Understanding Consumer Intention to Shop Online:
A Comparison of Three Intention-Based Models
Jae-Jin Park*
Doctoral Candidate
Missouri School of Journalism
Address: 1700 Forum Blvd. #305
Columbia, MO 65203
Telephone: 573) 446-6033
E-mail: [log in to unmask]
Glen T. Cameron
Professor
Missouri School of Journalism
Address: School of Journalism
University of Missouri-Columbia
Columbia, MO 65211
Telephone: 573) 884-2607
E-mail: [log in to unmask]
*Correspondence concerning this manuscript should be addressed to Jae-Jin Park.
A manuscript submitted to the Communication Theory and Methodology Division
of AEJMC for possible presentation at the convention to be held in Kansas
City, MO, July 30 – August 2, 2003.
Understanding Consumer Intention to Shop Online:
A Comparison of Three Intention-Based Models
Abstract
By using a sample of 733 consumers, this study employed path analysis to
compare three intention-based models (i.e., the theory of reasoned action,
the theory of planned behavior, and the technology acceptance model) in
terms of the extent to which each can be used to predict and understand the
consumer's online-shopping intention. This study found that the theory of
planned behavior provides a more robust theoretical basis for the study of
online shopping than does the theory of reasoned action. The significant
roles of usefulness and ease of use in this study confirm that the
technology acceptance model can be successfully applied to the domain of
online shopping beyond information systems usage.
Understanding Consumer Intention to Shop Online:
A Comparison of Three Intention-Based Models
Introduction
Although the Internet has a short history, it has dramatically changed the
way firms do business and the way consumers purchase products or services.
Besides working as an interactive communication channel, the Internet
serves as a market with the development of its technologies and its
interactivity (Hoffman and Novak, 1977). Because the Internet provides
users with rapid access to a variety of information and the ability to
easily compare competitors and prices, the market represented by the
Internet has "the potential to be a more efficient market than conventional
markets." (Hoffman and Novak, 1997, p.50)
These distinctive features of the Internet are leading to the increased
popularity of online shopping. Online shoppers are expected to increase
from 64.1 million in 2000 to more than 100 million by 2003 (Jupiter direct,
2002). The U.S. Department of Commerce estimated that online retail sales
reached $10.2 billion in the second quarter of 2002, up 24.2 percent from
the second quarter of 2001 (U.S. Census Bureau, 2002). Although reports
about the number of online shoppers and e-commerce revenues yield a variety
numbers, it is agreed that such numbers have dramatically increased every year.
The rapid growth of online shopping has led researchers to devote their
efforts to expanding our understanding of the ways consumers shop online.
Researchers have explored a variety of factors that might affect consumers'
decisions to shop online. Researchers have examined the roles of such
influences as Website design characteristics (Chau, Au, and Tam, 2000),
demographic characteristics (Donthu and Garcia, 1999; Fram and Grady,
1995), product or service types (Peterson, Balasubramanian, and
Bronnenberg, 1997; Phau and Poon, 2000), and emotional/motivational factors
(Goldsmith, 2000; Koufaris, 2002; Limayem, Khalifa, and Frini, 2000;
Swaminathan, Lepkowska-White, and Rao, 1999). Although those studies
provide critical insights into understanding the consumer's online-shopping
behavior, most research in this area has been exploratory. Little research
has contributed to theoretical developments in this area. The lack of a
strong theoretical framework to investigate the consumer's online-shopping
behavior may negatively influence the validity of research in this area.
Thus, it is important to revisit existing consumer-behavior theories to
test their applicability in the context of online shopping. Because online
shopping allows the consumer to use only two senses (i.e., sight and
sound), whereas offline shopping permits use of all five senses (i.e.,
sight, sound, feel, smell, and taste), and because online shopping requires
the consumer to extensively use a technology (i.e., specifically the
Internet in this study), the consumer's online shopping behavior needs to
be understood from different perspectives than does conventional shopping
behavior.
Based on the assumption that an individual's behaviors are under his/her
volitional control, the theory of reasoned action (TRA) was developed to
predict and understand an individual's psychological processes that
influence a certain behavior (Fishbein and Ajzen, 1975; Ajzen and Fishbein,
1980). The TRA posits that an individual's psychological responses
intervene between social pressures and personal factors (Abdul-Gader and
Kozar, 1995). As an extended model of the TRA, the theory of planned
behavior (TPB) was developed to explain an individual's behavior under
non-volitional as well as volitional control. The main difference between
the two theories is that the TPB accounts for non-volitional factors, such
as opportunities and resources required for performing a behavior, as
determinants of behavioral intentions or behaviors. Since the TRA and the
TPB have been successfully adopted to predict and understand a variety of
behaviors, they may also provide theoretical frameworks with studies that
investigate the consumer's online-shopping behavior. Based on these
theories, we can assume that before the consumer purchases products or
services through the Internet, s/he would have intention to perform the
behavior. This intention is based on how s/he personally feels about online
shopping, what important-others think about shopping online, and whether or
not s/he has opportunities or resources to shop online.
The technology acceptance model (TAM), which has its theoretical
background in the TRA, was developed to predict and understand an
individual's information systems acceptance in the workplace (Davis, 1989).
Compared to the TRA and the TPB, the distinctive feature of the TAM is that
it is specific and simple, identifying a salient belief set (i.e.,
perceived usefulness and perceived ease of use) as antecedents of
behavioral intention. Because online shopping also requires consumers to
extensively use computer technologies, it is necessary to involve the two
constructs of the TAM (i.e., usefulness and ease of use) in a study that
examines the consumer's intention to shop online.
Although numerous studies have successfully employed the three
intention-based models (i.e., the TRA, the TPB, and the TAM) in a variety
of situations, little research has been conducted to compare the
explanatory powers of the three models in predicting and understanding
consumers' online-shopping intentions. The lack of research in this area
led the authors to compare the three intention-based models in terms of the
extent to which each can be used to predict and understand the consumer's
online-shopping behavior. Results of this study will indicate whether all
three of the given models are adequate to explain the consumer's online
shopping behavior, will identify the relative strengths of each model, and
suggest which model would be more useful than the others. The authors hope
that answering these questions contributes to the development of the three
models, providing future research in the area of online consumer behavior
with a more robust theoretical foundation.
Literature review
The theory of reasoned action
Based on the studies of social psychology, Fishbein and Ajzen (1975) and
Ajzen and Fishbein (1980) developed the theory of reasoned action (TRA) to
predict and understand an individual's behavior. They assumed that human
beings are rational or motivation-based, so humans systematically use the
information available to them and consider the implications of their
behaviors "before they decide to engage or not engage in a given behavior."
(Ajzen and Fishbein, 1980, p.5)
According to the TRA (see Figure 1), behavioral intention (BI) is the
immediate antecedent of an individual's performance of a behavior. The TRA
posits that "most behaviors of social relevance are under volitional
control and are thus predictable from intention." (Ajzen and Fishbein,
1980, p.41)
[Insert Figure 1 here]
The TRA specifies that behavioral intention is a function of two
determinants: a personal factor termed "attitude toward the behavior" (A)
and a person's perception of social pressures termed "subjective norm"
(SN). The former refers to "[an] individual's positive or negative
evaluation of performing the behavior," (Ajzen and Fishbein, 1980,
p.6) whereas the latter refers to an individual's perception of
important-others' desire for him/her to perform or not perform a behavior
(Ajzen and Fishbein, 1980; Ajzen and
Madden, 1986). That is, the TRA assumes that an individual's behavioral
intention is a weighted function of his/her positive evaluations of
behaviors in question and his/her perception that significant referents
think s/he should perform them (Ajzen and Fishbein, 1980). That is,
Behavioral intention = w1A + w2SN
The TRA reaches its boundary at the determinants of attitude and
subjective norm, called behavioral and normative beliefs respectively, to
predict a behavior. In other words, attitude is a function of a set of
beliefs (Ajzen and Fishbein, 1980). A behavioral belief refers to an
individual's subjective probability that a behavior will lead to a certain
consequence (e.g., it saves time to use the Internet for shopping) (Davis,
Bagozzi, and Warshaw, 1989). According to the TRA, the strength of each
behavioral belief (bi) is multiplied by the evaluation of its consequence
(ei), and attitude is determined by summing the resulting products across
all salient behavioral beliefs. The evaluation is defined as "an implicit
evaluative response to the consequence." (Davis et. al., 1989, p.984) Thus,
attitude can be illustrated as
Attitude =
Subjective norm is also a function of a set of beliefs termed normative
beliefs. Normative beliefs "are concerned with the likelihood that
important referent individuals or groups would approve or disapprove of
performing the behavior." (Ajzen and Madden, 1986, p.455) According to the
TRA, to obtain an estimate of a subjective norm, each normative belief
(nbi) of an individual is first multiplied by his/her motivation to comply
with the referent (mci). Then, the cross-products are summed for all
salient referents. That is,
Subjective norm =
Only a few studies adopted the TRA to predict the consumer's
online-shopping behavior. Through a nationwide mail survey with a sample of
355 subjects, Yoh (1999) found that both attitude and subjective norm
significantly contribute to the prediction of Internet apparel shopping
intention, and the determinants of the factors (i.e., attitude and
subjective norm) are their corresponding beliefs. Shim and Drake (1990)
also revealed similar results in the context of electronic apparel shopping.
Although the psychological processes of the TRA have been demonstrated to
be applicable in understanding a variety of behaviors, it has been argued
that the predictive power of the TRA may be weak because it was designed to
predict only behaviors under volitional control. Although some behaviors
can be explained based on the logic of the TRA, under some circumstances,
an individual's behavior can be determined by non-volitional factors such
as opportunities and resources (e.g., time, money, and skills). For
example, even if a consumer has a favorable attitude and subjective norm
toward online shopping, s/he cannot afford it if s/he cannot access or
operate the Internet. In 1985, Ajzen proposed an extension of the TRA to
account for non-volitional factors that can be determinants of behavioral
intention or behavior as explained in the next section.
The theory of planned behavior
The theory of planned behavior (TPB) is an extension of the TRA. As Figure
1 presents, the only difference between the TRA and the TPB is that the TPB
takes account for non-volitional control, named "perceived behavioral
control," over the behavior. Perceived behavioral control is defined as the
individual's perception "as to how easy or difficult performance of the
behavior is likely to be." (Ajzen and Madden, 1986, p.457) As Figure 1
shows, perceived behavioral control influences behavior directly as well as
indirectly. The direct approach is based on the assumption that, holding
intention constant, the individual's effort or confidence to successfully
perform a behavior in question strongly influences performance of the
behavior (Ajzen, 1991). For example, a person who, although inexperienced,
is more confident in his/her ability to use the Internet is more likely to
succeed in shopping online than an unconfident inexperienced person (Ajzen,
1991; George, 2002). However, a direct relation between perceived
behavioral control and actual behavior becomes strong only under two
conditions: (1) when the behavior in question is completely non-volitional
and (2) when perceived behavior control accurately reflects actual control
(Ajzen and Madden, 1986).
With respect to indirect effects of perceived behavioral control on
behavior, the TPB posits that behavioral intention is an immediate
determinant of behavior and that behavioral intention is determined by
attitude, subjective norm, and perceived behavioral control with relative
weights (w). That is,
Behavioral intention = w1A + w2SN + w3PBC
Just as attitude and subjective norm are affected by behavioral and
normative beliefs respectively, perceived behavioral control is also a
function of beliefs, termed control beliefs. Control beliefs refer to the
individual's perception of the extent to which s/he possesses internal and
external factors that may increase or decrease the perceived difficulty of
performing the behavior. Internal factors include such variables as
individual differences, information, skills, abilities, power of will,
emotions, and compulsions, whereas external factors involve time,
opportunity, and dependence of others (Ajzen, 1985). To obtain an estimate
of perceived behavioral control, each control belief (cbi) is multiplied by
the perceived power of the control factor to facilitate or inhibit
performance of the behavior (ppi), and the resulting products are summed
across all salient control beliefs. That is,
Perceived behavioral control =
Recently, a few studies adopted the TPB to explain the consumer's
online-shopping behavior. George (2002) investigated predictors that
influence online shopping by using the secondary data of the tenth World
Wide Web user survey conducted in October 1998 by Graphics, Visualization
and Usability. Although he said that his study was based on the TPB, it
seemed to be based on the TRA rather than the TPB because he did not
investigate the effects of subjective norm and perceived behavioral control
on intention to shop online. The results revealed that actual online
shopping was a function of intention, and intention, in turn, was a
function of attitude. Limayem, Khalifa, and Frini (2000) demonstrated that
intention to shop online is a function of attitude, subjective norm and
perceived behavioral control. They also found that intention and perceived
behavioral control equally contribute to the prediction of actual online
shopping. These results imply that online shopping is low in volitional
control and requires adequate resources and opportunities to successfully
perform. Thus, we expect that the predictive power on intention of the TPB
would be greater than that of the TRA in the context of online shopping.
Previous studies also demonstrated that the TPB has a greater power in
predicting and understanding behaviors than does the TRA. They include
predicting unethical behavior (Chang, 1998) and students' getting an "A" in
a class (Ajzen and Madden, 1986). Note that because the TPB encompasses all
constructs of the TRA as an extension model of the TRA, this study presents
hypotheses based on the TPB.
Hypothesis 1: The TPB has better explanatory power than the TRA in
predicting the consumer's online-shopping intention.
Hypothesis 2a-2c: In the online-shopping context, there are positive
relationships between (a) behavioral belief and attitude, (b) normative
belief and subjective norm, and (c) control belief and perceived behavioral
control.
Hypothesis 3a-3c: In the online-shopping context, the consumer's intention
to shop online is a function of the following determinants: (a) attitude,
(b) subjective norm, and (c) perceived behavioral control.
The technology acceptance model
Based on the TRA, Davis (1989) developed the technology acceptance model
(TAM) (see Figure 2) to explain the individual's information systems (IS)
acceptance behavior. The most distinctive feature of the TAM is that it is
specific and simple. That is, the TAM uses a salient belief set (i.e.,
perceived usefulness (U) and perceived ease of use (EOU)) that is
consistently applicable across various situations. Perceived usefulness is
defined as "the degree to which a person believes that using a particular
system would enhance his or her job performance." (Davis, 1989, p.320) On
the other hand, perceived ease of use is defined as "the degree to which a
person believes that using a particular system would be free of effort."
(Davis, 1989, p.320)
[Insert Figure 2 here]
Consistent with the TRA, the TAM posits that IS usage is a function of
behavioral intention. However, in the TAM, behavioral intention to use IS
is determined by attitude toward using the system, which reflects an
individual's positive or negative evaluation of using the system, and a
belief, perceived usefulness, with relative weights (w). That is,
Behavioral intention = w1A+ w2U
As the TRA posits that attitude is a function of a belief set, the TAM
identifies that attitude toward using a system is determined by two
specific beliefs (i.e., perceived usefulness and ease of use) with relative
weights. With respect to the impact of usefulness on attitude, Davis et.
al. (1989) insisted that although the individual's attitude toward a
behavior may differ from attitude toward any possible reward as an outcome
of the behavior, positively perceived outcomes often improve the
individual's attitude toward achieving those outcomes. Davis et. al. (1989)
also insisted that ease of use directly influences attitude toward IS
usage. This is because individuals who believe that a system is easy to
interact with are more likely to increase their sense of self-efficacy.
This, in turn, influences attitude, effort persistence, and motivation to
use the system. Thus, attitude can be illustrated as
Attitude = w3U + w4EOU
Increasing ease of use may also increase performance. Increasing ease of
use maximizes efficiency of effort, allowing an individual to accomplish
more work for his/her effort. A change in weighted ease of use will also
proportionally affect usefulness. Thus, usefulness can be illustrated as
Usefulness = w5EOU
Although the TAM originates from the theoretical background of the TRA,
the two models have several different theoretical assumptions in explaining
the relationship between behavioral intention and other predetermined
factors affecting behavioral intention. First, contrary to the TRA, the TAM
does not include subjective norm as a determinant of behavioral intention.
Davis et. al. (1989) noted that although subjective norm is an important
factor in predicting individual behavior, the TRA has theoretical problems
in conceptualizing subjective norm. Specifically, they asserted that
It is difficult to distinguish direct effects of subjective norm on
behavioral intention from indirect effects via attitude; in addition,
attitude may influence subjective norm, for example, due to the false
consensus effect in which people project their attitudes to others. (p.986)
Second, a salient belief set should be elicited to explain each behavior in
the TRA. The TRA is not based on a specific behavior; therefore, each
behavior requires its distinctive and specific belief set (Ajzen and
Fishbein, 1980). On the other hand, the TAM uses a specific belief set
(i.e., usefulness and ease of use) that may be consistently applicable
across a variety of situations.
Third, the TAM considers its two beliefs (i.e., usefulness and ease of use)
as "distinct but related constructs." (Davis et. al., 1989, p.987) In
contrast, all behavioral beliefs of the TRA are multiplied by their
corresponding evaluations, and then the cross-products are summed.
Finally, the direct relationship between usefulness and behavioral
intention in the TAM violates the basic assumption of the TRA that attitude
mediates the relationship between beliefs and behavioral intention (Taylor
and Todd, 1995). According to Davis et. al. (1989), this violation
originates from the concept that within organizations, people evaluate
behaviors in terms of how they will increase their job performance instead
of on the emotional reactions the behavior may receive. The researchers
suggest that this is due to the importance of performance in achieving
rewards for the work.
The TAM has been predominantly applied to predict and understand user
acceptance of IS (e.g. Chau and Hu, 2001; Davis, 1989; Malhotra and
Galletta, 1999; Taylor and Todd, 1995). However, the results of TAM studies
are inconsistent in predicting causal links between constructs in the TAM,
especially between attitude and intention. Thus, some current TAM studies
(e.g., Gefen and Straub, 2000; Lederer et. al., 2000; Teo, Lim, and Lai,
1999; Venkatesh and Davis, 1996; Venkatesh and Morris, 2000) have measured
direct effects of usefulness and ease of use on technology usage intention
or actual technology usage by excluding attitude. This method of research
has been based on the assumption that attitude does not function as a
strong mediating variable between the two beliefs (i.e., usefulness and
ease of use) and behavioral intention as originally expected (Chau and Hu,
2001). In this study, however, attitude has been retained to test the TAM
because one of objectives of this study is to fairly compare validations of
the TRA, the TPB, and the TAM in predicting the individual's
online-shopping behavior. The three theories articulate that the effects of
beliefs on behavioral intention are mediated by attitude. Rich literature
(e.g., Davis et. al., 1989; Limayem et. al., 2000; Mathieson, 1991; Taylor
and Todd, 1995) about the TRA, the TPB, and the TAM has also demonstrated
that attitude takes an important role in the individual's decision to
perform a behavior in question. For example, in a model comparison study,
Mathieson (1991, p.187) reported that the "TAM explained attitude toward
using an IS much better than [the] TPB."
Since online shopping requires the consumer to extensively use Internet
technologies, the TAM can be applied to predict and understand the
consumer's intention to shop online. Considering the Internet as a system,
it can be expected that the more the consumers believe that using the
Internet will enhance their "shopping productivity" (e.g., perceived
usefulness) and will be easy to use (e.g., perceived ease of use), the more
they will use it for shopping (Koufaris, 2002, p.209).
As Table 1 shows, previous research has demonstrated that the TAM is
applicable to online shopping and World Wide Web studies. Based on the
assumption that the effect of ease of use on information technology usage
depends on the nature of the task, Gefen and Straub (2000) investigated
roles of usefulness and ease of use in two situations: an extrinsic task
(e.g., purchasing products online) and an intrinsic task (e.g., inquiring
about products through the Internet). They found that while both usefulness
and ease of use significantly influenced the intrinsic task, only
usefulness had a positive relationship with the extrinsic task. They
concluded that "extrinsic motivation is more important than intrinsic
motivation in IT adoption," (p.19) and many subsequent TAM studies (e.g.,
Chau and Hu, 2001; Koufaris, 2002) have supported this assertion.
[Insert Table 1 here]
Model comparison studies have been conducted to identify which model is
more likely to be suitable to predict user acceptance of IS. For example,
Davis et. al. (1989) compared the TRA and the TAM in terms of acceptance of
a word-processing program with 107 MBA students, and they found that the
TAM explained more variances of intention than did the TRA.
Results of previous studies comparing the TPB and the TAM have been
inconsistent. In a study investigating a physician's intention to adopt
technology, Chau and Hu (2001) found that the TAM is better than the TPB in
explaining the variance of behavioral intention. Mathieson (1991) also
found a similar result in a study of user acceptance of a spreadsheet. He,
however, insisted that it is difficult to determine that one model is
better than the other because the difference is not large. In addition, he
mentioned that while the "TAM supplies very general information about ease
of use and usefulness, [the] TPB delivers more specific information,
measuring the system's performance on various outcomes." (p.187)
On the other hand, Taylor and Todd (1995) found that although both the TAM
and the TPB have the same predictive power on actual usage behavior, the
TPB explains more variance of behavioral intention than does the TAM.
Riemenschneider (1997) also demonstrated that the TPB accounts for a small
firm's intention to adopt an IS slightly more than does the TAM.
Hypothesis 4: The TPB has better explanatory power than the TAM in
predicting the consumer's online-shopping intention.
Hypothesis 5: Ease of use is positively related to usefulness.
Hypothesis 6: Ease of use and usefulness are both positively related to
attitude toward online shopping.
Hypothesis 7: Usefulness and attitude are both positively related to
intention to shop online.
Methodology
Measures
Since the TPB was not developed to predict and understand a specific
behavior, analyzing a behavior in question requires eliciting its
distinctive and specific belief set (Ajzen and Fishbein, 1980). They
suggest three tactics for eliciting a belief set:
1. Beliefs may be the results of direct observation
2. [Beliefs] may be acquired indirectly by accepting information from
outside sources
3. [Beliefs] are self-generated through inference processes. (p.63)
They also asserted that it would be the simplest and most direct method to
ask subjects to describe the attitude object using "a free-response
format," (p.63) and that method has been applied most frequently in TPB
studies (e.g., Limayem et. al., 2000; Mathieson, 1991). However, the
authors of this study chose to indirectly elicit beliefs from extensive
study of prior literature in online shopping rather than from interviews
with potential subjects. This choice was made for two reasons: (1) to use
scales that have been validated in previous research and (2) to compare
three models (i.e., the TRA, the TPB, and the TAM) in a similar situation
by using two specific components of the TAM (i.e., usefulness and ease of
use) across the other models.
Four salient consequences that influence adoption of the Internet for
shopping were included to measure behavioral beliefs: usefulness, ease of
use, playfulness, and trust. Previous research has consistently
demonstrated the significant effects of both playfulness (Koufaris, 2002;
Moon and Kim, 2001; Webster, Trevino, and Ryan, 1993) and trust (Heijden
and Verhagen, 2002; Jarvenpaa and Todd,1997; Jarvenpaa and Tractinsky,
1999) on consumers' intentions to shop online or to use World Wide Web.
This study measures online shopping playfulness in terms of cognitive
enjoyment. More specifically, this study considers the extent to which the
consumer's cognitive curiosity is aroused when interacting with the
Internet for shopping and the extent to which the consumer finds the
interaction with the Internet for shopping intrinsically enjoyable and
interesting (Moon and Kim, 2001; Webster et. al., 1993). With respect to
trust, this study defines trust as "the subjective probability with which
consumers believe that [Web vendors] will perform a particular interaction
in a manner consistent with [consumers'] expectations." (Stewart, Pavlou,
and Ward, 2002, p.361) The definition of trust in this study involves three
dimensions: perception of security, perception of privacy, and perception
of product quality.
Based on Limayem et. al. (2000), three referent groups were included to
measure normative beliefs: friends, family, and media. Control beliefs
consisted of self-efficacy and technology facilitating conditions.
Self-efficacy refers to the individual's perception of "how well [s/he] can
execute courses of action required to deal with prospective situations."
(Bandura, 1982, p.122) According to social cognitive theory, self-efficacy
works as a crucial factor governing the individual's behavior (Bandura,
1986). In the online shopping context, self-efficacy can be defined as the
individual's perception of efficacy in purchasing products through the
Internet (Marakas, Yi, and Johnson, 1998). Self-efficacy has been found to
influence in online shopping behavior (Limayem et. al., 2000) and computer
usage (Compeau and Higgins, 1995). Facilitating conditions can be defined
as "objective factors, out there in the environment, that several judges or
observers can agree make an act easy to do." (Triandis, 1979, p.205) In the
context of online shopping, effective technological supports (e.g., fast
download speed, easy access to Web stores, and efficient transaction
process) of Web stores can be considered as a type of facilitating
condition. Limayem et. al. (2000) reported that technology facilitating
conditions were positively associated with Internet shopping behavior
directly and indirectly through intention.
Table 2 shows scales used to measure constructs of this study. All items
were measured with 7-point scales. As Fishbein and Ajzen (1975) suggested,
all items were adapted to specifically measure the consumer's online
shopping behavior. Appendix C shows sample items.
[Insert Table 2 here]
Measures of intention, attitude, subjective norm, perceived behavioral
control, usefulness, and ease of use were checked for reliability using
Cronbach's alpha and revealed reliable estimates: .99, .93, .83, .91, .90,
and .92 respectively. In the analyses, multiple items consisting of those
constructs were summed together for each construct, and these sums were
then divided by the numbers of the items included. Sets of behavioral
beliefs, normative beliefs, and control beliefs were not subject to
Cronbach's alpha of internal consistency because each belief construct was
estimated by summing all products of two items (e.g., a behavioral belief
(b) and its consequence (e)) across all beliefs in each construct. By the
same token, Harrison, Mykytyn, and Riemenschneider (1997, p.182) asserted that
Fishbein and Ajzen (1975) and Ajzen and Fishbein (1980) strongly
rule against calculating [reliability estimates for sets of beliefs] because
it is reasonable to assume that people will possess contrasting (positive
and negative) beliefs about the particular attributes of any reasonably
complex course of action. In a sense, then, internal consistency estimates
such as coefficient alpha for [_biei, _nbimci, and _cbippi] would not
be meaningful because such estimates assume a homogeneous
unidimensional construct.
Data collection
Zoomerang research company, a corporation specializing in sampling for
online surveys and online marketing research, hosted the online survey for
this study. Specifically, the company generated the sample for this study
from its database of registered panelists, sent invitation e-mails to
potential respondents, and automatically collected response data in a
spreadsheet format. The invitation e-mail contained an introduction to the
purpose of the study and the Website URL for the actual survey. The Website
was also provided by the company. Respondents could connect to the survey
Website by clicking the URL embedded in the e-mail.
The survey was conducted from the end of November to early December 2002.
A stratified random sample of 4,000 potential respondents was initially
drawn from Zoomerang's zSample database of registered panelists. The sample
for this study was generated to be representative of U.S. Internet users,
based on Harris Interactive's study depicting the profile of U.S. Internet
users (Greenspan, 2002). A week after sending the first invitation e-mail,
Zoomerang sent a reminder e-mail to potential respondents who had not yet
responded. At this time, another invitation e-mail was also sent to an
additional 500 panelists between the ages of 18 and 29. Because the company
allows researchers to check survey responses hours or even minutes after
deploying their surveys, the authors were able to note the low response
rate of this age group – statistically the largest group of Internet users
– from the first invitation e-mail. The additional 500 invitations helped
supplement this demographic so that the respondent group would match the
profile of the U.S. online population. Of a sample of 4,500 potential
respondents, a total of 807 respondents completed surveys.
Evaluation of assumptions
Prior to main analyses, data screening and an evaluation of assumptions
were preformed with a sample of 807 responses. Twenty eight cases were
deleted because of severe missing values, leaving 779 cases. In a handful
of cases, values randomly left out were replaced by the
expectation-maximization method through SPSS MVA.
The assumptions of normality and linearity were satisfactory. Thirty four
cases that had extremely low or high z scores (z _ _ 3.29, p < .001) were
found to be univariate outliers and deleted from analyses. Using
Mahalanobis distance, 12 cases were multivariate outliers (p < .001) and
also deleted. Multicollinearity and singularity were not a threat in this
data set. Analyses for this study were performed using data from 733
respondents.
Sample description
Of a sample of 4,500 prospective respondents, a total of 807 respondents
completed surveys, and 733 valid responses were used for analyses. The
overall response rate was 17.9 percent. The respondents' profile is
presented in Table 3. The total number of respondents within each
demographic group is different because of missing values.
Of the eligible respondents, slightly more than 50 percent were male. In
terms of age, both age groups of 18 to 29 and 40 to 49 showed high response
rate. The two age groups took approximately 57 percent of the respondents.
The majority of respondents were white (86 percent). For income,
respondents were fairly distributed within four household income categories
except the category of less than $20,000 (6 percent). The majority of
respondents (58 percent) reported their education level as college graduate
or higher. About 54 percent were married, and 46 percent were single.
[Insert Table 3 here]
Findings
Using AMOS 4.0 with a maximum-likelihood estimation, path analysis was
conducted to test the hypothesized models and paths in the models. To
assess models, this study used five indices: the x2 statistic, the
goodness-of-fit index (GFI), the adjusted goodness-of-fit index (AGFI), the
normed fit index (NFI), and the comparative fit index (CFI). Although the
x2 is the most common fit test, it was not significantly considered to
assess the fit of the model because the x2 is very sensitive to sample size
(Baumgartner and Homburg, 1996). It has been reported that the x2 test is
not a good index for model fit when the sample is large (Baumgartner and
Homburg, 1996; Marsh, Balla, and McDonald, 1988). A chi-square greater than
.05, a GFI greater than .90, an AGFI greater than .80, a NFI greater than
.90, and a CFI greater than .90 are considered as indicating a good fit
(Baumgartner and Homburg, 1996; Segars and Grover, 1993). Although AGFI
values exceeding .90 are preferable, the more liberal cutoff of .80 has
been used for a good model fit (Chau, 1997; Taylor and Todd, 1995).
In addition, path coefficients (e.g., direct, indirect, and total causal
effects) and predictive powers were examined for each model. Standardized
path coefficients were used for path coefficients. The total causal effect
of an independent construct on a dependent construct is the summation of
its direct and indirect effects. Predictive power was examined by using R2
for each dependent construct.
The theory of reasoned action
Path analysis to examine the theory of reasoned action resulted in a
less-than-adequate model fit. The path analysis revealed a significant x25
(176.96, p < .0001). Although the GFI was .92, other model fit indices were
not satisfactory (AGFI = .75, NFI = .87, and CFI = .87).
The coefficients for determination were .39 for intention, .45 for
attitude, and .29 for subjective norm. As shown in Figure 3, all paths in
this model were positively significant. Table 4 presents indirect and total
causal effects of each construct on intention to shop online (see Figure 3
for direct effects). Interestingly, the indirect effect (_ = .41, p < .01)
of behavioral beliefs on intention was greater than the direct effect (_ =
.07, p < .05) of subjective norm on intention. Attitude had the strongest
total causal effect (_ = .61, p < .01) on intention to shop online.
[Insert Figure 3 here]
[Insert Table 4 here]
The theory of planned behavior
Path analysis to examine the theory of planned behavior revealed an
adequate model fit. Although the x212 value was significant (231.87, p <
.0001), other model fit indices were satisfactory for a good model fit (GFI
= .92, AGFI = .82, NFI = .90, and CFI = .91).
The coefficients of determination were .43 for intention, .45 for
attitude, .29 for subjective norm, and .30 for perceived behavioral control
(see Figure 4). Compared with the TRA, the TPB – which adds the two
additional constructs of control beliefs and perceived behavioral control –
revealed a slight increase of explanatory power on intention (R2I = .43 and
.39 for the TPB and the TRA, respectively) and better model fit indices.
Thus, hypothesis 1 – the TPB has better explanatory power than the TRA in
predicting the consumer's online-shopping intention – is supported.
As summarized in Figure 4, all paths – except that between subjective norm
and intention – were positively significant. Thus, hypotheses 2a, 2b, 2c,
3a, and 3c are supported but hypothesis 2b is not.
As Table 4 presents, while behavioral belief and control belief constructs
(i.e., _biei and _cbippi) have significant indirect effects on intention to
shop online, the normative belief construct (i.e., _nbimci) has a
non-significant indirect effect on intention.
[Insert Figure 4 here]
The technology acceptance model
Path analysis to examine the technology acceptance model resulted in an
adequate model fit. Again the x21 value was significant (45.63, p < .0001),
and this time other model fit indices were also satisfactory for a good
model fit (GFI = .97, AGFI = .91, NFI = .97, and CFI = .97).
The coefficients for determination were .43 for intention, .47 for
attitude, and .42 for usefulness (see Figure 5). Although the TAM showed
somewhat-improved model fits, it revealed the same explanatory power on
intention as did the TPB (R2I = .43 for both the TPB and TAM). Thus,
hypothesis 4 – the TPB has better explanatory power than the TAM in
predicting the consumer's online-shopping intention – is not supported.
As summarized in Figure 5, all paths in the TAM model were positively
significant. Therefore, hypotheses 5, 6, and 7 are supported.
As shown in Table 4, both usefulness and ease of use had significant
indirect effects on intention. With respect to direct, indirect, and total
causal effects of usefulness and ease of use on attitude, this study found
that although the direct effect of usefulness (_ = .52, p < .01) was
greater than that of ease of use (_ = .22, p < .05), the total causal
effect of usefulness (_ = .52, p < .01) was smaller than that of ease of
use (_ = .56, p < .01). These findings suggest that usefulness takes an
important mediating role between ease of use and attitude toward online
shopping.
[Insert Figure 5 here]
Discussion
Model comparison
Three models (i.e., the TRA, the TPB, and the TAM) were compared in terms
of the power of each to predict and understand the consumer's
online-shopping behavior based on model fit indices and explanatory powers
in predicting intention to shop online. The fit statistics revealed that
the TRA provided a somewhat-inadequate model fit to the data. This finding
indicates that the TRA would be improved by including other variables. The
TPB, which adds perceived behavioral control to the TRA, provided a better
model fit and a moderately better explanatory power than did the TRA (R2I =
.43 and .39 for the TPB and the TRA, respectively). These results are
consistent with previous studies (Ajzen and Madden, 1986; Chang, 1998;
Madden, Ellen, and Ajzen, 1992) and indicate that non-volitional factors
such as the perception of ability to use the Internet for shopping and
Webstore features (e.g., site accessibility, product description, and
loading speed of Websites) significantly affect intention to shop online.
Thus, it can be assumed that the inclusion of perceived behavioral control
takes an important role in predicting and explaining the consumer's
online-shopping intention and provides a more robust theoretical basis for
the study of online shopping.
The TAM that has been typically adopted in IS studies (e.g., Davis, 1989;
Davis, et. al., 1989; Mathieson, 1991) was also successfully applied to the
domain of online shopping behavior. The TAM revealed an adequate model fit
to the data and the same amount of variance in explaining intention to shop
online as did the TPB (R2I = .43 for both models). With respect to the
variance in explaining attitude, the TAM was better than the TPB, but the
difference was not severe (R2A = .47 and .45 for the TAM and the TPB,
respectively). Previous studies comparing the two models had revealed
inconsistent results. While Mathieson (1991) and Chau and Hu (2001) found
that the TAM had more explanatory power in predicting behavioral intention
than did the TPB, Taylor and Todd (1995) demonstrated that the TPB
accounted for more variance than did the TAM. Results of the previous
studies, however, were similar in that the R2I differences were not severe.
These results make it difficult to determine which model is best to apply
to practical situations. In addition, the two models have the trade-off
between information provided by models and parsimony (Taylor and Todd,
1995). While the TPB provides more information, it is also more complex to
use. The TAM is more parsimonious and provides directive information, but
it explains only attitudinal components. Thus, this study concludes that
the best model to use may depend on variables of interest. If the research
focus is Internet technology usage in online shopping, the TAM would be
preferable to the TPB. If a focus of research encompasses social and
non-volitional factors in online shopping, the decomposed TPB would be
better than the TAM.
Effects on intention to shop online
This study found that, across the three models, attitude toward online
shopping had the strongest effect on intention to shop online. Recently,
several TAM studies (Gefen and Straub, 2000; Lederer et. al., 2000;
Venkatesh and Davis, 1996; Venkatesh and Morris, 2000) have adopted a
diminished TAM that excludes attitude. The reasoning for this was that
attitude did not function as a strong mediator between its antecedents
(i.e., usefulness and ease of use) and intention (Chau and Hu, 2001). In
order to test the role of attitude in the online-shopping context, this
study compared path coefficients and explanatory powers of the diminished
TAM and the original TAM. Table 5 shows the results. Although the total
causal effect of ease of use on intention increased in the diminished TAM,
the explanatory power of intention and the total causal effect of
usefulness on intention decreased in the diminished TAM (R2I = .37 and .43
for the diminished TAM and the TAM, respectively). It should be also noted
that the indirect effect of ease of use on intention dramatically decreased
in the diminished TAM, and the indirect effect of usefulness on intention
was greater than the direct effect of usefulness on intention in the TAM.
Thus, it seems reasonable to conclude that attitude takes an important
mediating role between its antecedents and intention in the online-shopping
context.
[Insert Table 5 here]
Interestingly, this study found that subjective norm was not a significant
predictor of intention to shop online in the TPB. As Davis et. al. (1989,
p.986) indicated that "subjective norm may influence behavioral intention
indirectly via attitude," several TPB studies (Chang, 1998; Malhotra and
Galletta, 1999) found significant direct effect of subjective norm on
attitude, although no direct effect of subjective norm on intention was
observed. This study also tested the causal link between subjective norm
and attitude in the online-shopping context by subtracting the link between
subjective norm and intention. The result revealed significant direct
effects of subjective norm on attitude (_ = .16 for both the TRA and the
TPB) and significant indirect effects of subjective norm on intention to
shop online (_ = .10 for the TRA and _ = .08 for the TPB, p < .01). It
should also be noted that the indirect effect of subjective norm on
intention in the alternative TRA that links subjective norm with attitude
instead of intention was greater than the direct effect of subjective norm
on intention in the original TRA. Supporting Davis et. al.'s (1989)
assumption, these findings indicate that subjective norm influences
intention through attitude indirectly rather than directly in the
online-shopping context. Thus, it can be assumed that consumers consider
how important-others evaluate shopping online when the consumers form their
own attitudes toward online shopping, and, their intention is, in turn,
determined by attitude. Since using the Internet for shopping is fairly
voluntary, personal, and individual, as compared to the use of technology
in the workplace, consumers may rely actively on individual desire to shop
through the Internet.
The TPB revealed direct effects of perceived behavioral control on
intention to shop online. As mentioned before, the addition of perceived
behavioral control to the TRA revealed a better model fit and explained the
increased variance in intention to shop online. These findings indicate
that online shopping is not totally under the consumer's volitional
control. In other words, resources and abilities to use the Internet for
shopping are required to successfully shop online.
As expected, attitude, subjective norm, and perceived behavioral control
in the TPB were direct functions of their corresponding beliefs. In the
TPB, normative beliefs did not have a significant indirect effect on
intention, whereas behavioral beliefs and control beliefs had significant
indirect effects on intention. In the TAM, usefulness had significant
direct and indirect effects on intention to shop online. These findings
imply that online shoppers have a utilitarian nature to improve their
shopping productivity by saving money and time and easily comparing
products. Ease of use also had a significant indirect effect on intention.
Interestingly, the indirect effect of ease of use on intention was
comparable to both the total causal effect of usefulness on intention and
the direct effect of attitude on intention. This result indicates that ease
of use takes an important role as an antecedent affecting the consumer's
attitude formation and perception of usefulness. The significant roles of
usefulness and ease of use in this study confirm that the TAM can be
successfully applied to the domain of online shopping beyond IS usage.
Limitations and future research
This study has a number of limitations, and these limitations would
provide guidelines for future research to improve our understanding of
online-shopping behavior. First, this study did not capture consumers'
actual shopping behavior. Since the positive relationship between intention
and actual behavior has been consistently demonstrated in the TRA, the TPB,
and the TAM studies (e.g., Davis et. al., 1989; George, 2002; Moon and Kim,
2001; Limayem et. al., 2000; Taylor and Todd, 1995), it is expected that
this limitation does not pose a serious problem in interpreting findings of
this study and that further studies that measure actual behavior in the
online-shopping context resolve the concern.
The second limitation is subject to the convenience sample comprising
panelists who registered with an online research company. Panelists may
possess better Internet skills and have experienced more online shopping
than the average user. This limit embodied in the sample makes it difficult
to generalize the findings of this study. Thus, it is hoped that future
research using different populations of consumers would test validations of
the findings and expand our understanding of online-shopping behavior.
The third limitation results from the method to elicit a distinctive and
specific belief set for the TPB. As mentioned before, this study adapted
beliefs validated in previous studies, instead of eliciting beliefs through
interviews with potential respondents as Ajzen and Fishbein (1980)
recommended. This might have led the TRA and the TPB to account for less
than fifty percent of the variance in intention; however, many studies in
journalism and mass communication also account for less variance than fifty
percent. These results imply the need to continue exploration of other
important variables that influence the consumer's intention to shop online.
Thus, it is suggested that future research uses in-depth interviews with
consumers as well as extensive study of related literature to expand the
scope of antecedents hampering or increasing the consumer's online-shopping
intention.
Beyond these limitations, more research is required to continue to test
how subjective norm affects intention in the online-shopping context. As
explained earlier, this study found the indirect effect of subjective norm
on intention through attitude instead of its direct effect on intention as
proposed by the TPB. The relationship between subjective norm and
attitude/intention may be different based on demographic characteristics,
level of Internet skills, and product category. For example, when
purchasing high-touchable products (e.g., clothing, flowers, jewelry, and
furniture) through the Internet, risks of product satisfaction rise because
the consumer cannot taste, smell, and touch the product when purchasing
online. Therefore, it is expected that the consumer will consider
important-referents' responses and will contemplate his/her ability to
successfully shop online more carefully than when purchasing low-touchable
products (e.g., books, music, and computers) through the Internet.
Appendix A: Tables
Table 1. TAM research in online shopping and World Wide Web
Authors
Constructs
Applications
Methodology
Findings
Gefen and Straub
BI, Intended inquiry,
Online shopping
Experiential
BI = U, Inquiry = U + EOU,
(2000)
U, & EOU,
survey
& U = EOU
Koufaris (2002)
BI, U, & EOU
Online shopping
E-mail survey
BI = U
Lederer et. al. (2000)
Usage,U, & EOU
WWW
E-mail survey
Usage = U + EOU
Moon and Kim (2000)
Usage, BI, A, U, EOU,
WWW
Survey
Usage = BI, BI = A + U + P,
& Playfulness (P)
A = U + EOU + P, U = EOU,
& P = EOU
Morris and Dillon
Usage, BI, A, U,
Netscape
Survey
Usage = BI, BI = A + U,
(1997)
& EOU
& A = U + EOU
Table 2. Summary of scales
Construct
Source
No. of items
Scales
Intention
Limayem et. al. (2000)
Three
7-point Likert
Attitude
Taylor and Todd (1995)
Four
7-point semantic
Subjective norm
Mathieson (1991)
Three
7-point Likert
Behavioral control
Taylor and Todd (1995)
Three
7-point Likert
Behavioral beliefs
Usefulness
Davis (1989) & Limayem et. al. (2000)
Six
7-point Likert
Ease of use
Davis (1989)
Four
7-point Likert
Trust
Jarvenpaa and Todd (1997)
Three
7-point Likert
Playfulness
Koufaris (2002) & Webster et. al. (1993)
Six
7-point Likert
Normative beliefs
Friend influence
Limayem et. al. (2000)
Two
7-point Likert
Family influence
Limayem et. al. (2000)
Two
7-point Likert
Media influence
Limayem et. al. (2000)
Two
7-point Likert
Control beliefs
Self-efficacy
Hollenbeck and Brief (1987)
Four
7-point Likert
Technology
Limayem et. al. (2000)
Four
7-point Likert
Table 3. Demographic characteristics of respondents
Variable
Description
Frequency
Percent
Gender
Male
384
53%
Female
343
47%
Age
18-29
244
33%
30-39
148
20%
40-49
175
24%
50-64
121
17%
Over 65
44
6%
Race
White
629
86%
African-American
34
5%
Hispanic
17
2%
Asian
21
3%
Other
28
4%
Income
Less than $20,000
41
6%
$20,000 - $40,000
178
25%
$40,000 - $60,000
173
24%
$60,000 - $80,000
136
19%
Over $80,000
183
26%
Education
Less than high school
9
1%
High school graduate
68
9%
Some college
229
31%
College graduate
242
33%
Some graduate school or higher
183
25%
Marital status
Single
334
46%
Married
395
54%
Note: Missing values lead to the variance in respondent numbers of each
demographic group.
Table 4. Effects on behavioral intention of the models
TRA
TPB
TAM
Indirect effect
åbiei
.41**
.33**
-
ånbimci
.04**
.02
-
åcbippi
-
.17**
-
Usefulness
-
-
.25**
Ease of use
-
-
.42**
Total causal effect
Attitude
.61**
.50**
.48**
Subjective norm
.07*
.03
-
Perceived behavioral control
-
.30**
-
åbiei
.41**
.33**
-
ånbimci
.04**
.02
-
åcbippi
-
.17**
-
Usefulness
-
-
.48**
Ease of use
-
-
.42**
* p < .05 and ** p < .01
Table 5. Path analysis results of the TAM and the diminished TAM^
The TAM
The diminished TAM
Direct effects
Ease of use to usefulness
.65**
.65**
Ease of use to attitude
.22**
-
Ease of use to intention
-
.34**
Usefulness to attitude
.52**
-
Usefulness to intention
.23*
.33**
Attitude to intention
.48**
-
Indirect effects on intention
Ease of use
.42**
.21**
Usefulness
.25**
Total causal effects on intention
Ease of use
.42**
.56**
Usefulness
.48**
.33**
^ The diminished TAM excludes attitude in the original TAM model.
* p < .05 and ** p < .01
Appendix B: Figures
Appendix C: Sample questionnaire items
Attitudinal constructs
Playfulness (Sample)
b: I would/do find it enjoyable to use the Internet for shopping.
e: How important is it to you to enjoy shopping?
b: I would/do find it exciting to use the Internet for shopping.
e: How important is it to you to be excited about shopping?
Usefulness (Sample)
b: Online shopping would/does enable me to save time.
e: How important is it to you to save time when shopping?
b: Online shopping would/does allow me to save money.
e: How important is it to you to save money when shopping?
Ease of use (Sample)
b: The process of interacting with the Internet for shopping would be/is
clear and understandable.
e: How important is it to you that your shopping process is clear and
understandable?
b: Learning to use the Internet for shopping would be/is easy for me.
e: How important is it to you that the system you use to shop is easy to
learn?
Trust (Sample)
b: Products purchased on the Internet would/do meet my expectations.
e: How important is it to you to purchase products that meet your
expectations?
Normative constructs (Sample)
nb: My friends would/do (oppose/support) my shopping online.
mc: Generally speaking, I don't want to do what my friends oppose.
nb: My family would/does (disapprove/approve) of my shopping online.
mc: Generally speaking, I want to do what my family approves.
nb: Media frequently encourage people to shop online.
mc: Generally speaking, I want to do what media encourage people to do.
Control constructs
Self-efficacy (Sample)
cb: I would be/am confident in my abilities to use the Internet for shopping.
pp: For me, being confident in my abilities to use a system for shopping is
(Extremely unimportant/Extremely important).
cb: It would be/is possible for me to shop online at the level I would like.
pp: For me, the possibility of shopping at the level I would like is
(Extremely unimportant/Extremely important).
Facilitating conditions (Sample)
cb: Online stores would be/are easily accessible (e.g., through search
engines, cyber malls, and Web ads). pp: For me, easy accessibility of
online stores (e.g., through search engines, cyber malls, and Web ads)
would be/is (Extremely unimportant/Extremely important).
cb: Products of online stores would be/are well described.
pp: For me, descriptions of products online would be/is (Extremely
unimportant/Extremely important).
Attitude
Using the Internet for shopping would be/is a (bad/good, foolish/wise, and
unpleasant/pleasant) idea.
I (dislike/like) the idea of using the Internet for shopping
Subjective norm
People who are important to me would/do (oppose /support) my shopping online.
People who are important to me would/do (disapprove/approve) of my shopping
online.
Media that promote opinions I value encourage people to shop online.
Perceived behavioral control
I am able to use the Internet to shop online.
Online shopping is entirely within my control.
I have the resources, the knowledge and the ability to use the Internet for
shopping.
Behavioral Intention
I intend to shop online in next six months.
It is likely that I will shop online in next six months.
I expect to shop online in next six months.
References
Abdul-Gader, A. H. & Kozar, K. A. (1995). The impact of computer
alienation on information technology investment decisions: An exploratory
cross-national analysis. MIS Quarterly, 19(4), 535-559.
Ajzen, I. (1985). From intentions to action: A theory of planned behavior.
In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior
(pp. 11-39). Heidelberg: Springer.
(1991). The theory of planned behavior. Organizational
Behavior and Human Decision Processes, 50(2), 179-211.
, & Fishbein, M. (1980). Understanding attitudes and
predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall, Inc.
, & Madden, T. J. (1986). Prediction of goal-directed
behavior: Attitudes, intentions, and perceived behavioral control. Journal
of Experimental Social Psychology, 22, 453-474.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American
Psychologist, 37(2), 122-147.
(1986). Social foundations of thought and action: A
social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall, Inc.
Baumgartner, H., & Homburg, C. (1996). Applications of structural equation
modeling in marketing and consumer research: A review. International
Journal of Research in Marketing, 13(2), 139-161.
Chang, M. K. (1998). Predicting unethical behavior: A comparison of the
theory of reasoned action and the theory of planned behavior. Journal of
Business Ethics, 17(16), 1825-1834.
Chau, P. Y. K. (1997). Reexamining a model for evaluating information
center success using a structural equation modeling approach. Decision
Sciences, 28(2), 309-334.
, Au, G., & Tam, K. Y. (2000). Impact of information
presentation modes on online shopping: An empirical evaluation of a
broadband interactive shopping service. Journal of organizational Computing
and Electronic Commerce, 10(1), 1-22.
, & Hu, P. J. (2001). Information technology
acceptance by individual professionals: A model comparison approach.
Decision Sciences, 32(4), 699-719.
Compeau, D. R., & Higgins, C. A. (1995). Computer self-efficacy:
Development of a measure and initial test. MIS Quarterly, 19(2), 189-211.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user
acceptance of information technology. MIS Quarterly, 13, 319-340.
, Bagozzi, R. P., & Warshaw, P. R. (1989). User
acceptance of computer technology: A comparison of two theoretical models.
Management Science, 35, 982-1003.
Donthu, N., & Garcia, A. (1999). The Internet shopper. Journal of
Advertising Research, 39(3), 52-58.
Fishbein, M. & Ajzen, I. (1975). Belief, attitude, intention, and
behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.
Fram, E., & Grady, D. (1995). Internet buyers: Will the surfers become
buyers? Direct Marketing, 57(10), 63-65.
Gefen, D., & Straub, D. (2000). The relative importance of perceived ease
of use in IS adoption: A study of e-commerce adoption. Journal of the
Association for Information Systems, 1(1), 1-28.
George, F. F. (2002). Influences on the intent to make Internet purchases.
Internet Research, 12(2), 165-180.
Goldsmith, R. E. (2000). How innovativeness differentiates online buyers.
Quarterly Journal of Electronic Commerce, 1(4), 323-333.
Greenspan, R. (April 17, 2002). Two-third hit the net. Insert online
address [Online]. Available:
http://cyberatlas.internet.com/big_picture/geographics/article
/0,,5911_1011491,00.html.
Harrison, D. A., Mykytyn, Jr., P. P., & Riemenschneider, C. K. (1997).
Executive decisions about adoption of information technology in small
business: Theory and empirical tests. Information Systems Research, 8(2),
171-195.
Heijden, H., & Verhagen, T. (2002). Measuring and assessing online store
image: A study of two online bookshops in the Benelux. Proceedings of the
35th Hawaii International Conference on System Sciences.
Hoffman, D., & Novak, T. (1997). A new marketing paradigm for electronic
commerce. The Information Society, 13(1), 43-54.
Jarvenpaa, S., & Todd, P. (1997). Is there a future for retailing on the
Internet? In Peterson, R. A. (Eds.), Electronic marketing and the consumer
(pp. 139-154). Thousands Oak, CA; Sage Publications.
, & Tractinsky, N. (1999). Consumer trust in an
Internet store: A cross-cultural validation. Journal of Computer-Mediated
Communication [Online], 5(2). Available:
http://www.ascusc.org/jcmc/vol5/issue2/jarvenpaa.htm.
Jupiterdirect (2002). ECommerce B2C: US consumer shopping, buying, and
demographics [Online]. Available:
http://jupiterdirect.com/item/0,,2317769_1,00.html.
Koufaris, M. (2002). Applying the technology acceptance model and flow
theory to online consumer behavior. Information Systems Research, 13(2),
205-223.
Lederer, A. L., Maupin, D. J., Sena, M. P., & Zhuang, Y. (2000). The
technology acceptance model and the World Wide Web. Decision Support
Systems, 29(3), 269-282.
Limayem, M., Khalifa, M., & Frini, A. (2000). What makes consumers buy from
Internet? A longitudinal study of online shopping. IEEE Transactions on
Systems, Man, and Cybernetics – Part A: Systems and Humans, 30 (4), 421-432.
Madden, T., Ellen, P. S., & Ajzen, I. (1992). A comparison of the theory of
planned behavior and the theory of reasoned action. Personality and Social
Psychology Bulletin, 18 (1), 3-9.
Malhotra, Y., & Galletta, D. (1999). Extending the technology acceptance
model to account for social influence: Theoretical bases and empirical
validation. Proceedings of the 32nd Hawaii International Conference on
System Sciences.
Marakas, G. M., Yi, M. Y., & Johnson, R. D. (1998). The multilevel and
multifaceted character of computer self-efficacy: Toward clarification of
the construct and an integrative framework for research. Information
Systems Research, 9(2), 126-163.
Marsh, H. W., Balla, J. R., & McDonald, R. P. (1988). Goodness-of-fit
indexes in confirmatory factor analysis: The effect of sample size.
Psychological Bulletin, 103(3), 391-410.
Mathieson, K. (1991). Predicting user intentions: Comparing the technology
acceptance model with the theory of planned behavior. Information Systems
Research, 2(3), 173-191.
Moon, J., & Kim, Y. (2001). Extending the TAM for a World-Wide-Web
context. Information & Management, 38, 217-230.
Peterson, R. A., Balasubramanian, S., & Bronnenberg, B. J. (1997).
Exploring the implications of the Internet for consumer marketing. Journal
of the Academy of Marketing Science, 25(4), 329-346.
Phau, I., & Poon, S. M. (2000). Factors influencing the types of products
and services purchased over the Internet. Internet Research, 10(2), 102-113.
Riemenschneider, C. K. (1997). Understanding IT adoption/acquisition in
small business: A comparison of three models. Unpublished doctoral
dissertation, The University of Texas at Arlington.
Segars, A. H., & Grover, V. (1993). Re-examining perceived ease of use and
usefulness: A confirmatory factor analysis. MIS Quarterly, 17(4), 517-525.
Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of
reasoned action: A meta-analysis of past research with recommendations for
modifications and future research. Journal of Consumer Research, 15(3),
325-343.
Shim, S., & Drake, M. (1990). Consumer intention to utilize electronic
shopping. Journal of Direct Marketing, 4(3), 22-33.
Stewart, D. W., Pavlou, P., & Ward, S. (2002). Media influences on
marketing communications. In J. Bryant & D. Zillmann (Eds.), Media effects:
Advances in theory and research (pp. 353-396). Hillsdale, NJ: Erlbaum.
Swaminathan, V., Lepkowska-White, E., & Rao, B. P. (1999). Browsers or
buyers in cyberspace? An investigation of factors influencing electronic
exchange. Journal of Computer-Mediated Communication [Online], 5(2).
Available: http://www.ascusc.org/ jcmc/vol5/issue2/swaminathan.htm.
Taylor, S., & Todd, P. (1995). Understanding information technology usage:
A test of competing models. Information Systems Research, 6, 144-176.
Teo, T., Lim, V., & Lai, R. (1999). Intrinsic and extrinsic motivation in
Internet usage. OMEGA: The International Journal of Management Science, 27,
25-37.
Triandis, H. C. (1979). Values, attitudes, and interpersonal behavior. In
Nebraska symposium on motivation, 1979: Beliefs, attitudes, and values (pp.
195-259). Lincoln, NE: University of Nebraska Press.
U.S. Census Bureau (2002). United States Department of Commerce News
[Online]. Available: http://www.census.gov/mrts/www/current.html.
Venkatesh, V., & Davis, F. (1996). A model of the antecedents of perceived
ease of use: Development and test. Decision Sciences, 27(3), 451-481.
, & Morris, M. G. (2000). Why don't men ever stop to
ask for directions? Gender, social influence, and their role in technology
acceptance and usage behavior. MIS Quarterly, 24(1), 115-139.
Webster, J., Trevino, L. K., & Ryan, L. (1993). The dimensionality and
correlates of flow in human-computer interactions. Computers in Human
Behavior, 9, 411-426.
Yoh, E. (1999). Consumer adoption of the Internet for apparel shopping.
Unpublished doctoral dissertation, Iowa State University.
|
