Individual Differences and Internet Shopping Attitudes and Intentions

C. Brad Crisp
Abilene Christian University, Department of Management Sciences, ACU Box 29353, MBB 254, Abilene, TX 79699-9353 Telephone: 325-674-2077 Fax: 325-674-2507

Sirkka L. Jarvenpaa*
University of Texas at Austin, Center for Business, Technology and Law, CBA 6.480, Austin, TX  78712 Telephone: 512-471-1751 Fax: 512-471-0587

Peter A. Todd
McGill University, Desautels Faculty of Management, Bronfman Building, 1001 Sherbrooke Street West, Montreal, Quebec H3A 1G5 Telephone: 514-398-4001 Fax.: 514-398-3876

Introduction

The Internet's World Wide Web (or Web) provides the promise of a new direct interactive shopping channel that is not bounded by time nor geography. Online store fronts are being constructed to market goods and services over the Internet. These store fronts can be established at low cost, and the product and service offerings can be updated on demand. Consumers around the world can have nearly instantaneous, 24 hour a day access to these store fronts using Web browsers. They can search for information about products and services, place an order for a product, receive follow-up service and in some cases, such as for information based products, actually take delivery over the Web. Individual retail store fronts can be grouped to form clusters of electronic malls.

Web retailing is an innovation that many predict will experience phenomenal growth, with total sales ranging from $2 to $20 billion by Year 2000 (Hoffman and Novak, 1996; Ives, 1997). Given these expectations, many businesses feel compelled to develop a Web presence (Champy et al, 1996; Economist, 1997). Others, however, are less enthusiastic about the commercial potential of the Web, pointing out the many challenges that the new retail form faces in terms of user acceptance (Alba et al, 1996).

Past forecasts on many other home information technology and services targeted at supporting retailing have suffered from over-optimism, technology determinism, and faulty assumptions of potential user communities (Klopfenstein, 1989). Thus, while it is clear that many see significant potential in the Web as a sales and marketing tool, much more needs to be understood about what drives commercial activity on the Web. In particular, “the need to understand Internet consumers is vital for both existing businesses and those who are considering a shop in cyberspace" (Fram and Grady, 1995, p. 63).

Despite the need to understand what drives Internet consumers, much of the empirical research of the Web as a retail channel has been descriptive, focusing on estimating the current Internet and Web user base, understanding what the users are buying, and how much they are spending (e.g., Georgia Institute of Technology, 1995; Gupta and Chatterjee, 1997). Emerging from this is a somewhat homogeneous picture of Web users and their buying behavior. For example, Hoffman, Kalsbeek and Novak  (1996) found that those who actually buy on the Web use the Internet at least once a week and are frequent Web users. Further, retailers are encouraged to target stereotypical Web users: "college educated white males in their early thirties, earning higher than average incomes and employed in the computer, education, and other professional fields" (Quelch and Klein, 1996; p. 61). Many merchants seem to have taken this stereotype to heart and target these high-tech professionals as evidenced, for example, by the fact that the bulk of advertising and retail offerings on the Internet are aimed at a high-tech user (Information Week, August 5, 1996).

However, the attention focused on this high-tech male user tends to be somewhat limiting and ignores other potentially important user segments. First, the Web population doubles every six months suggesting that, at a given point in time, nearly half of the users are inexperienced (Borenstein, 1996). Second, recent reports report a steadily growing female Internet population. It was noted that only 21% of the Web user base was female in 1995 (Fram and Grady 1995), 27% was estimated to be female in 1996 (Hoffman, Kalsbeek and Novak 1996), and that by 1997 41% of the Web population are women (Business Week, May 1997). Clearly women represent an important and growing segment of the Web user base. They are disproportionately more important in terms of potential retail activity. Women do the majority of shopping for most households and are the primary buying decision maker in a household (Dholakia et al, 1995).

In addition to a changing gender mix, the web user base is evolving in other ways that require the attention of marketers. It is attracting more mature individuals with 45% of the Web population being 40 or over. The Web population is more affluent and better educated than the population at large. More than 42% of the households who use the Web have incomes greater than $50,000 (versus 33% of the population at large), and 73% of the Web users have attended college (vs. 46% of the total population) (Business Week, 1997). Thus, in many ways, the Web user base has transformed itself significantly from the stereotypical view of the 18-24 year old male computer nerd that dominated only a few years ago.  The increased diversity of Web users suggests different market segments which may respond to commercial opportunities on the Web in various ways.  Thus, it is important to explore how individual differences between users may influence consumer attitudes, intention and acceptance of Web-based retailing.

To begin this process, this paper reports on a study which explored the effect of individual differences on the Internet shopping beliefs, attitudes, and intentions using a sample of 220 individuals, of whom 184 were female. The average participant was 43 years of age. The results suggest that attitude towards shopping on the Web was significantly influenced by attitude towards computers. Intention to shop on the Net was significantly influenced by prior Web experience, attitude towards computers, frequency of shopping in direct marketing channels, household size, and age, explaining 22% of the variability in shopping intention. Prior Web experience also seemed to affect the salience of some of the beliefs related to Internet shopping beliefs, although the Internet retail patronage model developed in an earlier paper seemed to hold well for both inexperienced and experienced users.

The next section reviews the theory of technology acceptance and the role of individual differences in the theory. We will advance hypotheses that integrate empirical findings from information systems and marketing on individual differences. The third section describes the methodology, and the fourth section reports on the results. The last two sections discuss the results and conclude the paper with recommendations for practice and theory.

Conceptual Background

The current study will explore how individual difference factors influence beliefs,  attitude and intention toward Internet shopping. Individual differences have been theorized and  found to be associated with the acceptance of new information technology (Zmud, 1979; Nelson, 1990) and new forms of retail marketing and shopping (Akaah, Korgaonkar, and Lund, 1995; Gehrt and Carter, 1992). We will place the individual difference factors in the context of attitude behavior models such as the Theory of Reasoned Action (TRA) (Fishbein and Ajzen, 1975; Ajzen and Fishbein 1980) and the Technology Acceptance Model (TAM) (Davis, 1989; Davis et al, 1989). To do this we employ an integrated model of Internet shopping behavior developed in Jarvenpaa and Todd (1997).

The model, shown in Figure 1, suggests that four factors product value, shopping experience, service quality and risk will, together, influence attitude and intention towards shopping on the Internet. These factors were derived from the literature on retail patronage and catalog shopping and then empirically examined. Each of the factors has a set of more detailed antecedent beliefs which together form an integrated Internet patronage model. The detailed antecedent belief factors are defined in Table 1 (see Jarvenpaa and Todd, 1997). The model shows how the beliefs related to value, experience, service and risk combine to influence attitude and intention towards Internet shopping. This set of relationships is consistent with the TRA which suggests that beliefs about the consequences of the behavior are key to the formulation of attitude towards the behavior (Fishbein and Ajzen 1975; Ajzen and Fishbein 1980). 

The model was useful in helping to understand the determinants of attitude and intention towards Internet shopping. Specifically, attitude, product value and service quality all had significant influences on intention to shop on the Internet. Product value, shopping experience and perceived risk all had significant effects on attitude. The model explained approximately 50% of the variance in attitude and intention towards Internet shopping. Further the detailed belief factors accounted for a significant portion of the variance in the four retail patronage factors. Perceived product value was significantly influenced by price, quality and variety. The shopping experience was significantly influenced by effort, playfulness and compatibility but not sociability. Service quality was significantly influenced by responsiveness, tangibility and reliability but not by empathy. Finally, perceived overall risk was significantly influenced by performance, personal, privacy and economic risks but not by social risks. Thus, overall the model appears to provide significant insight into attitude and intention towards shopping on the Internet.

To explore the relationships in the model further, this paper examines how individual difference factors might moderate the relationships found and act as determinants to the basic beliefs. Hence, this paper extends our earlier work on online retail patronage model to consider individual differences. In particular we examine how prior web experience, general attitudes towards computers and shopping, as well as demographic factors such as age, income, gender, education and household size affect beliefs, attitudes and intentions towards shopping on the Internet. The next section will present detailed hypotheses for the relationship of these individual difference factors to beliefs, attitude and intention.

Hypotheses: The Effect of Individual Differences

Attitude-behavior models such as the Theory of Reasoned Action and the Technology Acceptance Model, upon which the Integrated model of Internet Shopping is based, typically assert that the influence of external factors, such as individual differences, will be mediated through the belief, attitude and intention components of the model. Among the factors that have been considered in this regard are prior experience with related behaviors, attitude towards objects related to the action and demographic variables such as age, income, education, and gender. We explore each of these groups of factors in turn.

Prior Web Experience

One of the key sources of information used to shape beliefs is the person's own direct experience with a similar situation. The literature on TAM specifically and in the information systems literature in general suggests that an individual's prior computer usage impacts their beliefs about related systems and technology (Venkatesh and Davis, 1996; DeLone, 1988; Fuerst and Cheney, 1986; Igbaria et al, 1989; Lee, 1986; Igbaria et al, 1995). Taylor and Todd (1996) found the TAM model to predict intention and behavior for both experienced and inexperienced users although there were differences in the relative influence of the belief factors on intention and usage. Taylor and Todd in particular found that the relationship between intention and usage was stronger for the experienced than for inexperienced people, suggesting the users employ the knowledge accumulated from past experiences to form their intentions. Similarly, Thompson et al (1994) found that prior experience with computers not only had a direct effect on beliefs, attitude and intention, but also that the person's level of experience with the particular technology moderated the strength of the relationships between beliefs, attitude, and intention (Thompson et al, 1994). These findings on TAM are consistent with the arguments of the Theory of Reasoned Action (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975; Triandis, 1979). Past experience generates objective consequences from engaging in the behavior. This knowledge will help to reinforce behavior and shape and moderate beliefs, future intention and behavior (Fishbein and Ajzen, 1975).

In the retail literature, consumers' knowledge as well as past experience has also been found to influence their attitude and intention (Smith and Swinyard, 1983; Akaah et al, 1995; Korgaonkar, Lund, and Price, 1985). For example, in the early days of "computing in the home," the purchase of home computers was found to be a function of a person's prior computing experience (Venkatesh and Vitalari, 1987). Moreover, past consumer research has shown that the more positive the person's past experience, the more positive their attitude and more likely they are to engage in the behavior again (Bagozzi and Warshaw, 1990). For example, past favorable experiences with catalog sales have been shown to increase a person's likelihood of shopping via catalog again (Shim, 1992). The adoption of videotext based shopping has similarly been found to be a function of prior experience with mail-order and catalog shopping (Eastlick, 1993). Hence, the first hypothesis is that:

H1: Prior Web experience will positively affect as well as moderate beliefs, attitudes and intentions towards Internet shopping.

Computer Attitudes

Prior studies on videotext based shopping have found the predisposition toward computers and high-tech equipment in general to be a significant determinant of adoption and use of videotext shopping (Korgaonkar and Moschis, 1987; Korgaonkar and Smith, 1986; Shim and Drake, 1990; Shim and Mahoney, 1991). Although most of the past studies have found a consumers' computer orientation to influence attitudes toward videotext in a positive manner, Eastlick (1993) found a more ambiguous relationship between consumers' computer orientation and their attitudes toward videotext shopping. Her findings might suggest that those with favorable attitudes towards computers have higher expectations of electronic based shopping leading them to be both critical and positive at the same time. Research on new technology adoption has consistently found the attitudes towards computers to be a significant determinant of adoption and use although some have found the relationship to weaken as users gain experience with the technology (Thompson et al, 1994). Hence, we hypothesize that:

H2: Attitudes towards computers will affect beliefs, attitude and intention toward Internet shopping, especially for less experienced Web users.

Shopping Orientation and Direct Shopping Frequency

Shopping orientation refers to the degree to which an individual sees themselves as a shopper and takes pleasure or personal satisfaction from the act of shopping. Shopping orientation has been found to be among the most influential predictors of consumer patronage behavior (Darden and Howell, 1987). For some customers, shopping is a pleasurable activity and an important part of the person's life. However, others do not enjoy shopping. In the current study, we associate shopping orientation with shopping enjoyment which is the most important dimension of the shopping orientation (Bellenger and Korgaonkar, 1980; Carlson 1978).

Some past research suggests that those who generally do not have positive shopping orientation will have more positive attitudes towards direct shopping modes, such as catalog shopping. Direct shopping is associated with reduction in time spent on shopping, saving physical effort, flexibility in the timing of shopping, and reduced levels of aggravation (e.g., Jasper and Lan, 1992; Darian, 1987). By contrast, others have found the opposite: those who shop via catalog also have a high propensity to shop in general (Shim, 1992).

The above contradictory results were partly explained by Gehrt and Carter (1992) who suggested that while at times, shoppers seek for efficiency in catalog shopping; at other times shoppers seek recreational or enjoyment value. In a study of French catalog patrons, the recreational orientation actually dominated perhaps because of the newness of the innovation (Gehrt et al, 1992). Given that the Web is a new form of direct marketing, we expect that those with high levels of shopping orientation (i.e., shopping enjoyment) also exhibit higher levels of attitudes and intention to shop on the Internet.

H3: Those with more favorable attitudes towards shopping in general will have more favorable attitude and intention towards Internet shopping.

We also expect those who currently use direct marketing channels to have more positive attitude and intention towards Internet shopping than those who do not utilize direct marketing channels such as catalogs. In many respects, the Web is much like the other direct marketing channels, with most merchants' web sites being nothing more than electronic versions of their catalogs (Alba et al 1996). Merchants rarely take advantage of the Web technologies' interactive and multimedia features and even more rarely customize their interfaces to a particular customer. Moreover, as is the case with catalogs, electronic storefronts make it difficult for consumers to examine non-information based products before purchase. Further, because of its novelty, the electronic shopping channel comprises many merchants with whom the shopper has had no prior experience. Hence, we expect that:

H4: Those who more frequently use other direct shopping modes will have more favorable attitude and intention toward Internet shopping.

Demographics: Income, age, education, gender, and household size

Past literature suggests that the new modes of direct retailing are patronized more by high income, educated, younger shoppers than by low income, less educated, and older shoppers (Cunningham and Cunningham, 1973; Reynolds, 1974; Feldman and Star, 1968; Berkowitz et al, 1979; Gillette, 1976; Peterson et al, 1989; Fields and Greco, 1988; Shim and Drake, 1990; Korgaonkar and Smith, 1986; Eastlick, 1993). One explanation for these effects is that this particular profile of shopper places a higher value on convenience, is less price conscious, and is more tolerant of the risk associated with new retail innovations (Reynolds, 1974; Peterson et al, 1989). These consumers also are found to seek a more unique merchandise assortment that the new direct retailing forms often provide (Reynolds, 1974; Gehrt and Carter, 1992). These results apply to catalog, mail-order, and videotext shopping. With time, demographic effects, except for income, reduce as the novelty of the new mode wears off (Kilsheimer and Goldsmith, 1991; Shim, 1992; Akaah et al, 1995).

The effect of age has been found to be transitional in retail patronage studies. For example, as catalog shopping has become more commonplace, the older shoppers have been found to be as likely and sometimes even more likely to shop using catalogs than their younger counterparts (Lumpkin, Greenberg, and Goldstucker, 1985; Jasper and Lan, 1992). However, with new retail innovations, age has been found to be significant. Similarly, the technology innovation literature suggests an inverse relationship between age and acceptance of new technologies (e.g., Harrison and Rainer, 1992; Nickell and Pinto, 1986; Gattiker, 1992). Older individuals are more conditioned to their customary way of behaving (i.e., have routinized more of their behavior). This might suggest that older individuals have to exert more cognitive and emotional effort to learn new behaviors and disassociate themselves from their daily routines. For example, Zeithaml and Gilly (1987) found that those elderly people who did not adapt to using bank ATM's emphasized their old way of doing and chatting with tellers. Also, those with less education have been found to exhibit more computer anxiety and more negative attitudes towards computers (Igbaria and Parasuraman, 1989).

Relatively little is known about the effect of gender on direct marketing patronage behavior because most studies are dominated by female shoppers reflecting the fact that women are still the primary shoppers as well as the primary purchasing decision makers (Rodkin, 1991). Some studies that have involved both genders have found no gender difference on patronage behavior (Kilsheimer and Goldsmith, 1991), whereas others have found females to be more frequent users of catalogs than males, particularly if the individual is over 40 years of age (Lumpkin and Hawes, 1985; Darian, 1987). Males and females have also been found to differ in terms of their beliefs and motives affecting purchase behavior. Females are more motivated by convenience, time savings, and effort, whereas males are motivated by the merchandise assortment and service (Eastlick and Feinberg, 1994). For example, a study by Dholakia et al (1995) found that upscale males acknowledged the convenience of the catalog and computer-based shopping modes but preferred in-store shopping because they perceived the alternative forms as too expensive and as less fun and rewarding. The literature on new information technology innovations shows little evidence of significant differences between the sexes in the use of technology and attitudes towards technology-mediated activities (Gattiker, 1994) although some studies have found women to possess less positive attitudes toward technology than men (Rafaeli, 1986). Also, early work on women and information technology found women to be less optimistic about the benefits of computer technology than men (Gutek and Bikson, 1985). Hence, we hypothesize that:

H5: Higher income, younger, more educated, and single male shoppers will have more positive attitude and intention toward Internet shopping.

Methodology

Two hundred twenty household shoppers participated in the experiential survey study where they spent 1.5 hours in a university computer laboratory engaged in a variety of World Wide Web shopping activities and completed several questionnaires on individual differences factors, beliefs, attitudes and intentions toward shopping on the Internet. Each participant was paid $20. The computer lab was equipped with Pentium workstations, 17 inch monitors, and high speed Ethernet access to the university backbone network. Since even among World Wide Web users electronic shopping activity is relatively low, this approach ensured that respondents had a minimum level of  exposure to shopping on the World Wide Web, so that they could provide informed input.

Sampling Strategy.

The sample was targeted to match the demographics of the Web population along dimensions such as income, age, and education level; while targeting primary household shoppers, who are predominantly female (Blaylock and Smallwood, 1987; Dholakia et al, 1995). This was thought to be important since it has been argued that one of the key challenges of Web-based shopping is the male dominated user base (Burstein and Kline 1995).

Participants were solicited through two sources. A mail solicitation asking the primary shopper to participate in the study was sent to a random sample of 1500 local area households which met the demographic characteristics of World Wide Web users reported by O'Reilly and Associates (1995) (i.e., median household income between $50,000 and $75,000, professional occupations, and median age between 35 and 55). To ensure a reasonable proportion of World Wide Web users in the sample, 500 additional solicitations were sent to e-mail users among the non-academic staff at a Southwestern university.  The mailing generated over 300 responses from interested participants (a 15% response rate). Due to budget constraints we invited and scheduled 250 participants.  Over the course of the sessions 30 people canceled yielding a final sample of 220, which includes 70 university employees.

The demographic profile of the participants is summarized in Table 2. Aside from gender, which was deliberately skewed, the sample closely matches the demographic profile of Internet users reported by O'Reilly and Associates (1995).  The sample's World Wide Web experience is summarized in Table 3. Only 18 participants (8%) reported previously shopping through the Internet.  This is consistent with estimates of the proportion of Internet users who had made purchases at the time of the study (Advertising Age, 1995).

Experiential Tasks.

Web pages were created that outlined the shopping tasks and provided links to various shopping sites. Pointers to the shopping sites and instructions for study participants were available at Jarvenpaa's Website at the University of Texas. These included a link to the SHOPPER, a web page that indexes over 350 shopping sites, plus pointers to 5 electronic malls rated highly by various services such as Point Communications Corporation. The pages also contained pointers to catalogue retailers (JCPenney, LLBean and Land's End) and participants could access NetSearch facilities to search for particular products, services or shops. The intent was to create a selection of sites that were of good quality.

The participants performed three shopping activities.  The first was a familiarization task, where participants browsed among different shopping sites.  The second was a gift shopping exercise, to try to find a gift for a friend or relative.  The third task was a personal shopping exercise, where participants tried to find something to buy for themselves.  During each of these activities participants were told that they were not required to make a purchase and were personally responsible for any purchases they did make.  The tasks were meant to be typical of those that the consumer might routinely perform.  Their primary purpose was to provide the participants with a focus for their shopping activities. Table 4 summarizes the products and services that the consumers in our study most frequently considered buying as they completed the required shopping tasks. 

Procedures

Participants first completed a demographic questionnaire and a consent form. This was followed by a 5 minute overview of the study and introduction to World Wide Web and Netscape. Next, participants engaged in the familiarization task for 10 minutes, followed by the 20 minute gift shopping exercise. Next, they spent 10 minutes filling out an open ended survey about the shopping experience to that point. Then the personal buying activity lasted another 20 minutes, after which the participants completed a structured questionnaire. This was followed by an optional 30 minute focus group session which approximately half the participants attended. For all sessions, lab assistants were present to answer questions, help the participants find specific sites, and assist in navigation over the Web. The data were collected in groups of 10 to 24 over 5 days in January 1996.

Measures

The items used to measure prior Web experience, attitude towards computers, shopping enjoyment, direct shopping experience were constructed on the basis of past literature. These items are reported in Appendix A along with the items for the demographic factors. Scales to measure each of the beliefs were developed based on previous literature and existing scales where possible (see Jarvenpaa and Todd, 1997b). Measures related to service quality were based on the SERVQUAL scales of Parasuraman, Zeithaml and Berry  (1988; 1994a; 1994b), scales related to the shopping experience were based on prior work by Davis (1989) for effort, Webster et al (1991) for playfulness and Moore and Benbasat (1991) for compatibility. Measures of attitude and intention were based on the suggestions of Fishbein and Ajzen (1975).

Analysis Strategy

The data were analyzed using correlation and regression techniques. The structural equation modeling technique was not used because of the sample size and the relationships being investigated were exploratory as they have not been previously investigated in the Web shopping context.

Results

Results are presented in terms of correlation and regression analyses which look at the effects of the individual difference factors on beliefs, attitudes and intentions toward shopping on the Internet. Factor analysis was used to confirm the individual difference constructs of prior Web experience, computer attitude, shopping enjoyment, and direct shopping frequency (Appendix B). The constructs for beliefs, attitude, and intention were developed in an earlier paper (see Jarvenpaa and Todd, 1997). The criteria used to identify and interpret factors were that a given item should load .50 or higher on a specific factor and have a loading no higher than .35 on other factors. In general, items mapped closely with their intended factors. Table 5 provides a summary of descriptive statistics, intercorrelations, and reliability coefficients for the difference constructs, demographics, attitude, and intention. The reliability coefficients ranged in size from .61 for shopping enjoyment to .89 for prior Web experience. While some of the coefficients could be higher, they appear quite adequate given the exploratory nature of the study (Nunnally, 1967). Furthermore, the sizes of the coefficients are comparable with others reported in the literature (e.g., Korgaonkar, Lund and Price, 1985).

Intercorrelations Among Study Variables

Table 5 presents the intercorrelations among the individual difference constructs, demographic factors, attitude, and intention. As noted, attitude towards internet shopping is correlated with computer attitude (r=0.26, p< 0.01) and prior Web experience (r=0.21, p <0.01). Intention towards internet shopping is correlated with attitude (r=0.65, p <0.01), prior Web experience (r=0.32, p <0.01), computer attitude (r=0.27, p <0.01), and gender (r=0.16, p <0.05).

Table 6 presents the intercorrelations between the individual difference factors and the consumer's beliefs. Prior Web experience is positively correlated with compatibility (r=0.15, p <0.05) and social risk (r=0.16, p <0.05). Shopping enjoyment is negatively correlated with compatibility (r=-0.25, p <0.01). Direct shopping frequency is negatively correlated with price (r=-0.29, p <0.01), privacy (r=-0.18, p < 0.05), and performance risk (r=-0.15, p <0.05). It is positively correlated with product value, quality, service, social, effort, and compatibility, ranked from most significant to least (range of r=0.32 to 0.21, p <0.01), and with shopping experience (r=0.16, p <0.05) and performance (r=0.15, p < 0.05). Computer attitudes is positively correlated with shopping experience (r=0.31, p <0.01), compatibility (r=0.22, p <0.01), empathy (r=0.18, p <0.05), playfulness (r=0.17, p <0.05),  and product variety (r=0.16, p <0.05). Gender is correlated negatively with personal risk (r=-0.21, p <0.01) and positively with social experience (r=0.19, p <0.01).  Work tenure is positively correlated with service (r=0.19, p <0.01), and education is negatively correlated with effort (r=-0.15, p <0.05) and tangibility (r=-0.15, p< 0.05).

Results on Hypothesis Tests

Table 7 presents the regression results. The individual difference factors including the demographic variables explained 22% of the variance in intention, but only 6% of the variance in attitude. Examination of the model results for intention indicates that five of the individual difference factors yielded statistically significant standardized coefficient estimates – prior Web experience, age, household size, computer attitude,  and direct shopping frequency. Thus, the model results support the influence of these variables as predictor's of the consumer's intention to shop on the Internet. Examination of the model results for attitude indicates that only one of the individual difference factors yielded a statistically significant coefficient estimate – computer attitude (b=0.26, t=3.66). 

Hypothesis 1 predicted that prior Web experience will positively affect as well as moderate beliefs, attitude, and intention toward Internet shopping. The regression results noted above suggest that prior Web experience had a significant impact on intention to shop, but it did not affect attitude towards Internet shopping. In addition, prior Web experience was found to moderate shopping beliefs, attitude, and intention in several ways. Table 8 shows the beta coefficients for beliefs, attitude, and intention for two groups of consumers categorized on the basis of prior Web experience. Those with ‘no experience' had never before used the Web prior to the shopping study.

While the beta values for beliefs change with experience, the prior shopping model is still useful in explaining Internet shopping beliefs, attitude and intention (adjusted R² ranging from .47 to .75). As prior web experience increases: attitude becomes more significant in explaining intention to shop; shopping experience has a greater impact on attitude; personal risk and privacy have a larger influence on risk; and playfulness is a greater determinant of shopping experience. At the same time, effort and price have a larger impact on shopping experience and value, respectively, for less experienced shoppers. Attitude becoming more significant with experience is consistent with the literature and other technologies. Once experience is gained, consumers want merchants to offer a fun experience with low risk. 

Hypothesis 2 predicted that a consumer's attitude towards computers will affect beliefs, attitude and intention toward Internet shopping. Consistent with this hypothesis, the regression results in Table 7 suggest that computer attitude has a significant impact on attitude and intention toward Internet shopping. Also, while not presented, computer attitude was significant in regression analysis in explaining several beliefs, including shopping experience.

Hypothesis 3 predicted that those with more favorable attitudes toward shopping in general will have a more favorable attitude and intention towards Internet shopping. This hypothesis was not supported by the regression analysis. Shopping orientation (i.e., enjoyment) had neither positive nor negative effect on attitude and intention toward Internet shopping.

Hypothesis 4 predicted that those who more frequently use other direct shopping modes will have a more favorable attitude and intention towards Internet shopping. As indicated in Table 7, direct shopping frequency did have a significant positive impact on intention towards Internet shopping.

Hypothesis 5 predicted that higher income, younger, more educated, and single male shoppers will have more positive attitudes and intentions towards Internet shopping. The regression results in Table 7 show that this hypothesis is not supported. Age did have an impact on intention, but the coefficient is positive not negative (i.e., intention increases with age). Also, household size did have a significant impact on intention; however, it increases with household size. 

Discussion and Implications

The study investigated the effect of individual difference factors on beliefs, attitude and intention to shop on the Web. Overall, the individual difference variables explained only 22% of the variance in intention and less than 10% in attitude. The shopping beliefs explained about 50% of the variance in attitude and 52% in intentions (Jarvenpaa and Todd, 1997). Therefore, the beliefs should be directly related to the design characteristics of the online store fronts. The current results suggest that improving the store fronts, and thereby consumers' beliefs about Web shopping, should be a greater concern for retailers than simply waiting for Internet shopping (and therefore customer attitude and intention) to mature.

However, the maturing of this shopping mode and its users will have a significant impact. Our study also suggests that beliefs that impact attitude and intention to shop are likely to change somewhat from the time of initial adoption to frequent utilization of the new media. One can hypothesize that as access and exposure to the Web increases, people who have positive encounters with computer technology will develop more positive attitudes towards shopping on the Web.

The current study also suggests that prior Web experience makes a difference in favorable intention and favorable attitude towards computers make a difference in both attitude and intention toward Web shopping. Since among the general public, a vast majority has no knowledge of the Web, it seems important that the public be educated about Web shopping and understand that it is not a trivial task. Overall, their attitude and behavior will depend on their past experiences and their attitudes towards computers. Internet shopping is likely to experience negative reactions from consumers if it is aggressively hyped. Further,  marketing that does not take into account factors relating to the time and effort required to navigate on the Web may not be well received. An informal analysis of the open-ended comments on the most liked and least liked site indicated that the study's consumers found it effortful not only to browse a site once it was located but to locate appropriate sites in the first place (Jarvenpaa and Todd, 1996-1997).

The significant effects of individual differences also suggest that the potential market of Internet shopping might need to be segmented. The study found attitude and intention differences depending on the prior Web experience, computer attitudes, age, household size, and frequency of direct shopping. Merchants need to recognize these differences and identify methods to tailor their designs on the Web to these different individuals.

The study also discovered some unexpected effects. Affluent women, many with advanced degrees, were as likely to develop favorable attitude and somewhat even greater intention to shopping on the Web as their male counterparts. Also, shoppers from larger households were more likely to develop a favorable intention to shop on the Web. This last finding is similar to that of Lumpkin and Hawes (1985) who found the number of children living at home to be positively related to the catalog usage.  This likely reflects the time pressures facing shoppers in these households and their view that the Web may better help them to manage their time. 

This research should be extended to more clearly relate the views of different potential Internet market segments.  Using the attitude-behavior model as a framework, it would be useful to examine more carefully how the importance of different factors in the models change as a function of various individual differences.  Calibration of such models would be useful for developing an integration of the attitude behavior models and the individual difference literature.  More applied research might also begin to look at how different segments of the user population might respond to different site design strategies.  For example, those with a strong positive orientation towards shopping in general may be most attracted to Web based sites that embed the shopping experience in a rich community of interest, while those with a negative shopping orientation may be looking for sites that emphasize ease, efficiency and value.  Studies which look at how different user segments respond to alternative site design strategies will be important to ensuring the long run development of the Internet based retailing environment.

Conclusion

The Internet marketing has many problems requiring corrective action. For corrective action to succeed, it must be based on knowledge of potential customers' attitudes and intentions toward the new mode and the antecedents that underlie those attitudes. The objective of the current study was to examine the influence of demographic factors, prior Internet experience, computer attitudes, shopping orientation, and frequency of shopping in direct marketing channels as determinants of consumers' attitude toward Internet marketing and intention to patronize Internet online stores. The study involved an experiential survey; all participants were brought to a university computer laboratory where they browsed through electronic malls for 1.5 hours on the worldwide web and completed survey instruments comprised of items obtained from the literature. The study results indicate that attitudes towards computers affect attitudes towards Internet shopping. Prior Internet experience, attitudes towards computers, age, household size, and frequency of shopping via direct marketing channels affect intentions towards Internet shopping. Depending on their prior experience with the Internet, the consumers placed somewhat different emphasis on various beliefs about Internet shopping. Overall, however, the retail patronage model that we had developed in an earlier paper seems to hold for both inexperienced Web users as well as those with Web experience.

In conclusion, it is important to acknowledge the exploratory character of the results due to the immature and changing technology as well as the study limitations. The study's procedural and task limitations are discussed elsewhere (Jarvenpaa and Todd, 1997). Among the limitations that impact the relative effects of the individual differences variables are the following. Perhaps those users with experience with the Web but with negative attitudes towards Web shopping declined the invitation to participate in the study. The study had a limited number of males. The general attitude towards shopping was high and hence those that disliked shopping may have declined the invitation. Finally, our consumers were either from affluent township or from the university community and so the results may not reflect the full diversity of beliefs and attitude toward shopping.

Acknowledgements

This work has been supported by grants from the Graduate School of Business, University of Texas at Austin and the Natural Sciences and Engineering Research Council of Canada. We are grateful to Dave Hannah, Tom Shaw, Connie Todd, and Joe Wiseman for their assistance with the collection and coding of the data reported in this study.

Notes

* Contact person for correspondence

Originally published June 1997 - contact information for authors updated January 2007

This paper was originally available at the home page of C. Brad Crisp at the University of Texas, Austin, who has made the paper available to Information Research, with the agreement of his co-authors.

Appendix A

Appendix B: Factor Loadings