当前位置:首页 >> 自然科学 >>

applying the technology acceptance model and Flow Theory to Online Consumer Behavior


Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior
Marios Koufaris
Zicklin School of Business, Baruch College, CUNY, 55 Lexington Avenue, Box B11–220, New York, New York 10010 marios_koufaris@baruch.cuny.edu

n this study, we consider the online consumer as both a shopper and a computer user. We test constructs from information systems (Technology Acceptance Model), marketing (Consumer Behavior), and psychology (Flow and Environmental Psychology) in an integrated theoretical framework of online consumer behavior. Speci?cally, we examine how emotional and cognitive responses to visiting a Web-based store for the ?rst time can in?uence online consumers’ intention to return and their likelihood to make unplanned purchases. The instrumentation shows reasonably good measurement properties and the constructs are validated as a nomological network. A questionnaire-based empirical study is used to test this nomological network. Results con?rm the double identity of the online consumer as a shopper and a computer user because both shopping enjoyment and perceived usefulness of the site strongly predict intention to return. Our results on unplanned purchases are not conclusive. We also test some individual and Web site factors that can affect the consumer’s emotional and cognitive responses. Product involvement, Web skills, challenges, and use of value-added search mechanisms all have a signi?cant impact on the Web consumer. The study provides a more rounded, albeit partial, view of the online consumer and is a signi?cant step towards a better understanding of consumer behavior on the Web. The validated metrics should be of use to researchers and practitioners alike. (TAM; Flow Theory; Nomological Validity; Web Skills; Value-Added Search Mechanisms; Online Consumer Behavior)

I

1. Introduction
Consumer behavior on the Web has been the subject of considerable research in the last few years, but understanding it is made dif?cult by the fact that the main entities involved, consumers and businesses, have been transformed. First, every consumer is now also a computer user. The online consumer performs all the functions of a traditional consumer on a computer while interacting with a system, i.e., a commercial Web site. S/he, therefore, also exhibits all the characteristics of a computer user. Second, the physical
1047-7047/02/1302/0205$05.00 1526-5536 electronic ISSN

store has been transformed into a virtual store through information technology (IT). In the physical commercial world, the IT used for operations remains mostly in the background, invisible to the consumer. In e-commerce, however, the technology has been moved to the foreground and has become the store itself as a Web site. Such Web-based stores that use “networks and Internet technology for communications and transactions between various groups of stakeholders like businesses and consumers” have recently also been called
Information Systems Research, 2002 INFORMS Vol. 13, No. 2, June 2002, pp. 205–223

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

Net-enabled organizations (NEOs) (Straub and Watson 2001). There is a need to understand and measure such online consumer behavior and much of that burden falls on the IS ?eld (Straub and Watson 2001). To meet this goal, a multidisciplinary approach is ideal, and for this reason we take that approach in this study. In this study, we concentrate on new customer retention and unplanned purchases in business to consumer (B2C) electronic commerce. We also examine how certain emotional and cognitive responses to an initial Web store visit in?uence these two variables. In addition, we study the effects of some individual consumer differences as well as the impact of using different types of search mechanisms on the Web site. While models and constructs from marketing and psychology are prominent in this work, the research is relevant to the information systems ?eld due to its inclusion of traditional IS variables (Straub and Watson 2001), primarily the variables of the Technology Acceptance Model (Davis 1989).

2. Theoretical Framework and Study Measures
Discussion of metrics will accompany the building of theory-based hypotheses so that we can later examine a nomological network to test construct validity. Although this approach does not follow the conventional article format, we felt that it was justi?able given the stress on metrics in the special issue. There have been various attempts at building models of consumer behavior online. Examples include looking at “conversion” of Web surfers to Web customers (Berthon et al. 1996), studying consumer information acquisition and purchase decision online (Alba et al. 1997), and determining reasons for not shopping online (Peterson et al. 1997). Transaction cost theory has been used to explain the impact of IT on market structure (Malone et al. 1987) but also to understand individual consumer behavior. For example, buyer search costs are substantial transaction costs and they can determine buyer behavior and eventually market structures (Bakos 1997). An important question is whether online consumers think and act differently than their of?ine counterparts. If so, which metrics can we use to capture these

differences? For example, online consumers cannot depend on all ?ve senses to make purchases; instead, they must rely on limited product representations such as photographs and text descriptions. Yet, studies have shown that the representation of the product online, indeed the overall quality of the shopping experience, matters for both attitude towards shopping online as well as intention to buy (Burke et al. 1992, Jarvenpaa and Todd 1997a and b, Nowlis and McCabe 2000, Novak et al. 2000). NEOs, however, cannot rely on enticing product displays, exciting music, and “hip” clientele to convince their customers to buy. Instead, they may depend on well designed Web pages and powerful Web features, such as recommender systems and one-click checkouts. The online consumer may also have a different social and work environment than the of?ine consumer. For example, intention to buy online is in?uenced by the level of the consumer’s lack of leisure time, as in of?ine shopping, but also by the level of their wired lifestyle (Bellman et al. 1999). A key difference between online and of?ine consumer behavior is that the online consumer is generally more powerful, demanding, and utilitarian in her shopping expeditions. As a result, customer loyalty on the Web is low overall (Morrisette et al. 1999); although, as in the physical world, it is in?uenced by the availability of good, relevant content in an enjoyable context (Rice 1997, Eighmey 1997, Eighmey and McCord 1998). The locus of power seems to be shifting from the vendor to the consumer (Raman 1997) who does not favor traditional advertising and promotions online (Maignan and Lukas 1997). Research has also shown that certain site features, such as the availability of a FAQ section or promotions at the Web store entrance, can in?uence both traf?c on the Web site and overall sales (Lohse and Spiller 1998a and b). Unlike of?ine consumers, online consumers are concerned with those risks inherent in buying on the Web, risks such as credit card fraud and not receiving the right products (Bhatnagar et al. 2000). Perceived risk of online shopping and perceived ease of use of the Web site have been shown to in?uence attitude towards online purchasing (Heijden et al. 2001). The impact of perceived ease of use, however, seems to vary depending on the type of task the consumer is undertaking. Its effect is more signi?cant when consumers
Information Systems Research Vol. 13, No. 2, June 2002

206

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

are using a Web site to inquire about products rather than to purchase them (Gefen and Straub 2000). Online consumers clearly share some characteristics of their of?ine counterparts but also have unique needs and concerns that re?ect their online environment. Our study examines some of those similarities and differences in B2C commerce. We concentrate on new customers of a Web store and we measure their emotional and cognitive responses to determine what in?uences their intention to return and unplanned purchases. This is the basic nomological network that will allow us to test the metrics developed for the study. 2.1. Customer Intention to Return Customer retention is one of the primary goals of all companies (Pine et al. 1995, Reichheld and Sasser 1990) but store loyalty for NEOs can be low because of low switching costs (anonymous 1998). Also, for those customers who view shopping as a chance for getting out, socializing, and having fun (Morris 1987), a simply functional Web site can be very unappealing. If customers cannot participate in the activities that make shopping an enjoyable experience, they may stop using Web stores and return to the more enriching and enjoyable physical world (Rice 1997, Quelch and Takeuchi 1981). Ideally, one would measure customer loyalty by observing customers over time. Our study, however, used a cross-sectional questionnaire, so we used customer intention to return, not actual return visits, to measure loyalty. According to the Theory of Planned Behavior (Ajzen 1991) and its predecessor, the Theory of Reasoned Action (Fishbein and Ajzen 1975), behavioral intention can be a strong predictor of actual behavior. Therefore, we felt that intention to return is a satisfactory approximation of actual customer retention. 2.2. Why Customers Return When shopping on the Web, consumers perform certain tasks that can elicit both emotional and cognitive responses. Those responses can determine a new consumer’s intention to return for a second visit to a NEO. The double identity of the online consumer as a traditional shopper and a computer user implies that although attracting and retaining customers is a task that falls largely in the marketing realm, technology also
Information Systems Research Vol. 13, No. 2, June 2002

provides tools to assist companies with that task (Straub and Watson 2001). Therefore, to understand why such a consumer returns to a store, we need to look at her interaction with the Web site both as a store and as a system. First, we look at the Web consumer as a traditional shopper, albeit in a virtual environment. Therefore, we expect that her intention to return after a ?rst visit will be explained by more traditional psychological/marketing variables. One model from environmental psychology states that emotional responses to the environment mediate the relationship between the environment and one’s behavior (Mehrabian and Russsel 1974). According to this theory, physical and social stimuli in the environment in?uence the individual’s emotional state. Such stimuli can be as simple as wall colors and room temperature. Also, an individual’s personality traits can have an impact on his emotional state. That emotional state is de?ned in this model by three basic emotional responses: pleasure, dominance, and arousal. These three variables de?ne a person’s feelings that, in turn, in?uence behavior such as job performance and social interaction. Another model is ?ow (Csikszentmihalyi 1975 and 1977, Csikszentmihalyi and Csikszentmihalyi 1988), de?ned as “the holistic sensation that people feel when they act with total involvement” (p. 36). When people are in ?ow, they “shift into a common mode of experience when they become absorbed in their activity. This mode is characterized by a narrowing of the focus of awareness, so that irrelevant perceptions and thoughts are ?ltered out, by loss of self-consciousness, by a responsiveness to clear goals and unambiguous feedback, and by a sense of control over the environment” (Csikszentmihalyi 1977, p. 72). In the last few years, ?ow has also been studied in the context of information technologies and computermediated environments and has been recommended as a possible metric of the online consumer experience (Ghani et al. 1991, Trevino and Webster 1992, Webster et al. 1993, Ghani and Deshpande 1994, Hoffman and Novak 1996, Novak et al. 2000). While a valuable construct, we believe that ?ow is too broad and ill de?ned because of the numerous ways it has been operationalized, tested, and applied. We do, however, see value in some of the emotional and cognitive components used in ?ow research, namely, intrinsic enjoyment,

207

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

perceived control, and concentration/attention focus. The ?rst two variables also correspond to those of pleasure and dominance from environmental psychology (Mehrabian and Russel 1974). These constructs can be used as valid metrics for the online consumer experience. 2.2.1. Shopping Enjoyment. A common measure of ?ow is the level of intrinsic enjoyment of an activity, similar to the emotional response of pleasure from environmental psychology. Enjoyment in ?ow has, in fact, been measured using an adapted scale from environmental psychology (Novak et al. 2000). In the context of online shopping, we operationalized intrinsic enjoyment as shopping enjoyment and measured it with a four-item scale adapted from Ghani et al. (1991).1 Just as shopping enjoyment is important offline (Morris 1987, Forman and Sriram 1991, Blakney and Sekely 1994), it can be equally important online where it can have a signi?cant impact on attitude and intention towards online shopping (Jarvenpaa and Todd 1997a and b, Eighmey 1997). However, while shopping in the physical world can be a very enriching and emotionally ful?lling activity, shopping on the Web does not always provide the same experience because it is limited to mostly two-dimensional pictures and text. The importance of enjoyment in online shopping has been challenged in the past. At least one study found that recreational orientation was not different between nononline buyers, occasional online buyers, and frequent online buyers (Li et al. 1999). However, recreational orientation was de?ned as the consumer’s general attitude towards shopping, i.e., if shopping is something they enjoy overall. In our study, we look at the effect that a speci?c online shopping experience, and its enjoyment factor, can have on consumers. Past studies have indicated that shopping enjoyment can be an important determinant of online customer loyalty (Jarvenpaa and Todd 1997a and b, Rice 1997, Eighmey and McCord 1998). Also, ?ow research indicates that intrinsic enjoyment can positively impact the use of computer-mediated environments for e-mail use
1

(Trevino and Webster 1992), other software use (Webster et al. 1993), and Web use (Novak et al. 2000). We expect the effects on using a Web store to be similar. We, therefore, hypothesize that: Hypothesis 1a. Shopping enjoyment is positively related to intention to return. 2.2.2. Perceived Control. Perceived control has been used in several theoretical frameworks under different variations such as perceived locus of control (Rotter 1966), perceived control in achievement motivation theory (Atkinson 1964), self-ef?cacy (Bandura 1982), and perceived behavioral control in the theory of planned behavior (TPB) (Ajzen 1991). In ?ow research, perceived control has been de?ned as the level of one’s control over the environment and one’s actions. It is similar to Bandura’s self-ef?cacy and Ajzen’s perceived behavioral control in that it is speci?c to an action and it can vary for different situations or actions. In the same way, self-ef?cacy can be task- or situationspeci?c and perceived behavioral control in TPB is directed towards a speci?c behavior. Perceived control is also similar to the emotional response of dominance from environmental psychology, where it is de?ned as feeling “unrestricted or free to act in a variety of ways” in a speci?c situation and environment (Mehrabian and Russel 1974). In fact, an adapted scale for dominance has been used to measure perceived control in ?ow research (Novak et al. 2000). In our study, we measured perceived control with a four-item scale adapted from Ghani et al. (1991). For customers who shop on the Web, the information environment can be very different than for customers who shop in the physical world. The combination of less time available for shopping (Engel et al. 1990, Bellman et al. 1999), constant human cognitive resources available for information processing (Miller 1956), and an explosion of information and products available on the Web has led to more utilitarian customers demanding more control, less effort, and higher ef?ciency during shopping (Jarvenpaa and Todd 1997a and b, Clawson 1993, Tracy 1998). NEOs have responded to these challenges by providing site features such as search engines and recommendation agents to enable consumers to easily ?nd what they need, learn more about it, and quickly purchase it. All
Information Systems Research Vol. 13, No. 2, June 2002

A description of all the scales used in the study is available in the Appendix.

208

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

these site features result in Web customers enjoying high levels of control and convenience (Baty and Lee 1995, Hoffman and Novak 1996). Given the utilitarian nature of online consumers, we expect them to favor sites that provide them with a sense of perceived control, and so they would be more likely to remain loyal to them than other sites: Hypothesis 1b. Perceived control is positively related to intention to return. 2.2.3. Concentration/Attention Focus. For an individual to be in “?ow,” they must concentrate on their activity. Therefore, concentration has been a signi?cant correlate or measure of ?ow. It can also play a role in online consumer behavior. Web customers can have a short attention span because of their limited resources of time and information processing (Engel et al. 1990, Quelch and Klein 1996, Miller 1956) together with increased levels of control (Hoffman and Novak 1997, Sheth and Sisodia 1997). However, concentration can be critical for completing their purchases ef?ciently. Buying online is carried out in front of a computer at home, with numerous distractions (such as children, television, and the telephone) or at the of?ce (where distractions include work, colleagues, and phone calls.) A consumer can also be distracted by other online activities like e-mail, instant messaging, or other Web sites. Such distractions can limit online consumer concentration. When consumers shop in the physical world, they must allocate most of their attention to that task. They must walk or drive to the stores, look through the products, interact with sales people, and make purchases. They are not able to perform any other activities such as pay their bills or answer their e-mails. On the other hand, online customers can take advantage of the multitasking capabilities of their computers and do more than shop. They can go back and forth between the Web store and their e-mail, interrupt their shopping experience to do something else in their home or of?ce, or take advantage of download delays to perform other tasks. This can distract their attention and decrease their concentration in the purchasing task. Concentration as a measure of ?ow has been found to positively in?uence the overall experience of computer users (Novak et al. 1998) and their intention to
Information Systems Research Vol. 13, No. 2, June 2002

use a system repeatedly (Webster et al. 1993). Also, we know that interruptions that limit concentration reduce Web users’ satisfaction with online shopping (Xia and Sudharshan 2000). We therefore expect that high concentration, measured with a four-item scale adapted from Ghani et al. (1991), would have a positive impact on intention to return: Hypothesis 1c. Concentration is positively related to intention to return. However, the online consumer is also a computer user. Our model would be incomplete if we did not account for the reasons computer users adopt a technology such as a speci?c Web site. So, we turn to the most widely used theoretical model for explaining system usage, the Technology Acceptance Model (TAM) (Davis 1989). TAM has been tested in many empirical studies that include user acceptance of word processors (Davis et al. 1989), spreadsheets (Mathieson 1991), e-mail (Szajna 1996), voice mail (Straub et al. 1995), and telemedicine technology (Hu et al. 1999). By treating a Web store as a technology system and the Web consumer as a computer user, we can apply TAM and test how well it predicts user intention to use the technology, i.e., the Web store. In our case, that would translate into customer intention to return to the store. According to TAM, attitude towards technology affects use of the technology. Two belief variables have an impact on attitude: perceived ease of use and perceived usefulness of the technology. Empirical research has shown that attitude towards the technology is not a signi?cant mediating variable (Venkatesh and Davis 1996, Venkatesh 1999) so we do not include it. We believe that the TAM variables can also be successfully applied in the context of online consumer behavior. Viewing the store as a Web site or system, we expect that when customers believe that using the Web site will enhance their shopping productivity (perceived usefulness) they will be more likely to return. We anticipate the same effect when customers believe that the Web site is easy to use (perceived ease of use): Hypothesis 2a. Perceived usefulness of the Web store is positively related to intention to return. Hypothesis 2b. Perceived ease of use of the Web store is positively related to intention to return.

209

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

While TAM, environmental psychology, and ?ow theory may seem too discrete to combine in one study, we believe that they can complement each other successfully to form a hybrid theoretical perspective (Shaw and Jarvenpaa 1997). First, environmental psychology and ?ow share the common variables of enjoyment/pleasure and perceived control/dominance, the emotional responses of this study. The TAM variables (perceived usefulness and perceived ease of use) and concentration provide the cognitive responses. All three theories have at their core a number of emotional and cognitive responses to the environment that in?uence an individual’s behavior. The context of online shopping and the dual nature of the Web customer as a computer user allows them to come together to help us better understand online consumer behavior. 2.3. Unplanned Purchases While customers may enter a store planning to make speci?c purchases, they often end up making unplanned purchases. There are four types of unplanned purchases (Stern 1962): (1) Pure impulse: These are purchases that are made for purely hedonic reasons and are usually characterized by: (a) spontaneity; (b) power, compulsion, and intensity; (c) excitement and stimulation; (d) disregard for consequences (Rook 1987). (2) Reminder effect: A stimulus reminds the consumer to buy a product he needs. (3) Suggestion effect: The customer purchases a product because of a promotion. (4) Planned impulse: The customer shops without any speci?c product(s) in mind. Unplanned purchasing behavior on the Web can be interesting and complex. While shopping trips in the physical world are often constrained by time and geographical location, Web consumers can buy at any time from anywhere, possibly increasing the number of products they buy on impulse. The ability to click on banner ads and be transported immediately to the store, in front of the product advertised could make it easier for consumers to make unplanned purchases. Finally, the Web provides more privacy than the physi-

cal world, thereby possibly enabling consumers to indulge in impulsive shopping behavior they would ?nd embarrassing of?ine. Some of the same emotional and cognitive responses that impact a customer’s intention to return should in?uence the number of unplanned purchases made. We do not believe, however, that the belief variables of TAM will determine unplanned purchases. TAM predicts intention to use a technology and subsequently the amount of usage. It does not differentiate between speci?c kinds of use or the results of such use, e.g., making unplanned purchases at a Web store. We know from ?ow research that intrinsic enjoyment can increase a user’s exploratory behavior (Ghani and Deshpande 1994). Also, in their comprehensive study of impulse buying, Beatty and Ferrel (1998) found that in-store browsing behavior might increase the urge to buy impulsively. If online consumers enjoy their shopping experience, they might engage in more exploratory browsing in the Web store leading to more unplanned purchases. In addition, Beatty and Ferrel found an increase in impulse purchasing urges for shoppers with positive feelings during shopping. Beatty and Ferrel’s work has shown the signi?cance of positive emotional responses for unplanned purchases. Their impulsive nature implies that they rely a lot on consumer feelings. So, we believe that shopping enjoyment will induce unplanned purchases online. Hypothesis 3a. Consumers with higher shopping enjoyment are more likely to make unplanned purchases. Strategically placed promotions and displays throughout a store, can increase unplanned purchases and overall sales (Inman et al. 1990). Online, customers have high levels of control over what they see and do through the use of search engines, intelligent agents, and recommender systems (Baty and Lee 1995, Hoffman and Novak 1996 and 1997, Sheth and Sisodia 1997). This increased control allows consumers more choice over the advertising and promotional material they are exposed to, reducing unplanned purchases (Draft 1993, Rust and Oliver 1994, Shell 1994, Raman and Leckenby 1995, Burke 1997, Raman 1997, Lohse and Spiller 1998a). In this study, we measure perceived control. If it is
Information Systems Research Vol. 13, No. 2, June 2002

210

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

equivalent to actual consumer control, then, as discussed above, it may lead to a decrease in unplanned purchases. Alternatively, the perception of being in control could contribute to a positive experience that could increase unplanned purchases. We believe that perceived control online would re?ect actual control, so we expect that it would impact unplanned purchases negatively. Hypothesis 3b. Consumers with higher perceived control are less likely to make unplanned purchases. We know that exposure to marketing promotions increases unplanned purchases (Assael 1992). Consumers that are able to focus their attention at a Web store should also be more likely to notice marketing promotions on the site. If consumers are not paying full attention to the contents of the Web site when buying online, they are less likely to notice products that they might otherwise buy on impulse. Hypothesis 3c. Consumers with higher concentration are more likely to make unplanned purchases. 2.4. Determinants of Emotional and Cognitive Responses Once we know how certain emotional and cognitive responses to a Web store effect a consumer’s purchasing behavior, the next logical step is ?guring out how to elicit the right responses. To gain insight, the study looked at the effect of several variables that can determine a consumer’s emotional and cognitive responses. 2.4.1. Product Involvement. Product involvement and its measurement have been the source of considerable research and debate since Personal Involvement Inventory (PII) was ?rst proposed and analyzed by Zaichkowsky (1985). While there have been many variations on the de?nition of involvement (Zaichkowsky 1985, Greenwald and Leavitt 1984, Mitchell 1981, Park and Mittal 1985), it is generally accepted that involvement is: (a) a person’s motivational state (i.e., arousal, interest, drive) towards an object where (b) that motivational state is activated by the relevance or importance of the object in question (Mittal 1989). In the current study, the Revised Personal Involvement Inventory (RPII) proposed by McQuarrie and Munson (1992) was used to measure product involvement for
Information Systems Research Vol. 13, No. 2, June 2002

books, i.e., how interested consumers are in books and how important books are to them. While involvement with advertisements (Andrews and Durvasula 1991) and with the purchase process (Slama and Tashchian 1985) is important, the study focuses on involvement with the product. We are not concerned with how consumers are attracted to a Web site but with what happens while they visit it. Also, our three emotional and cognitive responses of perceived control, shopping enjoyment, and concentration give us a good indication of the involvement of the consumer with the purchase process. As a result, we only measure product involvement. A consumer’s involvement with the product(s) sold by a Web-based company can have an effect on the consumer’s experience and behavior. Involvement, measured simply as importance of the Web to the consumer, also had a strong effect on the primary antecedents of ?ow (Novak et al. 2000). We expect that customers with higher product involvement will have a more positive shopping experience due to their increased interest in the product. This relationship should be true for shopping enjoyment and concentration but not for perceived control. Perceived control is more of a result of the interaction with the Web site and its features, rather than the type of product purchased. Hypothesis 4a. Product involvement is positively related to shopping enjoyment. Hypothesis 4b. Product involvement is positively related to concentration. 2.4.2. Demographics. Demographic variables have been used extensively in market research to segment the customer population for better marketing strategies (Engel et al. 1990, Assael 1992). In some cases, demographics variables may determine some consumer behavior on the Web (Korgaonkar and Wolin 1999). For parsimony, we do not hypothesize on any possible differences in online consumer emotional and cognitive responses due to gender or age. 2.4.3. Web Skills. One of the most important antecedents of ?ow is the level of skills of the individual (Csikszentmihalyi 1975 and 1977, Ghani et al. 1991, Trevino and Webster 1992, Webster et al. 1993, Ghani

211

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

and Deshpande 1994, Hoffman and Novak 1996, Novak et al. 2000). Skills are measured as perceived by the user and not through observation or a standardized test. Therefore, they are similar to computer selfef?cacy, de?ned as “an individual judgment of one’s capability to use a computer” (Compeau and Higgins 1995). Computer self-ef?cacy has been found to affect computer use, often through its effect on the emotional state of the user by, for example, reducing his computer anxiety (Marakas et al. 1998). Similarly, skills have been consistently found (along with challenges, discussed below) to be signi?cant antecedents to ?ow. Therefore, we expect, that as online consumers perceive their Web skills to be higher, they will be more likely to have positive emotional and cognitive responses to the Web store they visit. Hypothesis 5a. Perceived skills are positively related to perceived control. Hypothesis 5b. Perceived skills are positively related to shopping enjoyment. Hypothesis 5c. Perceived skills are positively related to concentration. 2.4.4. Product Search Mechanisms. Web stores have responded to the call for customer control by providing various site features like internal search engines and recommender systems to enable consumers to easily ?nd what they need, learn more about it, and quickly purchase it. There have been very few attempts to create a framework or typology of Web-based customer decision-support systems (O’Keefe and McEachern 1998). We differentiated search mechanisms according to the two types of information they use: nonvalue-added and value-added. Nonvalue-added information consists of all information that is publicly available and standardized, such as objective information that describes the product sold. For a bookstore, such information can be the book title, author, and publisher. Value-added information consists of all information that is generated by the NEO. It is subjective and often proprietary to the entity that generates it. Value-added information can be created by the NEO (such as a weekly bestseller list), from a third party (such as hyperlinks to online reviews of products

by independent parties or Oprah’s Book Club selections), or from the customers themselves (such as customer reviews available to other customers). We know that product search may be undertaken simply because it is fun as in window shopping. Also, customers often engage in ongoing information gathering independent of speci?c needs or purchase decisions (Bloch et al. 1986). For them, value-added information can prove interesting and helpful. Whether it is reviews of books by other customers or stock recommendations by analysts, customers can enjoy this wealth of information about products and services not always available in the physical world. In general, we expect that the more enriched and satisfying the shopping experience is, the more likely customers are to experience high levels of perceived control, shopping enjoyment, and concentration. Because product information search can be a fun-seeking experience (Bloch et al. 1986), we expect that the use of value-added search mechanisms can make the shopping experience more ful?lling and enjoyable. Access to such information can improve consumer decision making through complex, nonlinear, and nondirected queries (Hoffman et al. 1995) and can be an important incentive for people to shop online (Jarvenpaa and Todd 1997a and b) by providing more control over their purchases. Such mechanisms may also help customers concentrate better because they require active customer participation and interpretation of the valueadded information. Hypothesis 6a. The use of value-added search mechanisms is positively related to perceived control. Hypothesis 6b. The use of value-added search mechanisms is positively related to shopping enjoyment. Hypothesis 6c. The use of value-added search mechanisms will be positively related to concentration. 2.4.5. Challenges. Along with individual skills, the challenges presented by an activity are the most important predictors of ?ow (Csikszentmihalyi 1975 and 1977, Ghani et al. 1991, Trevino and Webster 1992, Webster et al. 1993, Ghani and Deshpande 1994, Hoffman and Novak 1996, Novak et al. 2000). Positive challenges from the shopping experience on the Web can affect consumer responses to that experience. It is
Information Systems Research Vol. 13, No. 2, June 2002

212

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

important to stress here that we are looking at positive challenges, such as those presented by a sport or a game. We are not looking at negative challenges that are due to problems like dif?cult navigation, slow download time, or product unavailability. A Web site can be perceived by a consumer as positively challenging in the same way an electronic game is challenging. The process of discovering information and using special Web site features can feel challenging and ambitious with positive effects. Hypothesis 7a. The level of challenges of a Web store is positively related to perceived control. Hypothesis 7b. The level of challenges of a Web store is positively related to shopping enjoyment. Hypothesis 7c. The level of challenges of a Web store is positively related to concentration. The complete theoretical framework (Figure 1) is also a nomological network used to test the construct validity of our variables. Constructs are considered nomologically valid if, as measured in our study, they predict or are predicted by other constructs consistent with past research (Cronbach 1971, Bagozzi 1990, Straub et al. 1995).

traditional “brick and mortar” company Books-AMillion. Booksamillion.com was selected for several reasons: (1) The site is comprehensive and functional, eliminating any Web site problems. (2) Booksamillion.com is a relatively unknown online bookseller that guaranteed us enough ?rst-time customers for our sample. (3) The site, unlike booksellers like Amazon.com, primarily sells books. Books are a commodity—a fact that helps eliminate variance in product quality (the actual physical book, not the quality of its content). Though the site sells a few other items, such as coffee and magazines, the main focus of the store is books. 3.1. Study Design Obtaining permission to directly survey Booksamillion.com customers proved to be too dif?cult. Companies are reluctant to participate in research projects that interrupt customers’ shopping or the ?rm’s internal operations. So, we simulated actual customers by recruiting subjects, asking them to visit the Booksamillion.com site and then having them answer our questionnaire. The fact that subjects did not visit the Web site on their own volition is de?nitely a limitation of the study. However, while it is possible that their intention to return would differ when directed to visit the site by the researcher, we do not believe that the difference is large enough to compromise the generalizability of our results. To measure unplanned purchases we needed to know which purchases consumers had initially planned to make when they visited the store. We simulated an initial purchase plan by providing subjects with a $10 gift certi?cate. Subjects were asked to ?ll out the ?rst part of the questionnaire for a $10 gift certi?cate to an online bookstore, which they would visit upon completion. The ?rst part asked them questions unrelated to their particular visit, including how many (if any) books they intended to buy when they visited the online bookstore. The $10 gift certi?cate ensured us that some of the subjects would form an initial purchase plan of spending the gift certi?cate at the store. After visiting the site, we asked the subjects to immediately return and answer the second part of the questionnaire, where we asked about their shopping

3. Methods
To test our nomological network, an empirical study using an online questionnaire was administered to actual Web customers. The Web store selected for this study is Booksamillion.com, the online branch of the

Figure 1

Theoretical Framework

* All relationships are hypothesized positive except Perceived Control → Unplanned Purchases, which is negative.

Information Systems Research Vol. 13, No. 2, June 2002

213

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

experience and their actual purchases. Any purchases made in addition to what they had initially said they would buy were unplanned purchases. Unplanned purchases were calculated using the number of books purchased over their initial purchase plan, instead of the amount (in dollars) spent over the $10 gift certi?cate. We felt that it was a more accurate measure because customers most often go to a store with a plan to buy a number of books and not to spend an amount of money. Because our hypotheses deal with the likelihood of making unplanned purchases, we created a binary variable with the value zero if a customer made no unplanned purchases and the value one if a customer made some unplanned purchases. 3.2. Implementation The sample was recruited from the database of Dynamic Logic, Inc., an online market research ?rm that specializes in the measurement of the effectiveness of online ad. campaigns through the use of online questionnaires. Dynamic Logic also hosted the questionnaire using its patented technology that features randomized order of questionnaire items (to eliminate fatigue effect), live online results, and automated data collection. The sample was randomly selected from the Dynamic Logic database of registered participants who were e-mailed an invitation to participate in an online questionnaire for a $10 gift certi?cate to an online bookstore, $10 in cash (via check), and the chance to win $1,000 in cash. Each subject was asked which online bookstores he had visited in the past and if Books amillion.com was among them; the subject was screened out, leaving only new customers. They ?lled out the ?rst part of the questionnaire and were presented with a gift certi?cate code and a button. When they clicked it, a new window opened with the Books amillion.com site. They were instructed to visit Books amillion.com and then immediately return to the questionnaire to ?ll out the second part for an additional $10 in cash and a chance to win $1,000. We used a cash incentive for the second part of the questionnaire, instead of another gift certi?cate, so as not to in?uence the respondents’ answer to the question on their intention to return to the Web store. Our research instrument asked about the consumers’ experience during that speci?c visit to the store.

This increased the validity of our measures. Because we were not working with Booksamillion.com, we were unable to use log ?les to measure behavioral variables such as the number of purchases. We relied on self-reports for such measures. Subjects were not allowed to proceed unless they ?lled out the entire questionnaire (see Appendix C). 3.3. Sample and Descriptive Statistics The questionnaire ran for one week. The ?rst part of the questionnaire was completed by 300 subjects, and 280 of those ?lled out the second part, a retention rate of 93.3%. All 280 subjects were new customers to Books amillion.com. There were 12 international respondents, mainly from Canada. 63% of the sample was female, and about 50% of the subjects were 30 to 50 years old. 77% of the subjects had at least some college education, and 48% reported an income of over $50,000. Almost half of the subjects reported having children under the age of 18, and 68% reported spending over 15 hours per week online. Almost 50% of the subjects reported buying books online at least four times a year. When asked to rate loyalty to a speci?c online bookstore on a scale from one to six, many subjects (over 45%) reported low customer loyalty (values of one or two on a six-point scale). The majority of respondents (73.2%) reported that they intended to buy no books or that they did not know how many they would buy. There were 20% who wanted to purchase one book. After visiting the site, almost half of the subjects said they purchased at least one book, with almost 10% purchasing more than one. About 32% of the subjects made an unplanned purchase; 24% bought one more book than planned; 6% bought two more; and four subjects bought more than two extra books. Very few purchased crossselling items such as coffee (4 subjects) or magazines (17 subjects). Many respondents (65%) spent 20 minutes or less at Booksamillion.com, indicating an ef?cient purchasing process. Subjects seemed relatively satis?ed with their experience, with 33% reporting they would return to Booksamillion.com (giving six or seven on a seven-point scale). Only 14% said that they would not return (giving one or two on the same scale). 3.4. Results Based on prior research, we expected the ?ow factors as well as the TAM factors to be correlated. Therefore,
Information Systems Research Vol. 13, No. 2, June 2002

214

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

we factor analyzed our scales using principal components extraction with direct oblimin rotation, an appropriate method when there is reason to expect the factors to be correlated (Pedhazur and Pedhazur 1991). This analysis validated our instrument and the multiitem scales, as shown in the Appendix. All scales had acceptable Cronbach’s alpha values (Nunnally 1967), as seen in Table 1. The testing of the hypotheses that follows is intended to demonstrate further validation of the instrumentation, as discussed earlier. If the constructs perform as predicted by theory, then we can infer that the measurement of the constructs is nomologically valid. In testing the model, we were interested in seeing the difference in explanatory power between the part of the model that considers the online consumer as a shopper and that which considers the consumer as a computer user. Therefore, we tested the three ?ow variables separately from the TAM variables. We then tested the TAM variables independently. Finally we tested all variables together in a single model. Table 2 shows the linear regression model with the three ?ow variables. After removing four outliers that had standardized residuals with absolute values greater than 3.0 (Chatterjee et al. 1995), the ?nal model had an excellent ?t and explained a large percentage of variance with a R2 0.472. Only the coef?cient for

shopping enjoyment was statistically signi?cant. Therefore, Hypothesis 1a is supported but Hypotheses 1b and 1c are not.2 Table 3 shows the linear regression model with the TAM variables. After removing the same four outliers, the model had a great ?t and explained about the same variance (R2 0.489). Only the coef?cient for perceived usefulness was signi?cant, while that for perceived ease of use was not. Therefore, only Hypothesis 2a is supported. Comparing the two models, we saw very little difference in their predictive ability for intention to return. We ran a combined regression model, as seen in Table 4, after removing the same four outliers. The model had good ?t and, as expected, explained more variance (R2 0.546). The same coef?cients for shopping enjoyment and perceived usefulness were signi?cant, so the same hypotheses are supported. We ran a logistic regression on the binary variable of unplanned purchases to test the effects of perceived control, shopping enjoyment, and concentration. Table 5 shows the results. The chi-square value shows that

Table 2

Linear Regression Model for Intention to Return with the Flow Variables B 0.866** 0.104 0.705** 0.045 81.069 (p Std. Error 0.327 0.064 0.080 0.070 0.01)—where ** p ?

Variable Constant Perceived Control Shopping Enjoyment Concentration N *p 276; R 2 0.05, o p 0.472; F 0.1

Table 1

Summary Statistics and Cronbach’s Alpha Values for All Scales Mean S.D. 1.46 1.52 1.52 1.50 1.44 1.01 1.61 1.57 1.77 1.31 Cronbach’s Alpha

0.090 0.608** 0.037 0.01,

Concentration Shopping Enjoyment Perceived Control Skills Challenges Involvement Perceived Ease of Use Perceived Usefulness Intention to Return Value-Added Used

4.52 4.30 4.98 5.31 2.96 6.13 5.16 4.15 4.59 1.37

0.910 0.944 0.813 0.918 0.803 0.929 0.927 0.924 N/A N/A

Table 3

Linear Regression Model for Intention to Return with the TAM Variables B 1.130** 0.695** 0.119o 130.651 (p Std. Error 0.259 0.066 0.064 0.01)—where ** p b

Variable Constant Perceived Usefulness Perceived Ease of Use N *p 276; R 2 0.05, o p 0.489; F 0.1

0.621** 0.109o 0.01,

2 For all regression models in the study, low VIF values indicate low collinearity, the standardized residuals are normally distributed, and there is no heteroscedasticity.

Information Systems Research Vol. 13, No. 2, June 2002

215

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

Table 4

Linear Regression Model for Intention to Return with All Variables B 0.701* 0.021 0.400** 0.040 0.465** 0.002 64.986 (p Std. Error 0.308 0.069 0.088 0.066 0.074 0.073 0.01)—where p b

Table 5

Logistic Regression for Unplanned Purchases with the Flow Variables B Std. Error 0.614 0.123 0.139 0.111 Exp (B) 0.051 1.229 o 1.152 1.130

Variable Constant Perceived Control Shopping Enjoyment Concentration Perceived Usefulness Perceived Ease of Use N 276; R 2 o 0.05, p 0.1 0.546; F

0.018 0.345** 0.033 0.415** 0.002 0.01, * p

Constant Concentration Shopping Enjoyment Perceived Control Initial 2 Log Likelihood 17.724**

2.974 0.206o 0.142 0.123 350.136

2 Log Likelihood Chi-square df N ** p 3

332.413

the model has good ?t but none of the coef?cients is signi?cant. So, Hypotheses 3a, b, and c are not supported. We used linear regression models to test the effect of individual and environmental factors on perceived control, shopping enjoyment, and concentration. Table 6 shows the results of the regression model for concentration. The model had good ?t and explained about 21% of the variance. The coef?cients for product involvement, challenges, and skills were signi?cant and positive, providing support for Hypotheses 4b, 7c, and 5c. The coef?cient for value-added use was in the right direction, but not signi?cant. The regression model for shopping enjoyment had an even better ?t and explained more variance with a R2 0.281, as seen in Table 7. The coef?cients for product involvement, challenges, skills, and value-added use were highly signi?cant and positive, providing support for Hypotheses 4a, 7b, 5b, and 6b, respectively. The regression model for perceived control had a moderately good ?t (F 2.898, p 0.006) but no signi?cant coef?cients and a R2 0.070 that indicated the model explained too little variance. Therefore, we could not verify Hypotheses 7a and 8a. Figure 2 shows our nomological network but includes only the relationships that were supported by our data.

280; Cox and Snell R 2 0.061; Nagelkerke R 2 0.01, * p 0.05, o p 0.1

0.086—where

Table 6 Variable

Linear Regression Model for Customer Concentration B 0.047 0.407** 0.221** 0.138* 0.130o 0.088 0.014 0.008 10.321 (p Std. Error 0.646 0.083 0.056 0.056 0.070 0.074 0.176 0.007 0.01)—where ** p b

Constant Involvement Challenges Skills Value-Added Use Nonvalue-Added Use Gender Age N *p 280; R 2 0.05, o p 0.210; F 0.1

0.280** 0.216** 0.142* 0.116o 0.074 0.005 0.066 0.01,

Table 7 Variable

Linear Regression Model for Customer Shopping Enjoyment B 0.086 0.328** 0.358** 0.182** 0.240** 0.042 0.171 0.011o 15.095 (p Std. Error 0.640 0.082 0.055 0.055 0.069 0.073 0.174 0.007 0.01)—where ** p b

4. Discussion
An important contribution of our study is the testing and validation of metrics for consumer behavior on the

Constant Involvement Challenges Skills Value-Added Use Nonvalue-Added Use Gender Age N *p 280; R 2 0.05, o p 0.281; F 0.1

0.218** 0.338** 0.180** 0.207** 0.034 0.054 0.090o 0.01,

216

Information Systems Research Vol. 13, No. 2, June 2002

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

Figure 2

Relationships Supported by Empirical Study

All relationships are statistically signi?cant at the p

0.05 level.

Web as called for by Straub and Watson (2001). In addition to the fact that all the scales used in the study showed high reliability, those of shopping enjoyment, perceived usefulness, involvement, challenges, and skills also demonstrated high nomological validity. When tested in a nomological network derived from prior research, they generally behaved as demonstrated in past studies. Therefore, we believe that future research in online consumer behavior can use these metrics with some assurance. Testing them against other important factors not included in this study will advance our understanding of online consumer behavior. From a substantive point of view, we examined how emotional and cognitive responses to the shopping experience on the Web can in?uence online consumer behavior, speci?cally, intention to return and unplanned purchases for new customers. Looking at the online consumer both as a shopper and a computer user, by blending theoretical paradigms from psychology, marketing, and information systems, we con?rmed the dual nature of the online consumer as a traditional shopper and a computer user. Our results show that both enjoyment of the shopping experience (a psychological/marketing variable) and perceived usefulness of the Web site (an IS variable for system use) are important for a new customer’s intention to return. Furthermore, the signi?cance of perceived usefulness shows that the Technology Acceptance Model can be successfully applied, even when the behavior in question is not one of pure system usage. When tested
Information Systems Research Vol. 13, No. 2, June 2002

alone, without the ?ow variables, perceived usefulness explained 49% of variance of intention to return, demonstrating even further the robustness of TAM in this context. Our results also con?rmed prior TAM research that found that perceived usefulness was a more important predictor of intended system usage than perceived ease of use (Davis 1989, Hu et al. 1999). In accordance with research that has demonstrated the utilitarian nature of online consumers (Jarvenpaa and Todd 1997a and b, Clawson 1993), we found that a consumer’s belief about the usefulness of a Web store (a cognitive response) can determine their future visits. However, we also found that an emotional response to the Web site can also have the same effect. Even though consumers may not expect to be entertained when they shop online, if they do enjoy their experience, they are more likely to return to the Web store. These results show that online consumers are not purely utilitarian, valuing only ef?ciency in shopping, but they can also enjoy shopping online enough to make them return. If this means that online consumers are more similar to of?ine ones than previously thought, researchers of online consumers may successfully use previous research done of?ine, though with caution. Our results on unplanned purchases were surprising. We found no relationship between unplanned purchases and the ?ow variables of shopping enjoyment, concentration, and perceived control. A possible reason for the weak results could be our $10 gift certi?cate. One could argue that it is dif?cult to purchase anything with $10, even if it is a book. This may have prevented some of our subjects from buying anything at all in the ?rst place. Unfortunately, the high cost of such an incentive made it infeasible for our study. It is also possible that there are other variables that can explain unplanned purchases that we did not include in our framework. Future research needs to reexamine the relationships we tested as well as introduce new possible predictors. Our results also con?rmed some prior research on ?ow. We found that perceived Web skills and positive challenges are positively related with shopping enjoyment and concentration of online consumers. The importance of Web skills is also consistent with work done in computer self-ef?cacy (Marakas et al. 1998)

217

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

and stresses once again the “computer user” part of the dual identity of the online consumer. The more con?dent and comfortable consumers feel with the Web site, the more likely it is that they will enjoy it. The nonsigni?cance of the relationship of ?ow variables with unplanned purchases and intention to return (with the exception of shopping enjoyment) puts into question the use of ?ow in the context of online shopping. Unlike the study by Novak et al. (2000), which de?ned and measured the ?ow state for general Web users, we examined the effect of ?ow variables on consumer behavior during a particular store visit. It may be the case that for online consumers, not simply Web users, a multidimensional ?ow construct does not explain their behavior, while a simple construct like shopping enjoyment does. Therefore, we urge cautious use of ?ow in online consumer behavior research. We also found that the use of value-added search mechanisms may in?uence the experience of online consumers that can in turn affect their behavior. The use of a utilitarian tool such as a search engine is positively related with shopping enjoyment, an emotional response. However, we found no relationship between using value-added search mechanisms and perceived control or concentration. A possible implication is that Web site features that are intended to be mere tools for the consumer to use might have a strong emotional impact. An alternative explanation is that the more consumers enjoy themselves, the more they use valueadded search mechanisms. Our results also have important practical implications for companies. The dual nature of the online consumer as a traditional shopper and as a computer user means that appropriate interface, navigational structure, and other elements of human-computer interaction may be just as important to retaining customers as good customer service and lower prices. Many Webbased companies have taken large steps to increase the convenience to their customers by providing features like express checkouts and recommender systems. However, our results have shown that emotional experiences such as shopping enjoyment can help retain customers. Therefore, online stores should provide both utilitarian value as well as hedonic value to their customers, partly by providing them with value-added search mechanisms.

Our study was not without limitations. Because our subjects had registered with an online market research company and took multiple online questionnaires, they may be more Web-savvy than the average user. Unfortunately, to recruit subjects online, they must be comfortable enough with the Web to be recruited and to participate. Our “Web skills” scale shows a high mean of 5.3 on a seven-point scale, but also a high variance of 2.25. That justi?es some cautious generalizations of our results. The same applies to our scale for “perceived ease of use” with a mean of 5.16 and a variance of 2.6. An important variable missing from our framework is planned purchases of ?rst-time customers. How do customers decide to visit a speci?c Web store for a planned purchase, and what factors determine whether they will actually make that purchase at that store? Our study design did not capture actual customers as they came to a store for the ?rst time. Instead, we had to simulate new customers by directing our subjects to the Web store. A future study that captures actual new customers would be able to better explain the factors behind customer acquisition and planned purchasing. There are also many additional individual and environmental factors that can determine a consumer’s emotional and cognitive responses. In environmental psychology, physical stimuli like colors as well as personality traits can in?uence an individual’s emotional state (Mehrabian and Russsel 1974). In marketing, there have been studies on a variety of individual characteristics such as motivation and knowledge as well as a broad range of environmental variables including family, culture, and social class (Engel et al. 1990). There has also been some research on the antecedents of the two belief variables in the TAM (e.g., Venkatesh and Davis 1996). Our study did not consider any of these variables and we urge other researchers to do so. We consider this study a stepping stone on the road to understanding online consumer behavior. Many issues remain unresolved and many questions unanswered. We took a ?rst step at studying the consumer experience on the Web and found a number of metrics to be reliable and nomologically valid. Future research can use these metrics to explain how and why consumers think, act, and feel when shopping on the Web.
Information Systems Research Vol. 13, No. 2, June 2002

218

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

5.

Appendices
Structure Matrix from Principal Component Analysis with Direct Oblimin Rotation
FACTORS 1 (Concentration) 0.886 0.861 0.860 0.883 0.559 0.662 0.648 0.658 0.473 0.929 0.862 0.635 0.872 0.842 0.806 0.928 0.914 0.932 0.725 0.899 0.814 0.898 0.899 0.945 0.914 0.914 2 (Enjoyment) 3 (Perceived Control) 4 (Challenges) 5 (Perceived Usefulness) 6 (Perceived Ease of Use)

Appendix A.

ITEMS . . . I was absorbed intensely in the activity. . . . My attention was focused on the activity. . . . I concentrated fully on the activity. . . . I was deeply engrossed in the activity. . . . I found my visit interesting. . . . I found my visit enjoyable. . . . I found my visit exciting. . . . I found my visit fun. . . . I felt confused. . . . I felt calm. . . . I felt in control. . . . I felt frustrated. . . . challenged me to perform to the best of my ability . . . provided a good test of my skills . . . stretched my capabilities to the limits Using the site can improve my shopping performance. Using the site can increase my shopping productivity. Using the site can increase my shopping effectiveness. I ?nd using the site useful. Learning to use the site would be easy for me. My interaction with the site is clear and understandable. It would be easy for me to become skillful at using the site. I ?nd the site easy to use. I am very skilled at using the Web. I know how to ?nd what I want on the Web. I know more about using the Web than most users.

7 (Skills)

Appendix B.

Correlation Matrix
Perceived Ease of Use ValueAdded Used 0.06 0.22** 0.21** 0.15* 0.28** 0.21** 0.06 0.12*

Pearson Bivariate Correlations Perceived Usefulness 0.39** 0.68** Perceived Control 0.01 0.70** 0.55** Enjoyment 0.34** 0.66** 0.75** 0.58** Concentration 0.23** 0.43** 0.47** 0.25** 0.65** Involvement 0.02 0.31** 0.19** 0.12* 0.27** 0.33** 0.09 Intention to Return 0.27** 0.47** 0.62** 0.39** 0.62** 0.39** 0.09 0.18** 0.19**

Skills 0.10 0.24** 0.14* 0.06 0.18** 0.14*

Challenges Perceived Ease of Use Perceived Usefulness Perceived Control Enjoyment Concentration Skills Involvement Value Added where ** p 0.01, * p 0.05

0.01

Information Systems Research Vol. 13, No. 2, June 2002

219

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

Appendix C.

Survey Instrument

Tender Times

1. Intention to Return We measured customer intention to return with the following single item on a seven-point Lickert scale ranging from one—extremely unlikely to seven—extremely likely: “How likely is it that you will visit Booksamillion.com again in the future?” 2. Unplanned Purchases We used two questions to determine the number of unplanned purchases. The ?rst question asked how many books the customer intended to buy when she would be directed to the bookstore site with the $10 gift certi?cate. The possible answers were none, one, two, three, four or more, and “I don’t know how many” books. The second question asked how many books they actually bought as well as how many of the other items sold on Booksamillion.com, such as coffee and magazines, they purchased. We calculated the number of unplanned purchases by subtracting the ?rst question from the second one. For those subjects that reported “I don’t know how many books” in the ?rst question, we considered their initial planned purchases as zero. We then created a binary variable that takes the value zero for subjects who did not make unplanned purchases and the value one for those who did. 3. Search Mechanisms Used A multiple response question was used to determine which search mechanisms the customer used, and a single response question asked customers which one of the search mechanisms was the most helpful. Once again, due to the lack of resources, we depended on self-reports rather than Web logs to determine which search features customers used. We separated the search mechanisms available to the customers into value-added and nonvalue-added. We then created a variable that measured the use of each type of search mechanism by adding the number of search mechanisms the customers reported they had used. Value-Added Search Mechanisms Bestsellers Bargains Reviews Weekly Specials New and Notable Easter Great Movie Reads Campaign 2000 Oprah’s Book Club Seen in NY Times Graduation Mother’s Day Harry Potter Fans Autographed Books Popular Business All Literary Awards Music Bestsellers Nonvalue-Added Search Mechanisms Book Search Browse Subjects Upcoming Titles All Authors Road to El Dorado Left Behind Series eBook: Riding the Bullet Star Wars Pokemon MLB Showdown 2000 Bible Covers Accessories Magic the Gathering Yo Yos Furby Tae Bo Coin Collectibles

Don’t Sweat Series Coffee Magazines

Classi?cation of search mechanisms on Booksamillion.com 4. Multi-Item Scales Following are the multi-item scales used in the study and the source in prior literature from which they were adapted. All items used a seven-point Lickert scale: Concentration/Attention Focus (Ghani et al. 1991): During my last visit to Booksamillion.com . . . . . . I was absorbed intensely in the activity. . . . My attention was focused on the activity. . . . I concentrated fully on the activity. . . . I was deeply engrossed in the activity. Shopping Enjoyment (Ghani et al. 1991): During my last visit to Booksamillion.com . . . . . . I found my visit interesting. . . . I found my visit enjoyable. . . . I found my visit exciting. . . . I found my visit fun. Perceived Control (Ghani et al. 1991): During my last visit to Booksamillion.com . . . . . . I felt confused (reversed). . . . I felt calm. . . . I felt in control. . . . I felt frustrated (reversed). Web Skills (Novak et al. 1998): I am very skilled at using the Web. I know how to ?nd what I want on the Web. I know more about using the Web than most users. Challenges (Novak et al. 1998): Using Booksamillion.com challenged me to perform to the best of my ability. Using Booksamillion.com provided a good test of my skills. Using Booksamillion.com stretched my capabilities to the limits. Perceived Usefulness (Venkatesh and Davis 1996): Using Booksamillion.com can improve my shopping performance. Using Booksamillion.com can increase my shopping productivity. Using Booksamillion.com can increase my shopping effectiveness. I ?nd using Booksamillion.com useful. Perceived Ease of Use (Venkatesh and Davis 1996): Learning to use Booksamillion.com would be easy for me. My interaction with Booksamillion.com is clear and understandable. It would be easy for me to become skillful at using Books amillion.com. I ?nd Booksamillion.com easy to use. Product Involvement (McQuarrie and Munson 1992): We would like to know how interested you are in books. Please use the series of descriptive words listed below to indicate your level of interest in books: Important Irrelevant Means a lot to me O O O O O O O Unimportant O O O O O O O Relevant O O O O O O O Means nothing to me

220

Information Systems Research Vol. 13, No. 2, June 2002

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

Unexciting Dull Matters to me Boring Fun Appealing Of no concern to me

O O O O O O O

O O O O O O O

O O O O O O O

O O O O O O O

O O O O O O O

O O O O O O O

O O O O O O O

Exciting Neat Doesn’t matter to me Interesting Not fun Unappealing Of concern to me

References
Ajzen, Icek. 1991. The theory of planned behavior. Organ. Behavior and Human Decision Processes 50(2) 179–211. Alba, Joseph, John Lynch, Barton Weitz, Chris Janiszewski, Richard Lutz, Alan Sawyer, Stacy Wood. 1997. Interactive home shopping: Consumer, retailer, and manufacturer incentives to participate in electronic marketplaces. J. Marketing 61(3) 38–53. Andrews, J. C., S. Durvasula. 1991. Suggestions for manipulating and measuring involvement in advertising message content. R. H. Holman and M. S. Solomon, eds. Advances in Consumer Research, vol. 18. Association for Consumer Research. Anonymous. 1998. Repeat customers are scarce for most sites. Internet World 4(30) 14. Assael, Henry. 1992. Consumer Behavior and Marketing Action. PWSKENT, Boston, MA. Atkinson, J. W. 1964. An Introduction to Motivation. Van Nostrand, Princeton, NJ. Bagozzi, Richard P. 1980. Causal Methods in Marketing. John Wiley and Sons, New York. Bakos, Yannis. 1997. Reducing buyer search costs: Implications for electronic marketplaces. Management Sci. 43(12) 1676–1692. Bandura, A. 1982. Self-ef?cacy mechanism in human agency. Amer. Psychologist 37(2) 122–147. Baty, J. B., R. M. Lee. 1995. Intershop: Enhancing the vendor/customer dialectic in electronic shopping. J. Management Inform. Systems 11(4) 9–31. Beatty, Sharon E., M. Elizabeth Ferrell. 1998. Impulse buying: Modeling its precursors. J. Retailing 74(2) 169–191. Bellman, Steven, Gerald L. Lohse, Eric J. Johnson. 1999. Predictors of online buying behavior. Comm. ACM 42(12) 32–38. Berthon, Pierre, Leyland E. Pitt, Richard T. Watson. 1996. The World Wide Web as an advertising medium: Toward an understanding of conversion ef?ciency. J. Advertising Res. 36(1) 43–54. Bhatnagar, Amit, Sanjog Misra, H. Raghav Rao. 2000. On risk, convenience, and Internet shopping behavior. Comm. ACM 43(11) 98–105. Blakney, V. L., W. Sekely. 1994. Retail attributes: In?uence on shopping mode choice behavior. J. Managerial Issues 6(1) 101–118. Bloch, Peter H., Daniel Sherrell, Nancy Ridgway. 1986. Consumer search: An extended framework. J. Consumer Res. 13(1) 119–126. Burke, Raymond R. 1997. Do you see what I see? The future of virtual shopping. J. Acad. Marketing Sci. 25(4) 352–360. ——, Bari A. Harlam, Barbara E. Kahn, Leonard M. Lodish. 1992. Comparing dynamic consumer choice in real and computersimulated environments. J. Consumer Res. 19(1) 71–82. Chatterjee, Samprit, Mark S. Handcock, Jeffrey S. Simonoff. 1995. A

Casebook for a First Course in Statistics and Data Analysis. Wiley, New York. Clawson, Pat. 1993. Study: Consumers want interactive TV. Electronic Media (August 23) 24–25. Compeau, D. R., C. A. Higgins. 1995. Computer self-ef?cacy: Development of a measure and initial test. MIS Quart. 19(2) 189– 211. Cronbach, L. J. 1971. Test validation. R. L. Thorndike, ed. Educational Measurement. American Council on Eduaction, Washington, D.C. Csikszentmihalyi, Mihaly. 1975. Play and intrinsic rewards. Humanistic Psych. 15 41–63. ——. 1977. Beyond Boredom and Anxiety. Jossey-Bass, San Fransisco, CA. ——, Isabella Selega Csikszentmihalyi. 1988. Optimal Experience: Psychological Studies of Flow in Consciousness. Cambridge University Press, Cambridge, U.K. Davis, Fred D. 1989. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quart. 13(3) 319–340. ——, Richard P. Bagozzi, Paul R. Warshaw. 1989. User acceptance of computer technology: A comparison of two theoretical models. Management Sci. 35(8) 982–1003. Draft, Howard C. 1993. The destiny of direct. Direct Marketing (December) 65–69. Dykema, Evie Black, Kate Delhagen, Carrie Ardito. 1999. Consumers catch auction fever. Forrester Res. Inc., Cambridge, MA. Eighmey, John. 1997. Pro?ling user responses to commercial Web sites. J. Advertising Res. 37(3) 59–66. ——, Lola McCord. 1998. Adding value in the information age: Uses and grati?cations of sites on the World Wide Web. J. Bus. Res, 41(3) 187–194. Engel, James F., Roger D. Blackwell, Paul W. Miniard. 1990. Consumer Behavior. Dryden Press, Chicago, IL. Fishbein, M., I. Ajzen. 1975. Belief, Attitude, Intention, and Behavior: An Introduction to Theory and Research. Addison-Wesley, Reading, MA. Forman, Andrew M., Ven Sriram. 1991. The depersonalization of retailing: Its impact on the “lonely” consumer. J. Retailing 67(2) 226–243. Gefen, David, Detmar Straub. 2000. The relative importance of perceived ease of use in IS adoption: A study of e-commerce adoption. J. Association for Inform. Systems 1(1) http://jais.isworld .org . Ghani, Jawaid A., Satish P. Deshpande. 1994. Task characteristics and the experience of optimal ?ow in human-computer interaction. The J. Psych. 128(4) 381–391. ——, Roberta Supnick, Pamela Rooney. 1991. The experience of ?ow in computer-mediated and in face-to-face groups. J. I. DeGross, I. Benbasat, G. DeSanctis, and C. M. Beath, eds. Proc. 12th Internat. Conf. Inform. Systems, New York, 229–237. Greenwald, A. G., C. Leavitt. 1984. Audience involvement in advertising: Four levels. J. Consumer Res. 11(1) 581–592. Heijden, Hans van der, Tibert Verhagen, Marcel Creemers. 2001. Pre-

Information Systems Research Vol. 13, No. 2, June 2002

221

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

dicting online purchase behavior: Replications and tests of competing models. Proc. 34th Hawaii Internat. Conf. System Sci., Maui, HI. Hoffman, Donna L., Thomas P. Novak. 1996. Marketing in hypermedia computer-mediated environments: Conceptual foundations. J. Marketing 60(3) 50–117. ——, ——. 1997. A new marketing paradigm for electronic commerce. The Inform. Soc. 13(January–March) 43–54. ——, ——, Patrali Chatterjee. 1995. Commercial scenarios for the Web: Opportunities and challenges. J. Comput. Mediated Comm. 1(3). http://www.ascusc.org/jcmc/ . Hu, Paul J., Patrick Y. K. Chau, Olivia R. Liu Sheng, Kar Yan Tam. 1999. Examining the technology acceptance model using physician acceptance of telemedicine technology. J. Management Inform. Systems 16(2) 91–112. Inman, J. J., L. McAlister, W. D. Hoyer. 1990. Promotion signal: Proxy for a price cut? J. Consumer Res. 17(June) 74–81. Jarvenpaa, Sirkka L., Peter A. Todd. 1997a. Is there a future for retailing on the Internet? R. A. Peterson, ed. Electronic Marketing and the Consumer. Sage, Thousand Oaks, CA. ——, ——. 1997b. Consumer reactions to electronic shopping on the World Wide Web. Internat. J. Electronic Commerce 1(2) 59–88. Korgaonkar, Pradeep K., Lori D. Wolin. 1999. A multivariate analysis of Web usage. J. Advertising Res. 39(2) 53–68. Li, Hairong, Cheng Kuo, Martha G. Russel. 1999. The impact of perceived channel utilities, shopping orientations, and demographics on the consumer’s online buying behavior. J. Comput. Mediated Comm. 5(2). http://www.ascusc.org/jcmc/ . Lohse, Gerald L., Peter Spiller. 1998a. Quantifying the effect of user interface design features on cyberstore traf?c and sales. CHI, Los Angeles, CA, 211–218. ——, ——. 1998b. Electronic shopping: Quantifying the effect of customer interfaces on traf?c and sales. Comm. ACM 41(7) 81–86. Maignan, Isabelle, Bryan A. Lukas. 1997. The nature and social uses of the Internet: A qualitative investigation. J. Consumer Affairs 31(2) 346–371. Malone, Thomas W., Joanna Yates, Robert I. Benjamin. 1987. Electronic markets and electronic hierarchies: Effects of information technology on market structure and corporate strategies. Comm. ACM 30(6) 484–497. Marakas, George M., Mun Y. Yi, Richard D. Johnson. 1998. The multilevel and multifaceted character of computer self-ef?cacy: Toward clari?cation of the construct and an integrative framework for research. Inform. Systems Res. 9(2) 126–163. Mathieson, Kieran. 1991. Predicting user intentions: Comparing the technology acceptance model with the theory of planned behavior. Inform. Systems Res. 2(3) 173–191. McQuarrie, Edward F., J. Michael Munson. 1992. A revised product involvement inventory: Improved usability and validity. J. F. Sherry and B. Sternthal, eds. Advances in Consumer Research, vol. 19. Association for Consumer Research, Provo, UT. Mehrabian, Albert, James A. Russel. 1974. An Approach to Environmental Psychology. MIT Press, Cambridge, MA. Miller, George A. 1956. The magical number seven, plus or minus

two: Some limits on our capacity for processing information. Psych. Rev. 63(March) 81–97. Mitchell, A. A. 1981. The dimensions of advertising involvement. K. B. Monroe, ed. Advances in Consumer Research, Association for Consumer Research, Provo, UT. Mittal, Banwari. 1989. A theoretical analysis of two recent measures of involvement. T. K. Srull, ed. Advances in Consumer Research, vol. 16. Association for Consumer Research, Provo, UT. Morris, Betsy. 1987. As a favorite pastime, shopping ranks high with most Americans. The Wall Street J. July 30, p. 13. Morrisette, Shelley, James L. McQuivey, Nicki Maraganore, Gordon Lanpher. 1999. Are Net shoppers loyal? The Forrester Rep. March. Novak, Thomas P., Donna L. Hoffman, Yiu-Fai Yung. 1998. Modeling the structure of the ?ow experience among Web users. Abstr. INFORMS Marketing Sci. Internet Mini-Conf., MIT. ——, ——, ——. 2000. Measuring the customer experience in online environments: A structural modeling approach. Marketing Sci. 19(1) 22–42. Nowlis, Stephen M., Deborah Brown McCabe. 2000. Online vs. offline consumer decision making: The effect of the ability to physically touch merchandise. Second Marketing Sci. Internet Conf., USC, Los Angeles, CA, Nunnally, Jum C. 1967. Psychometric Theory, McGraw-Hill, New York. O’Keefe, Robert, Tim McEachern. 1998. Web-based customer decision support systems. Comm. ACM 41(3) 71–78. Park, C. W., Banwari Mittal. 1985. A theory of involvement in consumer behavior: Problems and issues. J. N. Sheth, ed. Research in Consumer Behavior, vol. I. JAI Press, Greenwich, CT. Pedhazur, E., L. Pedhazur. 1991. Measurement, Design, and Analysis: An Integrated Approach. Lawrence Erlbaum and Associates, Hillsdale, NJ. Peterson, Robert A., Sridhar Balasubramanian, Bart J. Bronnenberg. 1997. Exploring the implications of the Internet for consumer marketing. J. Acad. Marketing Sci. 25 329–346. Pine, Joseph B., Don Peppers, Martha Rogers. 1995. Do you want to keep your customers forever? Harvard Bus. Rev. 73(2) 103–113. Quelch, John A., Lisa R. Klein. 1996. The Internet and international marketing. Sloan Management Rev. 37(3) 60–75. ——, Hiroaka Takeuchi. 1981. Nonstore marketing: Fast track or slow? Harvard Bus. Rev. 59(4) 75–84. Raman, Niranjan V. 1997. A qualitative investigation of Webbrowsing behavior. Adv. Consumer Res. 24 511–516. ——, John D. Leckenby. 1995. Advertising in interactive media: Some issues, some suggestions. C. S. Madden, ed. Proc. 1995 Conf. Amer. Acad. Advertising, Waco, TX, 247–251. Reichheld, Frederick F., W. Earl Sasser Jr. 1990. Zero defections: Quality comes to services. Harvard Bus. Rev. 68(5) 105–111. Rice, Marshall. 1997. What makes users revisit a Web site? Marketing News 31(6) 12. Rook, Dennis W. 1987. The buying impulse. J. Consumer Res. 14(September) 189–199. Rotter, J. B. 1966. Generalized expectancies for internal versus external control of reinforcement. Psych. Monographs: General and Appl. 80(1, Whole N. 609).

222

Information Systems Research Vol. 13, No. 2, June 2002

KOUFARIS Applying the Technology Acceptance Model and Flow Theory to Online Consumer Behavior

Rowen, Mark. 1999. Internet Retailing: The New Age of Retail. Prudential Securities, New York. Rust, Roland T., Richard W. Oliver. 1994. The death of advertising. J. Advertising 23(4) 71–78. Shaw, T., S. L. Jarvenpaa. 1997. Process models in information systems. A. S. Lee, J. Liebenau, and J. L. DeGross, eds. Information Systems and Qualitative Research. Chapman and Hall, London, U.K., 70–100. Shell, Adam. 1994. A map of the information highway. Public Relations J. 50(January) 27. Sheth, Jagdish N., Rajendra S. Sisodia. 1997. Consumer behavior in the future. R. A. Peterson, ed. Electronic Marketing and the Consumer. Sage, Thousand Oaks, CA. Slama, M. E., A. Tashchian. 1985. Selected socioeconomic and demographic characteristics associated with purchasing involvement. J. Marketing 49(Winter) 72–82. Stern, Hawkins. 1962. The signi?gance of impulse buying today. J. Marketing 26(April) 59–62. Straub, Detmar, Richard T. Watson. 2001. Transformational issues in researching IS and Net-enabled organizations. Inform. Systems Res. 12(4) 337–345. ——, Moez Limayem, Elena Karahanna-Evaristo. 1995. Measuring

system usage: Implications for IS theory and testing. Management Sci. 41(8) 1328–1342. Szajna, Bernadette. 1996. Empirical evaluation of the revised technology acceptance model. Management Sci. 42(1) 85–92. Tracy, Bernadette. 1998. e-Tailing: What Web customers really want. Avertising Age’s Bus. Marketing 83(11) 39–41. Trevino, Linda Klebe, Jane Webster. 1992. Flow in computermediated communication: Electronic mail and voice mail evaluation and impacts. Comm. Res. 19(5) 539–573. Venkatesh, Viswanath. 1999. Creation of favorable user perceptions: Exploring the role of intrinsic motivation. MIS Quart. 23(2) 239– 260. ——, Fred D. Davis. 1996. A model of the antecedents of perceived ease of use: Development and test. Decision Sci. 27(3) 451–481. Webster, Jane, Linda Klebe Trevino, Lisa Ryan. 1993. The dimensionality and correlates of ?ow in human-computer interaction. Comput. Human Behavior 9(4) 411–426. Xia, Lan, D. Sudharshan. 2000. An examination of the effects of cognitive interruptions on consumer online decision processes. Second Marketing Sci. Internet Conf., USC, Los Angeles, CA. Zaichkowsky, Judith Lynne. 1985. Measuring the involvement construct. J. Consumer Res. 12(3) 341–352.

Charles Stein?eld, Special Issue Editor. This paper was received on January 10, 2000, and was with the author 10 months for 4 revisions.

Information Systems Research Vol. 13, No. 2, June 2002

223


赞助商链接
相关文章:
技术接受模型(Technology Acceptance Model,TAM)
采用技术接受模型对 Online Shopping 接受进行研究,但...and user acceptance of information technology[J]....Consumer reactions to electronic shopping on the ...
公司理财复习题
to the firm's creditors and stockholders is ...operating cash flow. B. net capital spending. C...are correct according to pecking-order theory? I...
曼昆微观经济学习题
and to reduce the cost of healthcare (via various...1. Draw a circular-flow diagram. Identify the ...Relate your answer to the theory of comparative ...
公司金融英文试题
= $758.46 Earnings before interest and taxes =...What is the amount of the cash flow to ...on its consumer loans of 10 percent per year. ...
Consumer Behavior消费者行为学 每章节的Q&A 以及重点_...
to the study of the cause and effect ...improved Many consumer theory models are put ...Online consumer behavi... 25页 1下载券 喜欢...
...I learned to stop worrying and love the flow教师...
Art.4. How I learned to stop worrying and love the flow教师_英语考试_外语...conversation, instead of just an audience member or consumer at the ...
...TheoryofAcceptanceandUseofTechnologyModel
ofUnifiedTheoryofAcceptanceandUseofTechnologyModel_...The objective of the present research is to ...online course management, content management and ...
软件工程选择题
UML notations that can be used to model the ...The data flow diagram: Answer: e depicts ...modeling a state is any consumer or producer of...
电子商务专业外文文献
theory is the law of e-commerce industry, the ...online transaction information transfer and trade. ...to achieve the automatic flow of information, and...
【论文正文】二维码支付的用户接受意愿影响因素研究
consumer attitude to use; (3) Age, sex, ...(Technology Acceptance Model, TAM)是由 Davis 教授...(the Unified Theory of Acceptance and Use of ...
更多相关标签: