98 0.05

5.4. Hypotheses testing

a Recommended values have been adapted from Bentler (1989) and Hair et al. (1998).

Fig. 2 and Table 10 show the verified results of all ten research hypotheses suggested in this study. First, Hypothesis 1, which suggested a positive (+) relationship between confirmation and intimacy, was supported because its t value is 5.191 (p < 0.01), which was statistically significant. Hypothesis 2 was also

Table 7 Results of confirmatory factor and correlation analyses. Construct

Average

Confirmation Satisfaction Perceived usefulness Familiarity Intimacy Continuance intention a

4.610 4.769 4.903 5.011 4.207 4.762

S.D. 1.418 1.466 1.463 1.679 2.369 1.947

pffiffiffiffiffiffiffiffiffiffi AVE 0.912 0.885 0.93 0.9 0.963 0.948

Confirmation

Satisfaction

Perceived usefulness

Familiarity

Intimacy

Continuance intention

0.783a 0.759 0.558 0.558 0.743

0.866a 0.67 0.519 0.721

0.811a 0.626 0.709

0.928a 0.815

0.899a

a

0.832 0.853 0.71 0.506 0.491 0.685

AVE.

Table 8 Results of multicollinearity analysis. Model

1 (cont.)a CON SAT PU FAM INT a

Unstandardized coefficient

Standardized coefficient

B

Standard error

b

.223 .171 .186 .034 .195 .407

.197 .056 .056 .045 .040 .035

CONT: a dependent variable.

.150 .169 .029 .190 .442

t-value

1.132 3.085 3.335 .751 4.826 11.745

Significance

.258 .002 .001 .453 .000 .000

Multicollinearity statistics Tolerance

VIF

.362 .333 .584 .554 .608

2.765 3.006 1.712 1.804 1.645

351

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

Perceived Usefulness 0.175 * (2.537)

0.308** (6.207)

0.635** (15.753)

0.686 ** (14.368)

0.304** (5.525)

Continuance Intention (80.9)

Satisfaction

Confirmation

0.159** (3.343)

0.562**

(10.946) 0.308** (5.191)

Familiarity

0.423** (10.612)

0.542** (10.002)

Intimacy

Standardized path coefficient (t-value) *)p<0.05, **)p<0.01

Fig. 2. Path diagram.

Table 9 Fit indices for measurement and structural model. Goodness-of-fit measure

Recommended valuea

Structural model

Chi-square/degree of freedom Goodness-of-fit (GFI) Adjusted goodness-of-fit (AGFI) Normalized fit index (NFI) Non-normalized fit index (NNFI) Comparative fit index (CFI) Root mean square error of approximation (RMSEA)

63.00 P0.90 P0.80 P0.90 P0.90 P0.90 60.10

2.28 0.92 0.89 0.95 0.96 0.97 0.05

a Recommended values have been adapted from Bentler (1989) and Hair et al. (1998).

supported because its t value is 10.612, which was statistically significant. Hypotheses 3–5, which addressed the relationships between confirmation and familiarity, familiarity and intimacy, and familiarity and continuance, respectively, were supported because their t values were statistically significant as well (H3: t = 10.946, p < 0.01; H4: t = 10.002, p < 0.01; H5: t = 3.343, p < 0.01). In addition, Hypotheses 6–10, which focused on the relationships between confirmation and perceived usefulness, perceived usefulness and satisfaction, perceived usefulness and continuance intention, confirmation and satisfaction, and satisfaction and con-

tinuance intention, respectively, were supported, with all their t values being statistically significant (H6: t = 15.753, p < 0.01; H7: t = 6.207, p < 0.01; H8: t = 2.537, p < 0.05; H9: t = 14.368, p < 0.01; H10: t = 5.525, p < 0.01). Behavioral intention towards web-based service continuance is predicted by perceived usefulness (b = 0.175), satisfaction (b = 0.304), familiarity (b = 0.159), and intimacy (b = 0.423), which jointly explain 80.9% of the variance in intention. As exhibited in prior research, satisfaction is one of the significant motivators of behavior intention. However, this study also reveals that the affective factors familiarity and intimacy are important indicators of continuance intention, and that the affective factors have a stronger effect than some cognitive factors, such as perceived usefulness. Such result shows that to understand a relationship between a consumer and a service, it would have positive factors on having a consistent relationship when it is viewed as ‘‘a consumer has a relationship with a service,’’ rather than ‘‘a consumer uses a service.’’ In an actual case, when reviewing behaviors of web-based service users, generally the usage time of web-based service is increased and users consistently use a certain service by adding ‘‘favorites,’’ ‘‘shortcuts,’’ or ‘‘smart phone applications,’’ rather than a single instant use; reasonably supports the relationship-oriented results. Further, web-based service’s lower barrier of entry than the other industries creates other similar services when

Table 10 Results of hypotheses testing. Hypothesis

Path

Standard coefficient

Standard error

t-value

Hypotheses acceptance

H1 H2 H3 H4 H5 H6 H7 H8 H9 H10

Confirmation ? intimacy Intimacy ? continuance intention Confirmation ? familiarity Familiarity ? intimacy Confirmation ? satisfaction Satisfaction ? continuance intention Familiarity ? continuance intention Confirmation ? perceived usefulness Perceived usefulness ? satisfaction Perceived usefulness ? continuance intention

0.308 0.423 0.562 0.542 0.686 0.304 0.159 0.635 0.308 0.175

0.059 0.04 0.051 0.054 0.048 0.055 0.048 0.04 0.05 0.069

5.191 10.612 10.946 10.002 14.368 5.525 3.343 15.753 6.207 2.537

Supported Supported Supported Supported Supported Supported Supported Supported Supported Supported

352

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

a new web-based service is created. Moreover, such trend will provide similar functions and user interface, resulting in emotional factors have greater influence on user’s selection and consistent usage than on usefulness and recognition factors. Ultimately, it shows that affective factors should be the first consideration of efforts to build users’ continuance intention. Comparing the results with prior ECM-based studies on continuance intention, the impact of perceived usefulness and satisfaction on intention are similar to what had been indicated in Bhattacherjee (2001b), although the explanatory power of the current research model is much higher than the previous model, possibly because familiarity and intimacy have direct effects which explain the error variance. Customer satisfaction is predicted by confirmation (b = 0.686) and perceived usefulness (b = 0.308), with 79.1% of the total variance explained. The effect of confirmation on satisfaction is much greater than perceived usefulness. However, the influence of perceived usefulness on continuance intention is significant in our study, which is obviously consistent with Bhattacherjee’s (2001b) findings: that perceived usefulness has a significant impact on continuance intention. Confirmation can determine familiarity (b = 0.423) and intimacy (b = 0.308). This result is consistent with those of prior studies, such as Ajzen (1991), Mathieson (1991) and Taylor and Todd (1995), which found causality between cognitive beliefs and affect. Finally, intimacy is determined by confirmation (b = 0.308) and familiarity (b = 0.423), which jointly explain 43.6% of the error variance in intimacy. This result reveals that deep emotion such as intimacy can be ultimately influenced by both cognitive and affective factors.

Original Research Model Perceived Usefulness

Satisfaction

Confirmation

Continuance Intention

Familiarity

Intimacy

Alternative model 1

Satisfaction

Confirmation

Continuance Intention

Familiarity

5.5. Alternative model settings and analysis We compared the results by setting and analyzing alternative models from the traditional research one. Fig. 3 presents both the research model and alternative models. Alternative Model 1 is a model with perceived usefulness removed, and in Alternative Model 2, all the paths in the research model have been set, along with the additional relationship sets familiarity, satisfaction and intimacy, and then satisfaction. To compare the models statistically, v2 difference tests should be conducted, for which the models should be in a mutually nested relation. Currently, however, Alternative Model 1 is not nested in the research model, while the research model is nested in Alternative Model 2. Therefore, direct comparison by v2 difference test is not available between the Research model and Alternative Model 1, while direct comparison by v2 difference test is available between the research model and Alternative Model 2. First, comparing the research model and Alternative Model 1, the v2 value and parsimony fit indices proposed as a result of analyzing Alternative Model 1 are shown in Table 11. The chief parsimony fit values proposed as a result of analyzing Alternative Model 1 are v2 = 354.135, d.f. = 140, p = .000, PGFI = .677, PNFI = .779, PCFI = .794, and AIC = 454.135. However, since Alternative Model 1 and the research model are not in a nested relation, direct comparison through the v2 value is not possible. Therefore we use the parsimony fit indices instead. As shown in Table 9, the parsimony fit values for the research model were PGFI = .688, PNFI = .780, PCFI = .798, and AIC = 559.316, higher than those in Alternative Model 1, showing that the original research model is superior to Alternative Model 1. In the path coefficient values of the research model shown in Fig. 2, those for confirmation ? perceived usefulness, perceived usefulness ? satisfaction, and perceived usefulness ? continuance intention were all shown to be significant. But since Alternative Model 1 is the one with all these significant path coefficients removed, goodness-of-fit is considered to be lower than for the Research Model.

Intimacy

Alternative model 2 Perceived Usefulness

Satisfaction

Confirmation

Continuance Intention

Familiarity

Intimacy Fig. 3. Research model and alternative models.

Table 11 Comparison: parsimony fit indices between research model and Alternative Model 1. Division 2

v parsimony

Parsimony fit indices

Fit Index

Research model

Alternative Model 1

d.f. P

433.316 190 .000

354.135 140 .000

PGFI PNFI PCFI AIC

.688 .780 .798 559.316

.677 .779 .794 454.135

2

v

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357 Table 12 Comparison of parsimony fit indices between research model and Alternative Model 2. Division

Fit Index

Research model

Alternative Model 2

v2 parsimony

v2

433.316 190 .000

419.371 188 .000

d.f. P

The v2 value and parsimony fit index presented as a result of Alternative Model 2 are shown in Table 12. The major fit index values proposed as a result of analyzing Alternative Model 2 are v2 = 419.371, d.f. = 188, p = .000, PGFI = .683, PNFI = .773, PCFI = .791, and AIC = 549.371. Since Alternative Model 2 and the research model are in a nested relation, they are directly comparable through the v2 difference test. The v2 value of Alternative Model 2 is 13.945 smaller, and the model is 2.0 smaller in terms of degrees of freedom, compared to the research model, as shown in Table 12. Since the v2 value is 5.99 when the significance level is .05 and the degree of freedom is 2.0, Alternative Model 2 is considered superior compared with the research model (Dv2 = 13.945, Dd.f. = 2). In Alternative Model 2, the degrees of freedom decreased as much as 2.0 compared to the research model, but considering that the decrease in v2 value was enough to counteract the decrease in degrees of freedom, Alternative Model 2 is superior to the research model. The reason why Alternative Model 2 is shown to be superior to the research model can be found in relation to the route coefficient. In the route coefficient value of Alternative Model 2, the familiarity ? satisfaction path (estimate = 0.006, C.R. = 0.141, p = 0.888) turned out to be insignificant, while the intimacy ? satisfaction path was significant (estimate = 0.109, C.R. = 3.254, p = 0.001). Therefore, compared to the research model, which does not include the significant path of intimacy ? satisfaction, Alternative Model 2 (which does include these routes) has greater goodness-of-fit. Based on the above result of analyzing the alternative models, we added the intimacy ? satisfaction path to the established research model. In an analysis of the corrected model, all route coefficient values were shown to be significant, especially the newly added intimacy ? satisfaction path (estimate = 0.111, C.R. = 0.029, p = 0.001). In addition, as a result of the v2 difference test on the established research model and correction model, the latter was 13.925 smaller in v2 value and 1.0 smaller in degree of freedom than the former (Dv2 = 13.925, Dd.f. = 1). However, values for fit indices were v2 = 419.391, d.f. = 189, p = 0.000, Chi-square/degrees of freedom = 2.22, GFI = 0.92, AGFI = 0.89, NFI = 0.95, NNFI = 0.965, CFI = 0.97, and RMSEA = 0.05, almost the same as the established model’s Chi-square/degrees of freedom = 2.28, GFI = 0.92, AGFI = 0.89, NFI = 0.95, NNFI = 0.96, CFI = 0.97, and RMSEA = 0.05. Also, in terms of R2 value, which indicates descriptive power for a model, both the established model and correction model yielded the same value of 80.9, suggesting no difference in descriptive power between the two models. Thus, in this study, weighing efficiency and improvement in goodness-of-fit against the cost incurred by the addition of hypotheses, we see that the established research model was optimal. 6. Discussion 6.1. Balanced theoretical framework for continuance intention The ECM has long been valuable in studies of users’ continued intention to use web-based services. The model mainly considers cognitive factors such as expectation and confirmation. In this paper, we argue that affective factors such as familiarity and intimacy should also be considered in explaining the post-adoption behavior of web-based services. Cognitive factors alone are insufficient,

353

because sometimes people’s feelings play a central role in the decision-making process, according to consumer research and social psychology. Hence, we proposed an extended ECM, which newly incorporates two affective factors: familiarity and intimacy, which are relatively more persistent than previously studied emotional factors, and have been regarded as essential for strengthening human relationships in research on consumer behavior. These two new factors significantly affected continuance intention in our analysis. Our new extended model makes three major contributions, as follows. First, this study proposes a balanced theoretical framework for continuance intention. Considering that cognition and emotion are two main paths to influence attitude and behavioral intention, we proposed a balanced research model to show that continuance intention is simultaneously affected by both cognitive and affective factors. The results of our empirical analysis show that these factors affect continuance intention conjointly, with the same result as in existing studies of other emotional factors affecting continuance intention. However, the influence of affective factors on users’ behavior is shown to be larger than that of cognitive factors. Second, this study also proposes new affective factors relevant to continuance intention, identifying two relatively persistent emotional factors created over time and clarifying the relationship between such emotional factors and continuance intention. Intimacy and familiarity are representative persistent feelings, based on a prior relationship with the user, which have been regarded as essential emotions for forming long-term human relationships in consumer behavior research. Hence, this study, treating users’ continued usage of a web-based service as an interpretable, long-term relationship with that service, clarified the relationship between these affective factors and continuance intention in an extended ECM. More specifically, as a result of analyzing the relationship of familiarity and intimacy to continuance intention empirically, this study found that both factors are clearly positively associated with continuance intention. This means that continuance intention is affected by longer-term emotions, as expected in this study, as well as by short-term, temporary emotions such as pleasure or playfulness. Third, we also found that intimacy affects continuance intention more strongly than perceived usefulness and familiarity due to the difference in the degree of emotional intensity between intimacy and familiarity, that is, because intimacy is stronger and more attitudinal than familiarity. 6.2. Implications for managers The results of this study show that intimacy is a more powerful predictor of users’ continuance intention than perceived usefulness, which was previously recognized as one of the most influential cognitive factors from the adoption stage to the post-adoption stage. Existing companies, particularly IT-related companies, have focused on developing high-quality, highly technological, stateof-the-art products and services. However, recently it has been demonstrated that customers’ long-term consumption patterns are not dependent simply on the functionality or usability of web-based services. Customers are interested in the story behind the product or how it can stimulate their emotions, not in rational suggestions. Repurchasing consumers are moving from a ‘‘function pursuit’’ to a ‘‘feeling pursuit.’’ Consumers today see few major differences in quality or function between sites, due to the highly competitive markets among top-tier websites. Consequently, the major differentiating factors are the feelings or experiences they have while engaging with the websites. Since this trend is directly relevant to the period postadoption, companies should not expect customers’ ongoing loyalty simply on the basis of functional quality. Now repurchase can be

354

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

expected only if there are constant relationships with customers. Therefore, it is necessary to change from a service-focused strategy to a relationship-focused one. At this time, managers of online shopping sites must consider what kinds of relationships customers really want. This study sheds light on this matter: customers prefer true and sincere relationships rather than just usefulness. They want closer relationships that include intimacy. Indeed, for a company to establish intimacy with its customers is a great achievement in the area of customer/service provider relationships. Intimacy can be classified either as customer-to-service intimacy or customer-to-customer intimacy. With respect to customer-toservice intimacy, this paper provides empirical evidence of the fact that successful web-based service companies value customer intimacy as a resource, a strategic source of sustainable management. Web-based service companies are willing to incur short-term costs in order to build long-term loyalty and satisfaction. Promoting customer-to-customer intimacy has become a crucial trigger of continuance intention in the context of social media, for which customer-to-customer intimacy really does matter. By using social media, companies can increase the level of intimacy with their customers. For example, if a social media user becomes close with others through Facebook or Twitter, the social media becomes a must-have service to the customer no matter what its quality, clearly increasing his or her continuance intention. Webbased service managers are thus wise to incorporate a social media component in order to promote intimacy with their customers. Intimacy can be even stronger than trust with respect to continuance intention, even though emotion is positively connected with trust (Dunn and Schweitzer 2005). There seems to be a subtle incongruence between trust and intimacy. By definition, trust is the willingness to accept vulnerability based upon positive expectations about another’s behavior. Intimacy is the ability to tolerate flaws in another’s personality and behavior. Hence, intimacy is stronger than trust in that intimacy accepts other undesirable outcomes, not just vulnerability. Intimacy is more durable than trust. Take the example of Apple’s iTunes, a representative successful web service. In terms of customer-to-service intimacy, the sense of closeness that customers feel for Apple has been based on much more than Apple’s actual leverage of customer data. They have nurtured that warm and fuzzy feeling from top-down branding. The iTunes store is an online media sales service offered through iTunes of Apple. The store sells music, audio, videos, movies, TV programs, and games for iPods, and distributes podcasts. Since April 28th, 2003, its market has grown explosively: Apple announced in 2010 that ten billion music items have been sold since the opening of the iTunes music store. This success has been caused by the fact that the familiarity and intimacy between customers and company have increased as Apple has made the most of this online product distribution channel, following up with related products such as the iPod, iPod Touch, iPhone, etc. iTunes has various functions to increase intimacy between itself and the user, and Genius is one of them. Genius analyzes users’ song choices and can make recommendations about playlists, either by showing a list of compatible songs already on the playlist, or by suggesting additional songs from the iTunes Store. The function has received good reviews from users, and it leverages customer data to create more perceived value in iTunes. As a result, Apple lovers will continue to feel intimacy with Apple, and barriers to exit will be higher than ever. The intimacy enhancement between Apple and its services has created customers loyal to Apple’s services who are the driving force behind its legendary success. Obviously, intimacy is a factor critical to success, although it is not often handled in a sophisticated manner, especially when it comes to web-based services, in terms of service quality or relationship management. In addition to objective factors such as product quality, managers need to consider how intimacy affects customers when planning customer-retention strategies. Manag-

ers can run customer relationship management (CRM) programs using customers’ basic profiles, and analyze past transaction data, but should also take into consideration customers’ emotions. Moreover, since customer intimacy can be fostered by individual care, not just by classifying customers by group or market segment, future CRM should be personalized. Any customer satisfaction evaluation system used as a key element in customer-related performance evaluation will need to include new evaluation items and measurement methods that can measure beyond satisfaction levels to include affective factors. 6.3. Emotional marketing The results of this paper imply that intimacy is a crucial component of emotional marketing. Emotional marketing originally aims to strengthen ties with a brand through emotional stimulation that appeals to customers’ feelings and sentiments, hence increasing brand loyalty by making the customer feel his or her own value in the brand. Recently, many companies have been proactive, using emotional marketing as a strategic means of ‘‘enticing new customers’’ and ‘‘marinating existing customers.’’ For example, the high-class coffee shop chain Starbucks created a new coffee culture through human-oriented emotional marketing. They converted a shop where people only drink coffee into a space of emotional experience. Customers feel comfortable in the pleasant and intimate setting because of the emotional stimulation that the coffee shop provides. As a result, Starbucks has been one of the world’s most successful companies in the last ten years. Web-based services will be no exception to this new marketing trend. Activities related to traditional channels, customer seminars, newspaper advertisements, etc. were once the norm in terms of marketing approaches; web-based service companies could also do these things. These were so-called customer-oriented marketing activities, but were essentially one-way, saying ‘‘our product is the best’’ or ‘‘this service should be introduced.’’ However, changes in customer needs require a new method to secure competitiveness: emotional marketing that stimulates customers’ sensibilities. It is time for emotional factors to be applied, such as familiarity and intimacy, which affect usage behavior. Emotional marketing focuses on pleasure, pride, and affection, rather than the product per se. Its success greatly depends on whether loyalty to the brand can be created or not. Therefore, the level to which relationships can be personalized should be considered. Customers remain loyal to the emotions and experiences they have while using web-based services; hence, stimulating customers’ emotions is an essential requirement of a successful strategy (Gobe 2001). Despite the substantial value of emotional marketing, emotional factors currently incorporated in the ECM seldom consider social bonds between customers and service providers. Moreover, they fail to reflect the multidimensional and long-term nature of the emotion concept (Plutchik 1982, Russell 2003). Thus, more elaboration was needed to identify significant affective factors which indicate social bonds formed over time between customer and service provider, in explaining continued use of web-based services. In this paper, we have shown that intimacy is a prominent social bond that clearly predicts user satisfaction and continuance intention.

7. Conclusion Gaining experience with a particular web-based service through repeat visits increases familiarity. However, ongoing visits do not guarantee an increase in positive affect, such as intimacy. The risk of becoming too familiar with a service, especially a web-based service that is likely to change and evolve, is that the user may ultimately

355

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

experience disappointment or an unexpected negative affect when the website is suddenly different from what he or she is used to. Our model goes beyond modifying Bhattacherjee’s (2001b) expectation–confirmation model (ECM), the most representative research model on continuance intention. We show that some affective factors, including intimacy, provide an alternative explanation as to how or why customers form continuance intention. We show that both familiarity and intimacy significantly determine continuance intention, even to the extent that temporal dissatisfaction may not actually affect continuance intention. This is easily observed: customers who have sufficient familiarity and/or intimacy come back to revisit a web-based service even when they encounter an unsatisfactory experience, such as a product claim or delay in delivery. These findings have meaningful practical implications: understanding the client’s affective factors has become crucial to the success of a web-based service. As the number of competitors increases, it becomes less and less feasible to maintain a competitive edge only by offering economic incentives. Rather, clients keep coming back not only because the service is useful, but also because they have a sense of intimacy with the service. The affective asset is truly a competitive advantage when the service provider is also satisfied with the clients, thus avoiding unnecessary economic expenditure. Finally, other long-term affective factors than intimacy and familiarity can be added in the future. These factors have already

been successfully used to predict service users’ overall subjective well-being, with implications for their behaviors in shopping and social networking (Urry et al. 2004). Also, affective factors are generally classified as positive (cheerful, happy, comfortable, etc.) or negative (nervous, fearful, angry, etc.). It would be interesting to consider negative affect factors, as well as positive factors, in examining the process of discontinuance intention. The well-established positive affect and negative affect scales (PANAS) can easily be used for this purpose (Watson et al. 1988). This study suffers from several limitations. First, when online shopping mall is selected as one of web-based services, and the term ‘‘relevant online shopping mall’’ is used in a survey. However, because there are numerous types of online shopping mall, there is a possibility of respondents having uncertain understanding. Even though we provided various meanings of online shopping mall to minimize such misunderstanding, we need to be careful to interpret the results of this study. Second, there may be some concerns about evaluating continuance intention in the context of online stores, because satisfaction may partially come from products rather than the online stores. Acknowledgement This work was supported by a grant from the Kyung Hee University in 2010.

Appendix A. Descriptive statistics Construct

N

Min

Max

Sum

AVE

S.D.

Confirmation Perceived usefulness Satisfaction Familiarity Intimacy Continuance intention

420 420 420 420 420 420

1.00 1.33 1.00 1.00 1.00 1.00

7.00 7.00 7.00 7.00 7.00 7.00

1936.25 2077.36 2011.25 2104.75 1766.75 1999.98

4.6101 4.9461 4.7887 5.0113 4.2065 4.7619

1.07376 1.03912 1.11890 1.19104 1.32829 1.22488

Appendix B. Results of exploratory factor analysis and reliability test Research variable

Measurement

Validity Factor 1

Confirmation

CONF1 CONF2 CONF3 CONF4

Satisfaction

SAT1 SAT2 SAT3 SAT4

Perceived usefulness

PU1 PU2 PU3 FAM1 FAM2 FAM3 FAM4

Intimacy

INT2 INT3 INT4 INT5

Continuance intention

CU1 CU2 CU3

Reliability Factor 2

Factor 3

Factor 4

Factor 5

Factor 6

.761 .841 .778 .780

Cronbach’s alpha 0.923

.723 .716 .717 .617

0.927

.800 .772 .568 .786 .856 .835 .852

0.815

0.939

.861 .630 .736 .730

0.886

.615 .731 .713

0.851

356

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357

References Acitelli, L. K., and Duck, S. Intimacy as the proverbial elephant. In D. Perlman and S. Duck (eds.), Intimate Relationships: Development, Dynamics, and Deterioration, Sage, Thousand Oaks, CA, 1987, 297–308. Ajzen, I. The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 2, 1991, 179–211. Anderson, E. W., and Fornell, C. Cross-category variation in customer satisfaction and retention. Marketing Letters, 5, 1, 1994, 19–30. Anderson, J., and Gerbing, D. Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 1988, 411– 423. Aronson, E. The Jigsaw Classroom. Sage Publications, Beverly Hills, CA, 1978. Atchariyachanvanich, K., Okada, H., and Sonehara, N. What keeps online customers repurchasing through the Internet? ACM SIGecom Exchanges, 6, 2, 2006, 47–57. Bagarozzi, D. A. Marital intimacy needs questionnaire: Preliminary report. American Journal of Family Therapy, 25, 3, 1997, 285–290. Bentler, P., and Chou, C. Practical issues in structural modeling. Sociological Methods and Research, 16, 1987, 78–117. Bhattacherjee, A. An empirical analysis of the antecedents of electronic commerce service continuance. Decision Support Systems, 32, 2, 2001, 201–214. Bhattacherjee, A. Understanding information systems continuance: An expectation–confirmation model. MIS Quarterly, 25, 3, 2001, 351–370. Bickmore, T., and Picard, R. W. Establishing and maintaining long-term human– computer relationships. ACM Transactions on Computer Human Interaction, 12, 2, 2005, 293–327. Bolton, R. N., and Lemon, K. N. A dynamic model of customers’ usage of services: Usage as an antecedent and consequence of satisfaction. Journal of Marketing Research, 36, 5, 1999, 171–186. Bornstein, R. F. Exposure and affect: Overview and meta-analysis of research, 1968– 1987. Psychological Bulletin, 106, 2, 1989, 265–289. Chelune, G. J., and Waring, E. M. Nature and assessment of intimacy. In P. McReynolds and G. J. Chelune (eds.). Advances in Psychological Assessment, vol. 6, Jossey-Bass, San Francisco, CA, 1984, 277–311. Chiu, C. M., and Wang, E. T. G. Understanding web-based learning continuance intention: The role of subjective task value. Information & Management, 45, 3, 2008, 194–201. Colombo, R. A., and Morrison, D. G. A brand switching model with implications for marketing strategies. Marketing Science, 8, 1989, 89–99. Cooper, R. B., and Zmud, R. W. Information technology implementation research: A technological diffusion approach. Management Science, 36, 2, 1990, 123– 139. Cordova, J. V., and Scott, R. L. Intimacy: A behavioral interpretation. Behavior Analyst, 24, 1, 2001, 75–86. Danaher, P. J., and Rust, R. T. Indirect financial benefits from service quality. Quality Management Journal, 3, 2, 1996, 63–75. Davis, F. D. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13, 3, 1989, 319–340. Davis, F. D., Bagozzi, R. P., and Warshaw, P. R. User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35, 8, 1989, 982–1003. Duck, S. Relating to Others. Open University Press, London, UK, 1988. Dunn, J. R., and Schweitzer, M. E. Feeling and believing: The influence of emotion on trust. Journal of Personality & Social Psychology, 88, 5, 2005, 736–748. Ecker, U. K. H., Zimmer, H. D., Groh-Bordin, C., and Mecklinger, A. Context effects on familiarity are familiarity effects of context—An electrophysiological study. International Journal of Psychophysiology, 64, 2007, 146–156. Festinger, L. A. A Theory of Cognitive Dissonance. Stanford University Press, Stanford, CA, 1957. Fornell, C., and Larcker, D. F. Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 1981, 39–50. Gefen, D. E-commerce: The role of familiarity and trust. Omega: The International Journal of Management Science, 28, 6, 2000, 725–737. Gefen, D., Karahanna, E., and Straub, D. Trust and TAM in online shopping: An integrated model. MIS Quarterly, 27, 1, 2003, 51–90. Gobbini, M. I., Leibenluft, E., Santiago, N., and Haxby, J. V. Social and emotional attachment in the neural representation of faces. Neuroimage, 22, 2004, 1628–1635. Gobe, M. Emotional Branding: The New Paradigm for Connecting Brands to People. Allworth Press, New York, NY, 2001. Granovetter, M. The strength of weak ties: A network theory revisited. In P. Marsden and N. Lin (eds.), Social Structure and Network Analysis, John Wiley and Sons, New York, NY, 1982, 105–130. Grayson, K., and Ambler, T. The dark side of long-term relationships in marketing services. Journal of Marketing Research, 36, 1, 1999, 132–141. Hair, J. F., Jr., Anderson, R. E., Tatham, R., and Black, W. C. Multivariate Data Analysis, 5th edition. Prentice-Hall, Englewood Cliffs, NJ, 1998. Hsu, M. H., and Chiu, C. M. Internet self-efficacy and electronic service acceptance. Decision Support Systems, 38, 3, 2004, 369–381. James, N. Care = organisation + physical labor + emotional labour. Sociology of Health and Illness, 14, 1992, 488–509. Jap, S. D., and Ganesan, S. Control mechanisms and the relationship life cycle: Implications for safeguarding specific investments and developing commitment. Journal of Marketing Research, 27, 2, 2000, 227–245.

Karahanna, E., Straub, D. W., and Chervany, N. L. Information technology adoption across time: A cross-sectional comparison of pre-adoption and post-adoption beliefs. MIS Quarterly, 23, 2, 1999, 183–213. Kelley, C. M., and Jacoby, L. L. Subjective reports and process dissociation: Fluency, knowing, and feeling. Acta Psychologica, 98, 1998, 127–140. Kim, H. W., Chan, H. C., and Chan, Y. P. A balanced thinking-feelings model of information systems continuance. International Journal of Human–Computer Studies, 65, 2007, 511–525. Komiak, X. S., and Benbasat, I. The effects of personalization and familiarity on trust and adoption of recommendation agents. MIS Quarterly, 30, 4, 2006, 941–960. Kwon, T. H., and Zmud, R. W. Unifying the fragmented models of information systems implementation. In R. J. Boland and R. A. Hirschheim (eds.), Critical Issues in Information Systems Research, John Wiley and Sons, New York, NY, 1987, 227–251. Lam, D., and Mizerski, D. The effects of locus of control on word-of-mouth communication. Journal of Marketing Communications, 11, 3, 2005, 215–228. Lam, S. Y., Shankar, V., Erramilli, M. K., and Murthy, B. Customer value, satisfaction, loyalty, and switching costs: An illustration from a businessto-business service context. Journal of the Academy of Marketing Science, 32, 3, 2004, 293–311. Laurenceau, J. P., Barrett, L. F., and Pietromonaco, P. R. Intimacy as an interpersonal process: The importance of self-disclosure, partner disclosure, and perceived partner responsiveness in interpersonal exchanges. Journal of Personality and Social Psychology, 74, 12, 1998, 38–51. Liao, C., Chen, J. L., and Yen, D. C. Theory of planning behavior (TPB) and customer satisfaction in the continued use of e-service: An integrated model. Computers in Human Behavior, 23, 2007, 2804–2822. Limayem, M., and Cheung, C. M. K. Understanding information systems continuance: The case of Internet-based learning technologies. Information & Management, 45, 4, 2008, 227–232. Lin, C. S., Wu, S., and Tsai, R. J. Integrating perceived playfulness into expectation– confirmation model for web portal context. Information & Management, 42, 2005, 683–693. Lowenthal, M. F., and Haven, C. Interaction and adaptation: Intimacy as a critical variable. American Sociological Review, 33, 1968, 20–30. Marsden, P. V., and Campbell, K. E. Measuring tie strength. Social Forces, 63, 2, 1984, 482–501. Maslow, A. H. The influence of familiarization on preference. Journal of Experimental Psychology, 21, 1937, 162–180. Mathieson, K. Predicting user intentions: Comparing the technology acceptance model with the theory of planned behavior. Information Systems Research, 2, 3, 1991, 173–191. Matlin, M. W. Response, competition, recognition, and affect. Journal of Personality and Social Psychology, 19, 3, 1971, 295–300. Min, Q., and Shenghua, X. An extended expectation confirmation model for information systems continuance. In International Conference on Wireless Communications, Networking and Mobile Computing, 2007, 3879–3882. Moss, B. F., and Schwebel, A. I. Defining intimacy in romantic relationships. Family Relations, 42, 1, 1993, 31–37. Oatley, K., and Jenkins, J. M. Understanding emotions. In Psychology, Psychiatry, and Social Science, Blackwell, Cambridge, MA, 1996. Oliver, R. L. A cognitive model for the antecedents and consequences of satisfaction. Journal of Marketing Research, 17, 4, 1980, 460–469. Oliver, R. L. A conceptual model of service quality and service satisfaction: Compatible goals, different concepts. In T. A. Swartz, D. E. Bowen, and S. W. Brown (eds.). Advances in Services Marketing and Management: Research and Practice, vol. 2), JAI Press, Greenwich, CT, 1993. Plutchik, R. A psychoevolutionary theory of emotions. Social Science Information, 21, 4–5, 1982, 529–553. Raj, S. P. Striking a balance between brand ‘popularity’ and brand loyalty. Journal of Marketing, 49, 1985, 53–59. Ratcliffe, M. J. Heidegger’s attunement and the neuropsychology of emotion. Phenomenology and the Cognitive Sciences, 1, 2002, 287–312. Rau, P. L. P., Gao, Q., and Ding, Y. Relationship between the level of intimacy and lurking in online social network services. Computers in Human Behavior, 24, 6, 2008, 2757–2770. Reis, H. T., and Shaver, P. Intimacy as an interpersonal process. In S. Duck (ed.), Handbook of Research in Personal Relationships, John Wiley and Sons, London, UK, 1988, 367–389. Rindfleisch, A., and Inman, J. Explaining the familiarity–liking relationship: Mere exposure, information availability, or social desirability? Marketing Letters, 9, 1, 1998, 5–19. Roca, J. C., Chiub, C. M., and Martíneza, F. J. Understanding e-learning continuance intention: An extension of the technology acceptance model. International Journal of Human–Computer Studies, 64, 2006, 683–696. Rogers, E. M. Diffusion of Innovations, 4th edition. Free Press, New York, NY, 1995. Russell, J. A. Core affect and the psychological construction of emotion. Psychological Review, 110, 1, 2003, 145–172. Schaefer, M. T., and Olson, D. H. Assessing intimacy: The PAIR inventory. Journal of Marital and Family Therapy, 7, 1981, 47–60. Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., and Sherif, C. W. Intergroup Conflict and Cooperation: The Robbers Cave experiment. Institute of Social Relations, University of Oklahoma, Norman, OK, 1961. Shuchman, A. Are there laws of consumer behavior? Journal of Advertising Research, 8, 3, 1968, 19–28.

Y. Lee, O. Kwon / Electronic Commerce Research and Applications 10 (2011) 342–357 Söderlund, M. Customer familiarity and its effects on satisfaction and behavioral intentions. Psychology and Marketing, 19, 10, 2002, 861–880. Sørebø, O., and Eikebrokk, T. R. Explaining IS continuance in environments where usage is mandatory. Computers in Human Behavior, 24, 5, 2008, 2357– 2371. Stern, B. B. Advertising intimacy: Relationship marketing and the services consumer. Journal of Advertising, 4, 1997, 7–19. Szymanski, D. M., and Henard, D. H. Customer satisfaction: A meta analysis of empirical evidence. Journal of the Academy of Marketing Science, 29, 1, 2001, 16–35. Taylor, S., and Todd, P. A. Understanding information technology usage: A test of competing models. Information Systems Research, 6, 2, 1995, 144–176. Thong, J. Y. L., Hong, S. J., and Tam, K. Y. The effects of post-adoption beliefs on the expectation–confirmation model for information technology continuance. International Journal of Human–Computer Systems, 64, 2006, 799–810. Timmerman, G. M. A concept analysis of intimacy. Issues in Mental Health Nursing, 12, 1991, 19–30. Tolstedt, B. E., and Stokes, J. P. Relation of verbal, affective, and physical intimacy to marital satisfaction. Journal of Counseling Psychology, 30, 1983, 573–580. Tomasi, A. Technology and intimacy in the philosophy of Georges Bataille. Human Studies, 30, 4, 2007, 411–428. Trauer, B., and Ryan, C. Destination image, romance and place experience—An application of intimacy theory in tourism. Tourism Management, 26, 4, 2005, 481–491. Tsai, H. T., and Huang, H. C. Determinants of e-repurchase intentions: An integrative model of quadruple retention drivers. Information & Management, 44, 3, 2007, 231–239. Urry, H. L., Nitschke, J. B., Dolski, I., Jackson, D. C., Dalton, K. M., and Mueller, C. J. Making a life worth living: Neural correlates of well-being. Psychological Science, 15, 6, 2004, 367–372.

357

Vatanasombut, B., Igbaria, M., Stylianou, A. C., and Rodgers, W. Information systems continuance intention of web-based applications customers: The case of online banking. Information & Management, 45, 7, 2008, 419–428. Venkatesh, V. Determinants of perceived ease of use: Integrating control, intrinsic motivation, and emotion into the technology acceptance model. Information Systems Research, 11, 4, 2000, 342–365. Venkatesh, V., and Davis, F. D. A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46, 2, 2000, 186–204. Venkatesh, V., Morris, M. G., Davis, G. B., and Davis, F. D. User acceptance of information technology: Toward a unified view. MIS Quarterly, 27, 3, 2003, 425–478. Verhoef, P. C., Francis, P. H., and Hoekstra, J. C. The effect of relational constructs on customer referrals and number of services purchased from multiservice provider: Does age of relationship matter? Academy of Marketing Science, 30, 3, 2002, 202–212. Watson, D., Clark, L. A., and Tellegen, A. Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1988, 1063–1070. Wellman, B. Physical place and cyber place. The rise of personalized networking. International Journal of Urban and Regional Research, 25, 2001, 227–252. Williams, A. A study of practicing nurses’ perceptions and experiences of intimacy within the nurse–patient relationship. Journal of Advanced Nursing, 35, 2, 2001, 188–196. Zajonc, R. B. Feeling and thinking: Preferences need no inferences. American Psychologist, 35, 1980, 151–175. Zajonc, R. B., and Helen, M. Affective and cognitive factors in preference. Journal of Consumer Research, 9, 1982, 123–131. Zhao, Y. Measuring the social capital of laid-off Chinese workers. Current Sociology, 50, 4, 2002, 555–571.