Year 2005 UCD— ITS— RR— 05— 07
The Intended and Actual Adoption of Online Purchasing:
A Brief Review of Recent Literature
Xinyu Cao
Patricia L. Mokhtarian
Institute of Transportation Studies ◊ University of California, Davis
One Shields Avenue ◊ Davis, California 95616
PHONE: ( 530) 752- 6548 ◊ FAX: ( 530) 752- 6572
WEB: http:// its. ucdavis. edu/
The Intended and Actual Adoption of Online Purchasing:
A Brief Review of Recent Literature
Xinyu Cao
Department of Civil and Environmental Engineering
One Shields Avenue
University of California, Davis
Davis, CA 95616
E- mail: xinyucao@ gmail. com
and
Patricia L. Mokhtarian
Department of Civil and Environmental Engineering
and
Institute of Transportation Studies
One Shields Avenue
University of California, Davis
Davis, CA 95616
voice: ( 530) 752- 7062
fax: ( 530) 752- 7872
e- mail: plmokhtarian@ ucdavis. edu
May 2005
Revised in April 2007
This research is funded by the University of California Transportation Center.
i
Table of Contents
List of Tables ............................................................................................................................... ........................ ii
List of Figures........................................................................................................................ .............................. ii
Executive Summary........................................................................................................................ .................... iii
1. Introduction................................................................................................................... ................................... 1
2. Theoretical Frameworks ............................................................................................................................... ... 3
2.1 Theory of Reasoned Action ( TRA)......................................................................................................... 3
2.2 Theory of Planned Behavior ( TPB) ........................................................................................................ 4
2.3 Technology Acceptance Model ( TAM)................................................................................................... 5
2.4 Transaction Cost Theory ( TCT).............................................................................................................. 6
2.5 Innovation Diffusion Theory ( IDT) ........................................................................................................ 7
2.6 Other theories....................................................................................................................... .................. 8
2.7 Summary........................................................................................................................ ........................ 8
3. Dependent Variables of Interest ...................................................................................................................... 10
3.1 E- shopping intention...................................................................................................................... ...... 10
3.2 Actual e- shopping behavior .................................................................................................................. 11
3.3 Summary........................................................................................................................ ...................... 12
4. Sampling Approaches ............................................................................................................................... ..... 13
5. Product Classifications ............................................................................................................................... ... 16
6. Methodologies ............................................................................................................................... ................ 21
6.1 Descriptive analysis .............................................................................................................................. 21
6.2 Correlational analysis....................................................................................................................... .... 21
6.3 Multivariate analysis....................................................................................................................... ..... 22
6.4 Summary........................................................................................................................ ...................... 24
7. Determinants of E- shopping Behavior............................................................................................................ 25
7.1 Perceived characteristics of the web as a shopping channel ................................................................. 25
7.2 Vendor and product characteristics ....................................................................................................... 28
7.3 Online consumer characteristics ........................................................................................................... 29
7.4 Summary........................................................................................................................ ...................... 31
8. Summary........................................................................................................................ ................................ 39
References..................................................................................................................... ..................................... 41
Appendix: Overview of Previous Research on E- shopping Intention and Adoption.......................................... 50
ii
List of Tables
Table 1. Dimensions of E- shopping Intention............................................................................................. 11
Table 2. Dimensions of E- shopping Behavior ............................................................................................ 12
Table 3. Product and Service Classification................................................................................................ 17
Table 4. Likely Consumer Decision Sequences.......................................................................................... 18
Table 5. Product Type Description.............................................................................................................. 19
Table 6. Summary of the Effect of Antecedents on the Intention and Usage of Online Shopping ............. 32
Table 7. Summary of the Findings of the Effect of Antecedents of Attitude toward Online Shopping ...... 35
Table 8. Summary of the Findings of the Effect of Antecedents of Risk Perception.................................. 36
Table 9. Summary of the Findings of the Effect of Antecedents of Trust ................................................... 37
Table 10. Summary of the Determinants of Online Shopping Discussed in Popular Press ........................ 38
List of Figures
Figure 1. Theory of Reasoned Action ........................................................................................................... 4
Figure 2. Theory of Planned Behavior.......................................................................................................... 5
Figure 3. Technology Acceptance Model...................................................................................................... 6
Figure 4. Transaction Cost Theory................................................................................................................ 7
Figure 5. Typology of Experiential Value ..................................................................................................... 8
iii
Executive Summary
The rapid growth of e- commerce is imposing profound impacts on modern society. On the supply side, the
emergence of e- commerce is greatly changing the operation behavior of some retailers and is increasing product
internationalization due to its geographically unlimited nature. On the demand side, the pervasiveness of
e- commerce affects how, where, and when consumers shop, and indirectly influences the way in which we live
our lives. However, the development of e- commerce is still in an early stage, and why consumers choose ( or do
not choose) online purchasing is far from being completely understood. To better evaluate and anticipate those
profound impacts of e- commerce, therefore, it is important to further refine our understanding of consumers’
e- shopping behavior.
A number of studies have investigated e- shopping behavior, and reviewing them is valuable for further
improving our understanding. This report aims to summarize previous e- shopping research in a systematic way.
In this review, we are interested primarily in the potential benefits and costs that the internet offers for the
business- to- consumer segment of e- commerce in the transaction ( purchase) channel. An overview of the 65
empirical studies analyzed in this report is provided in the Appendix.
Most previous studies fall into one or more of several theoretical frameworks, including the theory of reasoned
action, the theory of planned behavior, the technology acceptance model, transaction cost theory, innovation
diffusion theory, and others. Among them, social psychological theories ( the theory of reasoned action, the
theory of planned behavior, the technology acceptance model) were widely applied. As shown in the
applications of different theories, e- shopping behavior is not a simple decision process, and thus an integration of
various theories is necessary to deal with its complexities. We suggest synthesizing these theories through the
development of a comprehensive list of benefits and costs, using each of the key constructs of the pertinent
theories as a guide to identifying the nature of those benefits and costs.
The dependent variables mainly include e- shopping intention and actual e- shopping behavior ( a few studies used
attitudes toward e- shopping). E- shopping intention was measured by various dimensions. Among them, the
directly- stated intention to purchase online is the most frequently used measure. Although some studies used a
unidimensional measure, most adopted a latent construct to assess consumers’ e- shopping intentions. Actual
e- shopping behavior mainly includes three dimensions: adoption, spending, and frequency. Most studies
examined one or more of these three dimensions directly, while a few studies constructed a latent variable to
measure actual e- shopping behavior. When both behavioral intention and actual behavior are included in model
development, attention should be paid to the time precedence between intention and behavior.
With respect to sampling, a choice- based sampling approach is probably preferable given that online shopping
activity accounts for a minor proportion of all consumers, and a far smaller proportion of total retail sales. In
previous studies, most chose internet/ computer/ email users or students as their subjects. Generally, a student
sample is a natural choice for some particular products such as books. However, parameter estimates developed
from a student sample lack generalizability to a larger population because of its homogeneity. By contrast, a
more general sample of internet/ computer/ email users is more applicable for e- shopping behavior research.
iv
The characteristics of products strongly influence the degree to which they are suitable for selling online.
Mixing product categories in e- shopping behavior research tends to yield vague or inconsistent results. It is
therefore necessary to explicitly consider product characteristics when exploring consumers’ e- shopping
behavior. However, relatively little effort has been invested into product classification in the context of
e- shopping. Although Nelson’s dichotomized system ( search and experience goods) and Peterson et al.’ s
three- dimensional system ( cost, tangibility, and differentiability) provide useful guides for product type
classification, each has some shortcomings. Therefore, more research should focus on the construction of
product classification systems.
Different methodologies have been applied in previous research. Generally, descriptive analysis is used to
describe consumers’ e- shopping behavior; correlational analysis goes beyond descriptive analysis and attempts
to analyze how two variables are related; and multivariate analysis is mainly used to explain consumers’
behavior using many variables considered together. Therefore, although descriptive and correlational analyses
are important steps in helping to construct multivariate analyses, multivariate studies provide more information
than these other two types of analyses. Multivariate analysis is ideal to study e- shopping behavior in depth.
Among multivariate analysis techniques, multiple regression, structural equations modeling, and discrete choice
models were most frequently used.
Previous studies have identified various determinants of consumers’ e- shopping behavior. These determinants
mainly cover three essential elements: characteristics of e- shopping as a shopping channel, consumer
characteristics, and vendor and product characteristics. Among these characteristics, the former two have been
examined extensively in previous research, confirming their importance in understanding e- shopping behavior.
Specifically, the dimensions of channel characteristics of e- shopping include e- shopping service quality, relative
advantages, perceived risk of and confidence in e- shopping, and trust. Consumer characteristics include their
shopping orientations, personality, social and psychological characteristics, computer/ internet experience,
in- home shopping experience, and socio- demographics.
1
1. Introduction
New information and communication technologies ( ICTs) have had profound and pervasive impacts on modern
society – impacts that are still evolving with the ongoing development of new technologies and the continued
adoption/ adaptation of older ones. ICT applications and services are changing how and where we work, shop,
play, travel, and in other ways live our lives. In particular, the emergence and spread of electronic commerce
( e- commerce) have greatly influenced the behavior of both retailers and consumers. For example, the
Association of National Advertisers stated that 44% of US companies were selling online, and an additional 36%
would participate in e- selling by 2000 according to Nua Internet Surveys ( the Internet Economy Indicators,
2004); in their sample of more than 2,000 households across the U. S., Cole et al. ( 2003) found that about 66% of
internet purchasers have reduced their purchasing in traditional retail stores somewhat or a lot in 2002.
E- commerce refers to “ the buying, selling, marketing, and servicing of products or services over computer
networks” ( Wikipedia, 2004). E- commerce consists of several components: business to business ( B2B),
business to consumer ( B2C), and consumer to consumer ( C2C) ( such as eBay). Although the B2B segment
dominates e- commerce in terms of the dollar value of transactions made, B2C remains important for its potential
impacts on consumers’ shopping behavior. As a flexible, interactive, efficient medium, e- shopping is likely to be
a substitute for traditional shopping media, and may well dominate the exchange of certain products ( e. g., digital
assets such as music, software, movies, and the written word) in the future. The review presented here focuses
on the B2C segment of e- commerce.
B2C e- commerce has grown rapidly worldwide over the past few years. In Singapore, about 30% of internet
users aged 15 years and above have purchased online as of 2004, an 18- percentage point increase since 2000; and
the average amount spent online per internet shopper has increased from $ 336 in 2003 to $ 539 in 2004 ( IDA,
2005). In South Korea, the B2C volume of the e- commerce market increased from $ 0.61 billion in 2000 to $ 2.15
billion in 2001 ( http:// www. ecommerce. or. kr/ about/ ec_ market1. asp, accessed on April 24, 2005). In the UK,
one in ten adults has frequently purchased online; 25% of adults were frequent online purchasers in some areas;
and total e- commerce revenues are expected reach 20 billion pounds by 2005, accounting for 7.6% of retail sales
( CACI, 2000). According to the US Census Bureau ( 2004), adjusted retail sales through e- commerce reached
$ 16.1 billion in the first quarter of 2004 ( compared to $ 5.9 billion in the first quarter of 2000), or 1.8% of total
retail sales. However, this percentage has yet to reach the starting point ( 2.5%) of the classical sigmoid diffusion
curve presented by Rogers ( 1983). On the other hand, the development of e- commerce is still in a volatile stage.
One survey of American internet users and non- users found that fewer adult internet users bought online in 2002
than in the previous two years, and that more than half of internet purchasers spent less than they intended when
shopping online ( Cole et al., 2003). The Boston Consulting Group estimated that 65% of internet buyers,
accounting for 80% of the dollar value of prospective purchases, terminated their transaction before checking out
( Maravilla, 2001). This evidence highlights the importance of understanding the behavior of online purchasers.
A growing number of studies have explored the determinants of consumers’ electronic- shopping ( e- shopping)
behavior during recent years. Most studies either explicitly or implicitly embedded online purchasing behavior
into various theoretical frameworks, including the theory of reasoned action, the theory of planned behavior, the
2
technology acceptance model, transaction cost theory, innovation diffusion theory, and others. According to the
dependent variables of interest, previous studies can be generally classified into three categories: online
purchasing intention, online purchasing adoption, or both. Further, these studies used different sampling
approaches such as internet- based survey, paper and pencil survey, telephone interview, or using an available
database, and recruited their samples from different populations, including students, computer/ internet/ email
users, online purchasers, and the general population. With respect to the product, some studies dealt with all
products sold online, while others chose particular products ( or product categories). In addition, a variety of
statistical methodologies ( including descriptive as well as multivariate analyses) has been applied, to answer
different kinds of questions.
Previous works identified various antecedent factors of online purchasing intention and usage, and hence greatly
improve our understanding of the behavior of online purchasers. This understanding offers a useful foundation
to a number of different actors. To retailers, it provides a basis for implementing e- commerce promotion policies
and better designing e- retail sites; to academics, it serves as an important example of the impact of ICT in a key
realm of everyday life; and to transportation and urban planners, it presents a critical background from which to
assess the follow- on implications for travel and land use ( Mokhtarian, 2004). This report represents an effort to
summarize previous e- shopping research in a systematic way. Specifically, this report covers topics related to
consumer theories, dependent variables of interest, sampling approaches, product classification, methodologies,
and the determinants of e- shopping intention and adoption.
Marketing activities can occur through three types of channels: communication, transaction, and distribution.
Communication channels enable “ the exchange of information” between vendors and consumers; transaction
channels “ generate sales activities” between vendors and consumers; and distribution channels facilitate “ the
physical exchange of products and services” ( Peterson et al., 1997, p. 334). Although the internet will play a role
in all three channels, in this review, we are interested primarily in the potential for efficiency improvement and
other benefits, as well as drawbacks, that the internet offers in the transaction channel. Thus, we focus on studies
of actual or intended online purchase behavior, not just information- gathering activities.
The report is organized as follows: the next chapter summarizes the theories employed by researchers; Chapter 3
discusses the dependent variables ( the phenomena of interest typically measured); Chapter 4 presents various
approaches used by previous studies to recruit their samples; Chapter 5 highlights various product classification
systems that have been employed; Chapter 6 reviews the statistical methodologies used to describe and explain
the phenomena of interest; Chapter 7 recapitulates the statistically significant determinants of e- shopping
intention and actual adoption; and Chapter 8 summarizes the report. An appendix tabulates key facts with
respect to 65 empirical studies reviewed for this report.
3
2. Theoretical Frameworks
Theory plays a critical role in empirical research. According to the Merriam- Webster dictionary, theory refers to
“ a plausible or scientifically acceptable general principle or body of principles offered to explain phenomena”.
In deductive analysis, research hypotheses are constructed based on the available theory, and then are tested with
empirical data. Conversely, in inductive research, empirical examination further enriches and modifies theory.
Accordingly, theory and empirical analysis are interrelated and influence each other: theory guides empirical
observation and empirical observation improves theory. Over time, the interactions between theory and
empirical analysis enhance our understanding of the phenomenon of interest. In scientific research, once the
subjects of interest are determined, theory influences the research through the choice of explanatory variables,
hypothesis construction, and result interpretation.
As indicated in the Introduction, various theories have been explicitly or implicitly applied in research on
consumers’ e- shopping behavior, including the theory of reasoned action, the theory of planned behavior, the
technology acceptance model, transaction cost theory, innovation diffusion theory, and so on. The remainder of
this section will discuss each of these theories and their applications to e- shopping research in turn.
2.1 Theory of Reasoned Action ( TRA)
The TRA describes the psychological process behind conscious human behavior, and aims to explore the
determinants of that behavior ( Ajzen and Fishbein, 1980). According to the TRA, an individual’s behavioral
intention impacts the performance of her behavior, and her attitudes toward a behavior and subjective norms with
respect to the behavior compose two antecedent factors that determine her behavioral intention. Behavioral
intention measures “ how hard people are willing to try, [ and] how much of an effort they are planning to exert, in
order to perform the behavior” ( Ajzen, 1991, p. 181). Generally, individuals’ behavioral intentions have a
positive impact on the performance of the intended behavior. Attitudes toward a behavior constitute an
individual’s evaluation of the behavior. They are determined by her salient beliefs about the benefits and costs of
performing the behavior. An individual’s subjective norms are determined by her normative beliefs about the
feasibility of a behavior as evaluated by referent people, and her willingness to follow these beliefs. Further, the
TRA assumes that external factors such as an individual’s characteristics will affect her behavior only indirectly,
through their influence on the attitudes and subjective norms.
Vijayasarathy ( 2002) identified four types of salient beliefs that collectively determine an individual’s attitude
toward e- shopping: product perception, shopping experience, customer service, and consumer risk. The
individual’s normative beliefs were constructed based on her evaluation of the opinions of spouse, parents,
siblings, or friends about e- shopping. In this study, however, the author did not test the validity of the TRA, but
examined the influences of product type on the three key constructs: salient beliefs, normative beliefs, and
intention to purchase online. Cho ( 2004) and Verhoef and Langerak ( 2001) adapted the TRA to study e- shopping
behavior. Specifically, Cho assumed that attitude toward e- shopping is determined by perceived consequences
associated with e- shopping, past behavior, and attitudes toward other shopping channels, and that likelihood to
abort an intended online transaction is jointly determined by these three dimensions as well as the attitude toward
4
e- shopping. Verhoef and Langerak incorporated the constructs of innovation diffusion theory and hypothesized
that an individual’s e- shopping intention is determined by her perceptions of its relative advantage, compatibility,
and complexity. In these two studies, subjective norm was not considered as a determinant of behavioral
intention, in keeping with its less well- understood status ( Fishbein and Ajzen, 1975).
Figure 1. Theory of Reasoned Action
Source: Davis et al. ( 1989)
2.2 Theory of Planned Behavior ( TPB)
The TPB is an extension of the TRA. The major difference of the TPB from the TRA is its inclusion of perceived
behavioral control ( Ajzen, 1991). The TRA assumes that actual behavior is a motivational result of behavioral
intention, and it does not consider the influence of behavioral constraints on the link between intention and
behavior. In reality, most behavior is to some extent dependent on non- motivational factors such as availability
of resources and opportunities. For example, an individual with a high intention to engage in e- shopping may
not do so due to the lack of availability of the network or her inferior internet skills. These factors can represent
actual behavioral control; however, psychologists are more interested in the perception of behavioral control and
its influence on behavioral intention and actual behavior. Perceived behavioral control refers to an individual’s
perception of how difficult it is for her to perform a behavior ( Ajzen, 1991). As shown in Figure 2, the TPB
postulates that an individual’s behavioral performance jointly relies on and can be predicted by her behavioral
intention and perceived behavioral control.
Empirically, Hansen et al. ( 2004) applied both TRA and TPB. They found that TPB with an additional path from
subjective norm to attitude explains a higher proportion of variation in online grocery purchasing intention than
does TRA. Choi and Geistfeld ( 2004) used perceived risk and perceived self- efficacy to measure the
individual’s attitude and perceived behavioral control, respectively. Limayem et al. ( 2000) augmented the TPB
with two additional constructs: perceived consequences and perceived innovativeness. These two constructs
were assumed to influence both attitude and behavioral intention. In their models, subjective norms were
evaluated by an individual’s perception of the opinions of her family, friends, and media; and behavioral control
consisted of site accessibility, product description, transaction efficiency, navigation ability, speed, and
efficiency. Shim et al. ( 2001) adapted the TPB by incorporating the influence of past behavior and ignoring the
Attitude toward
the Behavior
Subjective Norm
Behavioral
Intention
Actual Behavior
5
impacts of subjective norms. Further, they assumed that perceived behavioral control indirectly affects intention
to shop online, through the intention to use the internet for information search.
Figure 2. Theory of Planned Behavior
Source: Ajzen ( 1991)
2.3 Technology Acceptance Model ( TAM)
TAM is another adaptation of TRA, which initially had mainly been used in explaining and predicting computer
acceptance ( Davis, 1993; Davis et al., 1989). As shown in Figure 3, the original TAM proposes two new beliefs
– perceived usefulness ( PU) and perceived ease of use ( PEOU) – and ignores the influence of subjective norm.
PU refers to an individual’s subjective evaluation of benefits ( in the initial application, specifically regarding job
performance) induced by using information technology. PEOU indicates “ the degree to which the prospective
user expects the target system to be free of effort” ( Davis et al., 1989, p. 985). PU and PEOU directly impact
attitude toward using. Additionally, PU has a direct effect on intention to use, in that one can be motivated to use
information technology to benefit one’s job performance, independently of one’s personal attitudes toward it
( Davis et al., 1989). Recently, Venkatesh and Davis ( 2000) proposed a second version of the TAM, which
incorporates additional constructs regarding social influence ( including subjective norm, voluntariness, and
image) and cognitive instrument process ( including job relevance, output quality, and result demonstrability).
Legris and his colleagues ( 2003) supported the usefulness of the TAM after reviewing a number of empirical
studies, but they pointed out that results based on the TAM are not totally consistent or clear. They recom-mended
the incorporation of factors related to human and social change processes, and the adoption of an inno-vation,
into the model.
The TAM is widely used in e- shopping research. Ahn et al. ( 2004) and O’Cass and Fenech ( 2003) directly
applied the TAM to investigate e- shopping behavior. On the other hand, most studies constructed an adapted
( simplified and/ or expanded) conceptual framework drawn from the TAM. Gefen and Straub ( 2000) proposed
that PU and PEOU directly affect intention to use and ignored the influence of the mediating variable, attitude
Perceived
Behavioral Control
Attitude toward
the Behavior
Subjective Norm
Behavioral
Intention
Actual Behavior
6
toward using. Adopting the same simplification, Liu and Wei ( 2003) additionally proposed that perceived risk is
an antecedent factor of intention to use. Henderson and Divett ( 2003) tested direct links from PU and PEOU to
actual use of e- shopping. Shang et al. ( 2005) adapted the TAM in two ways: adopting the same simplification as
Henderson and Divett, and adding the influences of two constructs ( cognitive absorption and fashion
involvement) on actual behavior.
Figure 3. Technology Acceptance Model
Source: Davis et al. ( 1989)
By contrast, Childers et al. ( 2001) and Koufaris ( 2002) augmented the TAM with the enjoyment of e- shopping, a
positive predictor of attitude toward using. Chen et al. ( 2002) incorporated innovation diffusion theory in the
TAM. Specifically, they assumed that the compatibility between using electronic stores ( e- stores) and an
individual’s belief and need affects her attitude toward using. Chen and Tan ( 2004) further expanded this model
by adding two more links, from perceived trust and perceived service quality to attitude toward using. Van der
Heijden et al. ( 2003) hypothesized that attitude toward using is determined by trust in e- stores and perceived risk
in addition to PU and PEOU. Shih ( 2004) proposed a more complex modeling structure based on the TAM.
2.4 Transaction Cost Theory ( TCT)
Williamson ( 1985) defines a transaction as a process by which a good or service is transferred across a
technologically separable interface. “ In classical economic theory, it is assumed that information is symmetric in
the market. Since both buyers and sellers are assumed to have the same amount of information, the transaction
can be executed without cost. In reality, however, markets are often inefficient. In order to proceed with a
transaction, consumers must conduct activities such as searching for information, negotiating terms, and
monitoring the on- going process to ensure a favorable deal. The costs involved with such transaction- related
activities are called transaction costs” ( Liang and Huang, 1998, p. 31).
TCT can explain various problems of economic organizations ( Rindfleisch and Heide, 1997). Its basic principle
is that individuals would like to conduct transactions in the most efficient way ( Williamson, 1985). That is, the
lower the transaction costs, the more likely individuals are to conduct the transaction. Transaction costs are
Perceived
Usefulness
Perceived Ease
of Use
Attitude toward
Using
Behavioral
Intention to Use
Actual
System Use
7
determined by several constructs, including uncertainty and asset specificity as shown in Figure 4. Since
information in the market is always asymmetric, the outcomes of a transaction may not follow, or may even be
contrary to expectations, leading to uncertainty. Transactions are encouraged through reducing uncertainty, as
one form of lowering the transaction costs. Asset specificity refers to the lack of transferability of the assets from
one transaction to the other. Since “ assets with a high amount of specificity represent sunk costs that have little
value outside of a particular exchange relationship” ( Rindfleisch and Heide, 1997, p. 41), higher asset specificity
is associated with lower transaction costs for the exchange relationship to which the specificity applies, and
higher transaction costs for other exchange relationships.
Figure 4. Transaction Cost Theory
Source: Liang and Huang ( 1998)
Liang and Huang ( 1998) applied TCT to investigate consumers’ intention to shop online. In addition to the
modeling structure presented in Figure 4, they further assumed that e- shopping intention is also directly
influenced by uncertainty and asset specificity. Teo and Yu ( 2005) proposed that buying frequency is also a
( negatively- associated) predictor of transaction costs, with trust replacing asset specificity.
2.5 Innovation Diffusion Theory ( IDT)
Compared to traditional shopping, e- shopping is an innovative application of information technology by retail
industries. Therefore, IDT can be applied to explore consumers’ e- shopping behavior. Generally, the cumulative
adoption of an innovation follows a sigmoid curve, with adoption growing slowly in its initial years, growing
steeply as it reaches its half- way point, and growing slowly again as it nears its saturation level ( maximum
penetration). The rate of adoption is mainly dependent on five attributes of an innovation: relative advantage
( the extent to which an innovation is perceived to be better than the one it substitutes for or competes with),
compatibility ( the extent to which an innovation is perceived to be consistent with the experiences and
requirements of potential adopters), complexity ( the extent to which an innovation is perceived to be difficult to
use), trialability ( the extent to which an innovation can be experimented with on a limited basis) and
observability ( the extent to which the utility of an innovation is visible to the public). Relative advantage,
compatibility, trialability, and observability of an innovation are found to be positively related to its rate of
adoption, while complexity is negatively associated with its rate of adoption ( Rogers, 1983). Dearing et al.
( 1994) further suggested that applicability and reliability are important for diffusion of risky innovations.
Uncertainty
Asset Specificity
Transaction
Costs
Transaction
Intention
8
Innovation diffusion models are commonly used in aggregate studies although there are some recent efforts at
the disaggregate level ( Roberts and Lattin, 2000). In the context of e- shopping, most studies consider the
influences of the constructs derived from IDT on disaggregate e- shopping behavior, rather than model the overall
diffusion of e- shopping at an aggregate level. As mentioned earlier, Verhoef and Langerak ( 2001) explored the
impacts of relative advantage, compatibility, and complexity of e- shopping; and Chen et al. ( 2002) and Chen and
Tan ( 2004) introduced the compatibility of e- shopping into their TAM. Finally, Eastin ( 2002) examined the
influences of compatibility, reliability, complexity, and relative advantage on the frequency of e- shopping.
2.6 Other theories
In addition to the theories discussed above, various constructs from other theories have been introduced to
explore the determinants of consumers’ e- shopping behavior. Grazioli and Jarvenpaa ( 2000) investigated the
role of perceived internet deception ( from social exchange theory) in the adoption of e- shopping. Swaminathan
et al. ( 1999) viewed e- shopping as a social influence process and examined the influences of the social
characteristics of actors, namely, e- shopping vendors and consumers. Li et al. ( 1999) applied channel theory and
studied the impacts of retail channel knowledge and utilities. Foucault and Scheufele ( 2002) applied social
influence theory, social learning theory, and use and gratifications theory in their empirical work.
Mathwick et al. ( 2001) applied experiential value theory to the preference for and intention to use a certain
retailer’s internet site ( and, separately, catalog). This theory suggests that an activity such as shopping can offer
intrinsic as well as extrinsic benefits, and may involve an active ( participatory) or reactive ( passive) role of the
individual. The combination of these two dimensions yields ( p. 41) “ a value landscape divided into four
quadrants framed by intrinsic/ extrinsic sources of value on one axis and active/ reactive value on the other”, as
shown in Figure 5.
Figure 5. Typology of Experiential Value
Intrinsic Value Playfulness:
• intrinsic enjoyment
• escapism
Aesthetics:
• visual appeal
• entertainment
Extrinsic Value Consumer Return on Investment ( CROI):
• economic utility
• efficiency
Service excellence
Active Value Reactive Value
Source: Adapted from Mathwick et al. ( 2001)
2.7 Summary
Various theories have been applied in previous research to explain and predict consumers’ e- shopping behavior.
Overall, more attention has been paid to theories in the social psychology area: TRA, TPB, and TAM. The
empirical studies based on these theories highlight the importance of consumers’ attitudes, since attitudinal
factors explain most of the variation in e- shopping behavior. However, the explanatory power of these theories
varies from one study to another.
9
E- shopping behavior is a complicated decision process. First, consumers make a shopping decision based on
their family needs, budget limitations, and other constraints impinging on them. Accordingly, they are likely to
minimize transaction costs and maximize compatibility with needs. Second, e- shopping behavior is a social
influence process and it is affected by social influence ( e. g., social norms), vendor and consumer characteristics,
and third parties ( e. g., competitive offerings) ( Bagozzi, 1974). Third, e- shopping can be viewed as an innovation
and its adoption is impacted by its intrinsic attributes as well as by mass media and word of mouth ( Mahajan et
al., 1990). When multiple retail channels are present for the same transaction, the adoption of e- shopping can
become a substitution for traditional retail channels; alternatively, it may complement or supplement these
channels. Therefore, no single theory appears capable, as is, of capturing the complexities of e- shopping
behavior: a comprehensive integration of several theories becomes necessary. For example, Konana and
Balasubramanian ( 2005) incorporate elements of TRA, TPB, and TAM as well as economic factors ( perceived
financial gains) into their model of online investing.
In the current context, we suggest, as an initial approach, that integration across theories can be viewed in terms
of the utility maximization theory of economics, in which the utility of an alternative is positively associated
with its benefits and negatively related to its costs. The evaluation, weighting, and way of combining benefits
and costs ( or, advantages and disadvantages, pros and cons) may vary by individual, but each individual is
assumed to balance the benefits of each alternative against its costs, and choose the one with the highest net
benefits, or utility. To briefly place each of the theories discussed in Sections 2.1- 2.5 in this context, we can note
that:
( 1) In the TRA, both attitudes and subjective norms can be positive with respect to the behavior ( benefits) or
negative ( costs). For example, in Cho ( 2004), an inability to physically examine the good and concerns
over delivery and return can be viewed as liabilities or disadvantages of e- shopping, while the saving of
effort it makes possible can be viewed as a benefit.
( 2) Similarly, in the TPB, measures of perceived behavioral control ( such as the site accessibility and
transaction efficiency of Limayem et al., 2000) represent costs or benefits of adoption, depending on
whether the measure is positively or negatively oriented, and whether a shopping alternative rates high or
low on it. Cost measures can in some cases represent prohibitively high costs or outright constraints, and in
other cases may fall on a continuum constituting gradually stronger disincentives to adopt. The converse is
true for benefit measures.
( 3) In the TAM, usefulness and ease of use are two categories of benefits ( or, if negatively oriented, costs)
which, broadly construed, can contain most or all of the constructs of the other theories. For example, a
negative subjective norm can be viewed as reducing the ease of use of an alternative.
( 4) In TCT, uncertainty is a cost and asset specificity is a benefit with respect to the asset in question.
( 5) In IDT, as implied by their descriptions, relative advantage, compatibility, trialability, and observability are
benefits while complexity is a cost.
Thus, a first- cut synthesis of these theories can be achieved through the development of a comprehensive list of
benefits and costs, using each of the key constructs of the pertinent theories as a guide to identifying the nature of
those benefits and costs.
10
3. Dependent Variables of Interest
Various dimensions of consumers’ e- shopping behavior have been explored in previous studies. Generally, these
dimensions can be classified into two categories: behavioral intention and actual behavior. Some studies treated
e- shopping intention as the dependent variable ( e. g., Choi and Geitsfeld, 2004); some studies chose actual
e- shopping behavior as the dependent variable ( e. g., Eastin, 2002); and a few studies examined the determinants
of multiple dependent variables including both behavioral intention and actual behavior ( e. g., Grazioli and
Jarvenpaa, 2000; Liang and Lai, 2002). In addition, several studies considered attitudes toward e- shopping as
the ultimate dependent variable ( e. g., Childers et al., 2001; Vijayasarathy and Jones, 2000). The remainder of
this chapter will provide an overview of measures of e- shopping intention and actual e- shopping behavior.
3.1 E- shopping intention
As discussed previously, behavioral intention measures how hard an individual is intended to try, or the strength
of intended effort to perform a behavior ( Ajzen, 1991). Behavioral intention does not perfectly correlate with the
actual behavior. An individual may engage in a less intended choice due to the presence of constraints
( behavioral control factors). In addition, of course, stated intentions often differ from true intentions due to
social desirability bias ( the tendency to provide the response that is socially expected) or consistency bias ( the
need to appear consistent to the analyst, which can result in stated intentions that are consonant with
previously- expressed attitudes, when in reality one’s actual behavior will be dissonant from those attitudes).
Generally, however, behavioral intention tends to have a positive association with the actual choice of that
behavior. That is, the stronger an individual’s behavioral intention, the more likely she is to perform the
behavior.
E- shopping intention can measure an individual’s conative beliefs with respect to adopting or using e- shopping
( e. g., Belanger et al., 2002), as well as with respect to planning to abort an online transaction before checkout
( Cho, 2004). Previous studies have adopted various scales to measure respondents’ e- shopping intention: a
two- point scale and multi- point qualitative scales. Among the latter, five- point scales and seven- point scales
have been most commonly used. Further, e- shopping intention has been constructed in different ways. In some
studies, a single question was used to ask respondents to report their e- shopping intentions ( e. g., Chen and Tan,
2004, Goldsmith, 2002; Koufaris, 2002; Liang and Huang, 1998; Liao and Cheung, 2001; Phau and Poon, 2000;
van den Poel and Leunis, 1999; Verhoef and Langerak, 2001). In most studies, however, e- shopping intention
was evaluated by a latent construct. That is, several dimensions related to e- shopping intention were
individually measured in the survey and then integrated into a scalar intention, generally through factor analysis.
Table 1 presents an overview of these dimensions. Clearly, the intention to purchase products from online retail
sites is the most direct indicator of e- shopping intention. This purchasing intention was expressed in various
ways such as likelihood, probability, expectation and so on, and was measured at different time points such as
currently, at the next visit, or in the future. In some studies, e- shopping intention was also assessed by an
integration of near- term and long- term intentions ( e. g., Choi and Geistfeld, 2004; Jarvenpaa et al., 2000; van der
Heijden et al., 2003; van der Heijden and Verhagen, 2004).
11
Chen et al. ( 2002) and Limayem et al. ( 2000) theoretically considered e- shopping intention as an antecedent
factor of actual e- shopping behavior, which in principle is the expected direction of causality. Limayem et al.
( 2000) employed a longitudinal survey, the first survey measuring respondents’ attitudes, online purchasing
intention, and so on, and the second survey asking for the number of online purchases they made since the first
survey. However, in Chen et al. ( 2002), an individual’s e- shopping intention was measured at the current time
point, while her previous e- shopping behavior was used as the dependent variable. This model inverted the time
precedence between intention and actual behavior and thereby violated temporal consistency. Since individuals’
intentions may be influenced by their prior e- shopping experience and/ or change over time ( Liang and Huang,
1998, Vijayasarathy, 2002), the validity of such model results is in great doubt. More reasonably, some studies
used previous e- shopping behavior to explain consumers’ current intention of purchasing over the internet
( Goldsmith, 2002; van den Poel and Buckinx, 2005).
Table 1. Dimensions of E- shopping Intention
Intention Dimensions References
Purchasing from e- shopping sites Ahn et al. ( 2004); Belanger et al. ( 2002);
Foucault and Scheufele ( 2002); Gefen ( 2000);
Gefen and Straub ( 2000, 2004); Grazioli and Jarvenpaa ( 2000);
Hansen et al. ( 2004); Huang ( 2000); Jarvenpaa et al. ( 2000);
Liang and Huang ( 1998); Liang and Lai ( 2002);
Liao and Cheung ( 2001); Limayem et al. ( 2000); Liu et al. ( 2004);
Liu and Wei ( 2003); Phau and Poon ( 2000);
Ranganathan and Ganapathy ( 2002); Shih ( 2004);
Shim et al. ( 2001); Teo and Yeong ( 2003); Teo and Yu ( 2005);
van den Poel and Buckinx ( 2005); van den Poel and Leunis ( 1999);
van der Heijden et al. ( 2003); van der Heijden and Verhagen ( 2004);
Vijayasarathy ( 2002); Verhoef and Langerak ( 2001).
Using e- shopping sites Ahn et al. ( 2004); Chen et al. ( 2002); Chen and Tan ( 2004);
Mathwick et al. ( 2001).
Recommending others to use e- shopping
sites
Ahn et al. ( 2004); Grazioli and Jarvenpaa ( 2000);
Liu et al. ( 2004).
Using credit card on e- shopping sites or
paying online
Belanger et al. ( 2002); Gefen ( 2000);
Gefen and Straub ( 2000, 2004); Shih ( 2004).
Returning to e- shopping sites Grazioli and Jarvenpaa ( 2000); Jarvenpaa et al. ( 2000);
Koufaris ( 2002); Liang and Lai ( 2002); Liu et al. ( 2004); Mathwick
et al. ( 2001); van der Heijden et al. ( 2003); van der Heijden and
Verhagen ( 2004).
Creating an account on e- shopping sites Belanger et al. ( 2002).
Providing information to e- shopping sites Gefen and Straub ( 2000).
Positive remarks on e- shopping sites Liu et al. ( 2004).
Willingness to pay more online Srinivasan et al. ( 2002).
Collecting information using the internet Vijayasarathy ( 2002).
3.2 Actual e- shopping behavior
In previous studies, e- shopping behavior was defined in different ways. Most studies considered only purchase
behavior. On the other hand, some studies treated both purchases and information gathering as e- shopping
behavior ( e. g., Chen et al., 2002). The measures of actual e- shopping behavior mainly include three dimensions:
12
the adoption of e- shopping, the amount spent online, and the frequency of using e- shopping, as shown in Table 2.
The measurement periods span from a few months to several years, or even longer ( for example: have you
ever … ?). The adoption of e- shopping was measured as a binary variable, yes or no. E- shopping frequency was
assessed on either a numerical scale or an ordinal scale. For example, Limayem ( 2002) asked respondents to
report the number of purchases they made over the internet; Eastin ( 2002) measured the frequency of e- shopping
on a six- point ordinal scale: never, at least once, once a month, several times per month, once a week, and daily.
In most studies, these three dimensions were investigated separately. By contrast, several studies adopted a
latent construct of e- shopping behavior generated by factor analysis. Specifically, Chen et al. ( 2002) used two
items regarding e- shopping frequency to construct a behavior scale; Kim et al. ( 2000) extracted actual
e- shopping behavior from four variables measuring e- shopping spending and frequency; and Corbitt et al. ( 2003)
integrated four dimensions of e- shopping behavior: frequency, spending, the ratio of online purchases to total
purchases in value, and the intention to increase/ decrease purchasing in the future.
Table 2. Dimensions of E- shopping Behavior
Behavior Dimensions References
E- shopping adoption Bellman et al. ( 1999); Bhatnagar et al. ( 2000); Donthu and Garcia ( 1999);
Foucault and Scheufele ( 2002); Goldsmith and Goldsmith ( 2002);
Grazioli and Jarvenpaa ( 2000); Koufaris ( 2002); Koyuncu and Bhattacharya ( 2004);
Liang and Lai ( 2002); Lohse et al. ( 2000); O’Cass and Fenech ( 2003);
Shang et al. ( 2005); Sim and Koi ( 2002); Sin and Tse ( 2002).
E- shopping spending Bellman et al. ( 1999); Corbitt et al. ( 2003); Lohse et al. ( 2000);
Forsythe and Shi ( 2003); Henderson and Divett ( 2003); Kim et al. ( 2000);
Lunn and Suman ( 2002); Shang et al. ( 2005); Swaminathan et al. ( 1999).
E- shopping frequency Blake et al. ( 2003) ; Chen et al. ( 2002); Corbitt et al. ( 2003); Eastin ( 2002);
Forsythe and Shi ( 2003); Henderson and Divett ( 2003); Kim et al. ( 2000);
Koyuncu and Bhattacharya ( 2004); Li et al. ( 1999); Limayem et al. ( 2000);
Lunn and Suman ( 2002); Miyazaki and Fernandez ( 2001); Shang et al. ( 2005);
Swaminathan et al. ( 1999).
3.3 Summary
The dependent variables used in e- shopping behavior research mainly relate to e- shopping intention and actual
e- shopping behavior. E- shopping intention was measured by various dimensions. Among them, the intention to
purchase online is the most frequently used measurement. Although some studies used a unidimensional
measurement, most adopted a latent construct to assess consumers’ e- shopping intentions. Actual e- shopping
behavior mainly includes three dimensions: adoption, spending, and frequency. Most studies examined one or
more of these three dimensions directly, while a few studies constructed a latent variable to measure actual
e- shopping behavior.
In most studies, either e- shopping intention or actual behavior was chosen as the dependent variable of interest.
A few studies investigated both jointly. When behavioral intention and actual behavior are simultaneously
included in an empirical study, attention should be paid to the time precedence between intention and behavior.
13
4. Sampling Approaches
Distributing surveys randomly to members of the general population is a traditional sampling approach. If truly
random, such a sample can be a good representation of the population. However, with respect to an e- shopping
survey, non- response bias may be a major concern with the traditional approach. For example, those who do not
have access to computers or are unaware of e- shopping may be less likely to respond, leading to a higher
representation of existing online consumers in the sample than in the population as a whole. Even so, since only
a minority of the general population purchases online, such a sample may not contain a substantial number of
e- shopping occasions. Accordingly, the lack of e- shopping observations in the sample tends to limit the
statistical robustness of research results. Therefore, from the standpoint of efficiency, a choice- based sampling
approach is probably more appropriate for e- shopping surveys, at least with the current state of adoption and use.
Among the studies reviewed here, 34 used data collected from computer/ internet/ email users ( e. g., Shim et al.,
2001; Swaminathan et al., 1999; van den Poel and Leunis, 1999). Obviously, online consumers must be
computer and internet users, and are probably email users. Therefore, understanding these users’ e- shopping
behavior is somewhat equivalent to understanding the e- shopping behavior of the general population, aside from
descriptive statistics on the prevalence of such behavior in the population. Also, this sampling approach has
several advantages. First, compared to the general population, these users are more likely to be aware of
e- shopping even if they do not purchase online. Thus, they are likely to be more representative of potential
near- term future buyers. Second, internet users are essentially the only ones who have the choice to shop online;
it would produce statistically inconsistent estimates ( Thill, 1992) to include those who do not have a choice in an
e- shopping adoption model ( on the other hand, current non- users can provide important insight into intentions,
and more generally are essential to understanding the choice to become a computer/ internet/ email user, as a
necessary precursor to e- shopping adoption). Third, by excluding non- users, this sampling approach is more
likely to capture an adequate number of e- shopping observations and will achieve a better balance between
internet buyers and non- buyers. Finally, e- vendors’ marketing policies mainly target internet users rather than
the whole population. Therefore, understanding the e- shopping behavior of internet users will provide e- vendors
direct information on how to improve site design and marketing strategies.
When the survey subjects were internet users, some studies used the internet as the data collection tool, which
acted as a natural filter to exclude ineligible respondents. These studies adopted different channels to recruit
their samples. Most studies used email to announce their surveys. Specifically, some researchers sent an
invitation email to a mailing list, with a URL link to the survey web site. The mailing list was obtained in various
ways. Chen and Tan ( 2004) distributed their emails to 1,865 registered users of a non- profit organization and
three e- shopping- related news groups. Limayem et al. ( 2000) chose 6,110 email users from four internet- based
directories. O’Cass and Fenech ( 2003) recruited their sample through a major Australia internet service provider.
Shang et al. ( 2005) obtained their mailing list from a major computer magazine in Taiwan. Van den Poel and
Leunis ( 1999) collected email addresses from the Belgian White Pages. Eastin ( 2002) harvested more than
1,300 email addresses from 50 e- commerce- related news groups ( randomly selected from the population of 127
e- commerce news groups in a USENET organization). To increase response rate, he explicitly removed business
accounts and sent two emails to recipients: one introduced the background and objective of the survey and the
14
other contained a link to the internet- based survey. Li et al. ( 1999) drew their sample from about 50,000 internet
users registered in the company’s CyberleagueTM online panel. After an invitation email was distributed to these
users, they sent two email reminders at five- day intervals to those who did not respond.
Moreover, several studies adopted different announcement strategies. For example, Ahn et al. ( 2004) posted the
banner of their survey on e- shopping sites and provided a URL link to the survey site; and Teo and Yeong ( 2003)
announced the survey through local newsgroups, URL links from faculty homepages, and personal emails. On
the other hand, van den Poel and Leunis ( 1999) argued that an internet- based survey is expected to be more prone
to self- selection bias than one requiring little effort to access, and hence directly distributed the surveys to
recipients by email.
Some studies extracted their analysis sample from available databases. Bellman et al. ( 1999) and Lohse et al.
( 2000) used data from the Wharton Virtual Test Market ( WVTM), which were collected to understand the
demographics and attitudes of online consumers, and determinants of e- shopping behavior. As stated in Bellman
et al. ( 1999, p. 32), “ the Wharton Forum on Electronic Commerce began recruiting a panel of Web users from
around the world in October 1997. As part of the participant registration process, 10,180 people completed a
survey asking 62 questions about online behavior and attitudes about Internet communication and privacy issues,
as well as routine demographic questions. People were attracted to the survey site ( wvtm. wharton. uppen. edu)
through an online banner advertising campaign targeting specific segments of Web users worldwide, links
provided by members’ sites, and word of mouth. Survey respondents self- selected themselves as panel members,
and their answers to the survey questions were self- reported.”
Some studies analyzed data from the Graphic, Visualization, and Usability ( GVU) Center of the Georgia
Institute of Technology ( e. g., Bhatnagar et al. 2000; Forsythe and Shi, 2003). “ Since its beginning in 1994, the
GVU WWW User Survey has accumulated a unique store of historical and up- to- date information on the growth
and trends in Internet usage. It is valued as an independent, objective view of developing Web demographics,
culture, user attitudes, and usage patterns. Recently the focus of the Survey has been expanded to include
commercial uses of the Web, including advertising, electronic commerce, intranet Web usage, and
business- to- business transactions” ( http:// www. cc. gatech. edu/ gvu/ user_ surveys, accessed on March 4, 2005).
The GVU has conducted 10 surveys up to 1998. Internet shopping is one component of these surveys. The
population for the GVU surveys is internet users. In addition to the seventh GVU WWW User Survey, Hoffman
et al. ( 1999) also analyzed the data from the 1997 Nielsen Media Research and CommerceNet Internet
Demographics Study. The Nielson study focused on U. S. web users and studied the use of the internet for
shopping and purchasing.
In addition, some studies collected data from known online consumers, for the purpose of examining their future
e- shopping intentions. Ranganathan and Ganapathy ( 2002) collected data from 214 individuals in Illinois, who
had completed at least one online purchase in the last six months. Srinivasan et al. ( 2002) sent an email
invitation to 5,000 consumers randomly selected from a list of online consumers compiled by a market research
firm. Koufaris ( 2002) simulated actual consumers by first asking the respondents to visit a specific online
15
bookstore ( booksamillion. com) and then asking them to complete the questionnaires. This sampling approach
leads to a more homogeneous sample.
Finally, 22 studies chose students ( either undergraduates or graduates) as their subjects ( e. g., Belanger et al.,
2002; Choi and Geistfeld, 2004). The student sample is often criticized due to its higher- than- average proportion
of younger adults. However, most studies did not discuss the acceptability and appropriateness of using students
in this area of research. Actually, a student sample has some advantages for an e- shopping study, since those
respondents will tend to be harbingers of future adoption patterns in the population at large. That is, the
e- shopping behavior of today’s over- 50 adults is apt to tell us less about the future than that of today’s
20- somethings. Further, compared to traditional consumers, online consumers tend to be young and
well- educated ( OECD, 1999). Therefore, the profile of a student sample is in some ways closer to that of the
online consumer population than is the profile of the general population more broadly. Further, if books are one
product type chosen for detailed study, a student sample is a natural group of consumers to analyze. However,
these studies have limited generalizability of parameter estimates to a larger population because of their
homogeneous samples. For example, cost- conscious student consumers may weight price more heavily in their
purchase decisions than would the population of online shoppers as a whole. They may also be more risk- taking,
more innovative, and more trusting than their elders.
16
5. Product Classifications
The characteristics of products greatly determine the degree to which they are suitable for marketing online
( Peterson et al., 1997). That is, these characteristics are important factors that influence consumers’ internet
purchases of specific products. Accordingly, to better understand consumers’ e- shopping behavior, we must
differentiate various types of products.
A number of different classifications have been proposed in the literature. In one early study, Copeland ( 1924)
grouped products into three categories: convenience goods, shopping goods, and specialty goods. Convenience
goods refer to “ goods which the customer usually purchases frequently, immediately, and with the minimum of
effort”, for example, newspapers and many grocery products ( AMA, 1948, p. 206). Shopping goods are “ goods
which the customer in the process of selection and purchase characteristically compares on such bases as
suitab[ ilit] y, quality, price, and style”, such as furniture and apparel ( AMA, 1948, p. 215). Specialty goods refer
to “ goods on which a significant group of buyers characteristically insists and for which they are willing to make
a special purchasing effort”, for example, specific brands of gourmet foods and computers ( AMA, 1948, p. 215).
Online consumers’ satisfaction with delivery and post- delivery services differs along these three product types.
Specifically, Thirumalai and Sinha ( 2005) found that, on average, convenience and shopping goods consumers
tend to be more satisfied with product order fulfillment process than specialty goods consumers. However, this
classification system is less able to evaluate the online- purchasing suitability of products.
On the other hand, Nelson ( 1970) distinguished two types of qualities of a product: search qualities and
experience qualities. Search qualities refer to the traits that consumers can fully ascertain prior to use, while
experience qualities indicate those that cannot be determined until using the product. A particular product may
possess both search qualities and experience qualities. If search qualities dominate the attributes of a product,
such as software, it is called a search good. Conversely, if experience qualities of a product, such as apparel,
outweigh its search qualities, it is called an experience good. Consumers can identify the important attributes of
a search good based on an information search over the internet. In such cases, if an e- shopping transaction is at
least as efficient and cost- effective as traditional store shopping, the likelihood of conducting an online
transaction will greatly increase. On the other hand, if a product is an experience good, the available information
may not be sufficient for consumers to engage in an online transaction. They may have to evaluate its attributes
by physical inspection or personal trial, and hence the probability of adopting e- shopping is likely to decrease.
Of course, neither the character of the product nor the choices facing the consumer are purely binary; various
hybrid processes are possible, such as using online searches and in- store trials to narrow the choice set and then
making the final purchase based on price information obtained over the internet. Bhatnagar and Ghose ( 2004)
comment that in the context of the internet, some goods that are traditionally search goods become experience
goods, and conversely.
Although Nelson’s classification is useful, Peterson et al. ( 1997, pp. 335- 336) proposed a more detailed
classification system
“ in which products and services are categorized along three dimensions that are more relevant in the context
of the Internet: cost and frequency of purchase, value proposition, and degree of differentiation. Goods
17
vary along the first dimension from low- cost, frequently purchased goods ( e. g., consumable products such
as milk) to high- cost, infrequently purchased goods ( e. g., durable products such as stereo systems). … In
general, when purchase fulfillment requires physical delivery, the more frequent the purchase and the
smaller the cost ( e. g., milk), the less likely there is to be a good ‘ fit’ between a product or service and
Internet- based marketing.
“ Goods vary along the second dimension according to their value proposition, whether they are
tangible and physical or intangible and service related. Internet- related marketing is particularly well suited
to certain types of intangible or service- related goods ( i. e., those with a digital basis). To the extent that the
value proposition is intangible, the greater the frequency of purchase or use of a good, the greater the
advantage of the Internet as a transaction and distribution channel.
“ The third dimension reflects the degree to which a product or service is differentiable. In particular, it
reflects the extent to which a seller is able to create a sustainable competitive advantage through product
and service differentiation. Internet- based marketing can result in extreme price competition when
products or services are incapable of significant differentiation. … When products or services are capable of
significant differentiation, the Internet can serve as an effective segmentation mechanism for guiding
consumers to their ideal product or service”.
As shown in Table 3, they presented an example classification of products or services along the three dimensions,
dichotomized for simplicity of exposition.
Table 3. Product and Service Classification
Cost & Frequency Value Proposition Differentiation Examples
Tangible or Potential high Wines, soft drinks, cigarettes
physical Potential low Milk, eggs
Potential high Online newspapers and magazines
Low cost,
frequently
purchased goods Intangible or
informational Potential low Stock market quotes
Tangible or Potential high Stereo systems, automobiles
physical Potential low Precious metal ingot of known weight & purity
Potential high Software packages
High cost,
infrequently
purchased goods Intangible or
informational Potential low Automobile financing, insurance
Source: Peterson et al. ( 1997).
Once a product category is determined, consumers need to decide which brand to choose, from which channel( s)
to gather product information, and over which channel to complete the purchase ( Peterson et al., 1997). ( Note
that although Peterson et al. characterized the choice of information- gathering channel as either- or, in reality
both the internet and store or other conventional channels can be used). Combined with the classification system,
Peterson et al. ( 1997) proposed some logical consumer decision sequences. As shown in Table 4, low- cost,
frequently- purchased, and tangible or physical products are more likely to be purchased in traditional retail
stores. Conversely, low- cost, frequently- purchased, and intangible or informational products are more likely to
be purchased over the internet. And both channels are suitable for high- cost, infrequently purchased products.
Phau and Poon ( 2000) applied this classification system in an empirical study. They found that product and
service type classification significantly influence the consumer choice between a retail store and an e- shopping
18
mall. In general, products and services that have a low cost, are frequently purchased, have intangible value
proposition, and/ or are relatively high on differentiation are more amenable to be purchased over the internet.
Specifically, flowers, videos/ music, paid subscription to online newspapers and financial reports, computer
software, and consultancy services were found to be more likely to be purchased over the internet; while milk,
eggs, vegetables, car loans, and insurance were found to be more likely to be purchased in traditional retail
stores.
Table 4. Likely Consumer Decision Sequences.
Cost & Frequency Value Proposition Differentiation Likely Decision Sequences*
Potential high
Brand choice likely after retail search.
Price search on the internet is unlikely.
Tangible or Final acquisition likely in retail store.
physical
Potential low
Brand choice likely after retail search.
Price search on the internet is unlikely.
Final acquisition likely in retail store.
Potential high
Brand choice likely after internet search.
Price search in retail channels is unlikely.
Final acquisition likely on the internet.
Low outlay,
frequently
purchased goods
Intangible or
informational
Potential low
Brand choice likely after internet search.
Price search in retail channels is unlikely.
Final acquisition likely on the internet.
Potential high
Brand choice likely after search of both
channels.
Price search likely in both channels.
Final acquisition may occur in either channel
( the need for personal product inspection may
strongly influence the decision in this case).
Tangible or
physical
Potential low
Brand choice likely after search of both
channels.
Price search likely in both channels.
Final acquisition may occur in either channel.
Potential high
Brand choice likely after search of both
channels.
Price search likely in both channels.
Final acquisition may occur in either channel
( if prices are comparable, the internet may be
convenient for final delivery of such products).
High outlay,
infrequently
purchased goods
Intangible or
informational
Potential low
Brand choice likely after search of both
channels.
Price search likely in both channels.
Final acquisition may occur in either channel.
Source: Peterson et al. ( 1997).
* The authors considered only two shopping media ( the internet and traditional retail store) and three decision
dimensions ( information search leading to brand choice, price search, and final acquisition).
Based on cost outlay and tangibility, Vijayasarathy ( 2002) classified four types of products, as shown in Table 5.
He found that consumers’ salient beliefs and normative beliefs about e- shopping were significantly different
19
between tangible and intangible products, and that a product’s tangibility had a significant influence on
consumers’ intentions toward e- shopping. In particular, consumers’ intentions to shop online for intangible
products were higher than their intentions to shop for tangible products. However, cost and the interaction
between cost and tangibility were not significantly associated with consumers’ intentions.
Table 5. Product Type Description
Tangible Intangible
Low cost A physical product that is relatively inexpensive
( less than $ 50). Examples: groceries, clothing,
toys, health and beauty products, pet supplies,
sporting goods and garden supplies
Non- physical items and services that are
relatively inexpensive ( less than $ 50).
Examples: computer software, music, movie/
concert/ theater/ sporting event tickets, online
banking and online brokerage/ trading services
High cost A physical product that is relatively expensive
( greater than $ 300). Examples: computer
hardware, household appliances, furniture, and
consumer electronics
Non- physical items and services that are rela-tively
expensive ( greater than $ 300). Examples:
mortgage or automobile financing, insurance
for real estate/ auto/ life, vacation/ travel plan-ning,
and airline tickets
Source: Vijayasarathy ( 2002).
While the classification system of Peterson et al. is clearly an improvement over that of Nelson in some ways, the
dimensions of search and experience should not be neglected. The experience dimension of Nelson, for example,
is synonymous with neither the value proposition nor the differentiation dimensions of Peterson et al. Consider
soft drinks: while there is clearly differentiation between soft drink brands, there is virtually no differentiation
within a brand. A specific soft drink not only need not be experienced before being purchased, it cannot be
experienced beforehand in the same way that a specific garment can be tried on before a purchase decision is
made ( of course that brand of soft drink can be experienced before making a later purchase of the same brand.
The fact that the previous experience is completely transferable to the present purchase context is precisely the
point). Similarly, the content of a specific book is essentially identical from one platform to the next, and so
( with some obvious exceptions, such as collector’s items, preferences for hardback versus paperback,
accompanying information such as commentary or background) once having determined to buy a certain book,
the need to experience a particular version of that book before purchase is minimal. On the other hand, consider
wine: while there is clearly variation by vintage within brand, many wine purchases are made without first
having tasted the specific brand- vintage combination being purchased; in fact the opportunity to do such tasting
is generally rather limited. Combining these observations, we suggest that the need/ ability to experience the
specific purchase in advance is a relevant product classification dimension. All else equal, products with either a
low need or a low ability to be tried in advance are more likely to be purchased over the internet. Empirically,
Cho ( 2004) found that concern for the risks involved in the inability to physically examine products increases the
likelihood of aborting an online transaction.
For different products, the internet shows diverse suitability as a shopping medium. Therefore, mixing product
categories in e- shopping behavior research tends to yield vague or inconsistent results. For example, a consumer
may be more likely to purchase software online, but less inclined to acquire clothing over the internet; her overall
e- shopping intention will be some unknown mixture of the high intention and the low intention. Further, when
20
respondents are asked to report their overall e- shopping intention without specifying any product categories,
they might overstate or understate their intention because one or several products they considered may be more
or less suitable for e- shopping. Many studies measured intention in this way ( e. g., Belanger et al., 2002; Huang,
2000; Ranganathan and Ganapathy, 2002).
On the other hand, some studies chose particular product categories with which to study e- shopping behavior.
This specification reduces the heterogeneity resulting from different product types. Books/ textbooks are the
most common choice of online products in these studies ( e. g., Foucault and Scheufele, 2002; Gefen, 2000, Gefen
and Straub, 2000 & 2004; Liu et al., 2004). Apparel and used laptops were chosen at least once each ( Goldsmith
and Goldsmith, 2002; Grazioli and Jarvenpaa, 2000). Shim et al. ( 2001) focused on search products ( books,
software, and videos) to investigate the influences of intention to search on e- shopping intention. Bhatnagar et al.
( 2000) studied the adoption of 14 types of online products and services. Hansen et al. ( 2004) and Henderson and
Divett ( 2003) respectively investigated the intention and adoption of e- shopping for groceries.
In conclusion, consumers’ decisions to participate in e- shopping activities differ by product type. However,
relatively little effort has been invested into product classification in the context of e- shopping. As a result,
confounding suitable products and unsuitable products in e- shopping surveys hinders our understanding of
consumers’ e- shopping behavior.
21
6. Methodologies
Previous studies have employed a number of approaches to empirically analyze consumers’ e- shopping behavior.
Researchers use specific methodologies to answer specific questions in which they are interested. These
approaches can be classified into three categories, based on the complexity with which e- shopping behavior and
their precedent factors are presented: descriptive analysis, correlational analysis, and multivariate analysis. The
purpose of the overview of methodologies presented here is not to critique approaches used in a specific study,
but to summarize the types of methodologies ever used and the strengths and weaknesses of the approaches
themselves.
6.1 Descriptive analysis
Descriptive studies analyze e- shopping behavior at the aggregate level; for example, what percent of internet
users in the sample purchased online, or had an intention to e- shop ( Cole et al., 2003)? The strength of
descriptive analyses is in showing us what happened for a particular sample at a particular time, which provides
a clear picture of observed behavior. Accordingly, descriptive analysis is a very important step in helping us
understand what is going on. However, descriptive studies convey little information about why consumers
behave as they do and hence do not attempt to explain their behavior. Specifically, descriptive studies can
provide only a simple accounting of consumers’ e- shopping experiences on average. But this simplicity may
mask the complex interactions among the factors that explain consumers’ e- shopping behavior.
On the other hand, some descriptive studies do attempt to explain consumers’ behavior ( Hoffman et al., 1999;
Lim, 2003; Raijas, 2002). These researchers asked respondents to report the reasons why they did ( or did not)
use e- shopping. Based on the aggregated results, these studies illuminate consumers’ behavior to some extent.
However, these findings are unable to explain consumers’ behavior at the individual level; for instance, a
consumer’s e- shopping choice may be a tradeoff among various determinants ( see Chapter 7), and it is desirable
to know the relative weights of each determinant in the decision.
6.2 Correlational analysis
Here, a correlation refers loosely to the relationship between two variables, where ( in our context) one is taken to
be dependent and the other explanatory. Correlational methods used in previous studies mainly include Pearson
correlations ( both the dependent and explanatory variables are continuous), t- tests ( the dependent variable is
continuous and the explanatory variable is binary), and chi- square tests ( both variables are categorical) ( e. g.,
Corbitt et al., 2003; Donthu and Garcia, 1999). Through correlational analysis, researchers can test a specific
hypothesis about whether a certain variable affects e- shopping behavior. As a result, correlational studies
attempt to explain rather than merely describe the observed behavior ( or intention). However, the observed
correlation of one variable with e- shopping behavior may act as a proxy for another variable. For example, high
income consumers may be found to be more likely to adopt e- shopping, but this observed correlation may be a
surrogate for the correlation between e- shopping adoption and education status ( since higher education is
correlated with higher income and higher educated consumers have more opportunities to access the internet).
22
Therefore, the weakness of correlational analysis is its lack of ability to account for the influences of other
factors on e- shopping behavior.
6.3 Multivariate analysis
Multivariate analysis permits the researcher to test specific hypotheses about the direction and magnitude of
influence some variable( s) have on e- shopping behavior, while controlling for other variables. Previous studies
have applied multiple regression, structural equations modeling ( SEM), discrete choice modeling, and
discriminant analysis. The purpose of these kinds of approaches is to explain rather than describe consumers’
behavior. Therefore, multivariate studies can provide a more robust understanding of e- shopping behavior,
although the appropriateness of the particular method chosen and the validity of the statistical results are highly
dependent on other assumptions.
Multiple regression is the most common approach used in previous e- shopping studies ( e. g., Henderson and
Divett, 2003; Liao and Cheung, 2001). Analysis of variance ( ANOVA) is a special type of regression model, in
which the dependent variable is continuous and the explanatory variables are categorical. Multiple regression is
easy to understand and implement. However, the results of regression models may be threatened by the
interactions among explanatory variables and possible correlations between some explanatory variables and
residuals. Alternatively, SEM is capable of solving these problems. SEM allows for multiple simultaneous
directions of causality, and distinguishes the direct effect and the indirect effect as well as the total effect of an
explanatory variable on each dependent variable. Most applications of SEM in e- shopping behavior research are
tied to some specific theories such as TAM in Ahn et al. ( 2004), TPB in Choi and Geistfeld ( 2004), TRA in Cho
( 2004), and TCT in Liang and Huang ( 1998), although some researchers established their own modeling
structures ( e. g., Gefen and Straub, 2004; Goldsmith, 2002). Srinivasan et al. ( 2002) applied seemingly unrelated
regressions ( SURs) to study multiple dependent variables. Theoretically, both multiple regression with two or
more equations and SUR are specific types of SEM ( Long, 1983). The major difference between multiple
regression and SUR is that seemingly unrelated regression assumes that the error terms of several regression
equations are correlated, while multiple regression assumes independent errors across equations. The difference
between SUR and SEM is that in SUR, only the error terms are allowed to be correlated; the explanatory
variables in each equation are still assumed to be uncorrelated with the error terms, while in SEM this
assumption is relaxed.
Some studies also used multiple analysis of variance ( MANOVA) ( e. g., Goldsmith and Goldsmith, 2002;
Kaufman- Scarborough and Lindquist, 2002). MANOVA is used to examine the main and interaction effects of
categorical variables on multiple continuous dependent variables. In addition, Vijayasarathy ( 2002) applied
analysis of covariance ( ANCOVA) and multiple analysis of covariance ( MANCOVA) to study e- shopping
behavior. In contrast to ANOVA and MANOVA, ANCOVA and MANCOVA are used to test the effects of both
continuous and categorical explanatory variables on single and multiple continuous dependent variables,
respectively. All of these approaches are also special cases of SEM, in the same way that ANOVA is a special
case of multiple regression.
23
When the dependent variable was categorical, some researchers employed discrete choice models, which are tied
to utility maximization theory ( UMT). UMT is extensively used in the development of predictive models of
human behavior. It assumes that when a consumer is confronted with a set of choices, she attempts to choose the
alternative with the highest utility to her ( Ben- Akiva and Lerman, 1985). “ Utility” refers to a scalar
measurement of the attractiveness of an alternative, and is a function of a vector of attribute values. A decision
maker explicitly or implicitly uses this single index to compare different alternatives. In the context of random
utility theory, what an analyst observes is consumers’ behavior and the utility for each alternative is not known to
her with certainty. As a result, the utility is treated as a random variable. Consumers’ choices are predicted
probabilistically: the probability that an individual chooses an alternative is the probability ( evaluated from the
perspective of the analyst having only incomplete information) that the utility of this alternative exceeds the
utilities of other alternatives in the choice set. Discrete choice models deal with only categorical dependent
variables. When consumers make e- shopping decisions, they may be confronted with a binary choice, such as
whether or not an online purchase is made ( e. g., Bellman et al., 1999; Bhatnagar et al., 2000; Lohse et al., 2000;
van den Poel and Buckinx, 2005), or multiple categorical choices, such as the frequency of e- shopping ( treated
as a nominal variable by Koyuncu and Bhattacharya, 2004) or conventional store vs. internet vs. catalog.
Several studies applied discriminant analysis to study e- shopping behavior ( e. g., Ranganathan and Ganapathy,
2002; Sin and Tse, 2002). The purpose of discriminant analysis is to find the best functions of a set of
explanatory variables by which we can determine which variables discriminate between two or more predefined
groups. In the context of e- shopping behavior, discriminant analysis can be used to determine which variables
are the best predictors of consumers’ e- shopping choice. In most studies, consumers’ e- shopping choices were
pre- classified into two groups such as internet buyers and internet non- buyers.
Finally, various modeling approaches were used to segment e- shoppers and non- e- shoppers. Such approaches
provide further insight into the relationship of e- shopping behavior to various explanatory variables, by
identifying groups of people who weight various characteristics, such as risks and benefits, differently in their
shopping decisions. Bhatnagar and Ghose ( 2004) developed a latent segmentation model to classify internet
consumers into three segments, based on their sensitivity with respect to e- shopping benefits, product risks, and
security risks. Brengman et al. ( 2005) factor- analyzed responses to 21 ( of an original 38) psychographic
statements into six underlying factors: internet convenience, perceived self- inefficacy, internet logistics, internet
distrust, internet offer ( referring to the comparative prices, quality, and selection of goods for internet versus
store shopping), and internet window shopping. They then applied cluster analysis to the factor scores to identify
four e- shopper segments ( tentative shoppers, suspicious learners, shopping lovers, and business users) and four
internet user but non- e- shopper segments ( fearful browsers, positive technology muddlers, negative technology
muddlers, and adventurous browsers).. Clearly, the non- adopter segments differ on their propensity to become
adopters, while all eight segments differ in the types of marketing approaches to which they would be most
receptive.
24
6.4 Summary
Different methodologies have been applied to answer different levels of questions. Generally, descriptive
analysis is used to describe consumers’ e- shopping behavior; correlational analysis goes beyond descriptive
analysis and attempts to analyze how two variables are related; and multivariate analysis is mainly used to
explain consumers’ behavior using many variables considered together. Therefore, although descriptive and
correlational analyses are important steps in helping to construct multivariate analyses, multivariate studies
provide more information than these other two types of analyses. Further, if the sample is somewhat biased,
multivariate analysis is the least affected by the bias ( Sommer and Sommer, 1997). Therefore, multivariate
analysis is ideal to study e- shopping behavior in depth.
25
7. Determinants of E- shopping Behavior
A general purpose of studies relevant to e- shopping behavior is to understand and predict consumers’ e- shopping
behavior as well as, in some cases, to improve the design of e- shopping sites. Accordingly, identifying the
factors influencing e- shopping behavior becomes very important to achieving the goal. Although previous
studies have investigated a large number of potential antecedent factors, this chapter summarizes only those
factors that were empirically found to be significantly associated with e- shopping behavior. It is worth noting
that Chang et al. ( 2005) has reviewed the literature ( published before 2004) related to the adoption of online
shopping, which provided a guideline for our work. They classified the antecedent factors of e- shopping
behavior into three categories: perceived characteristics of the web as a sales channel, online consumer
characteristics, and vendor and product characteristics. This review adapts their work and incorporates some
new studies published in 2004 and 2005. As a convenience, we have appended to the end of this chapter their
tables summarizing the empirical findings with respect to influences on ( 1) online shopping intention and usage,
( 2) attitudes toward online shopping, ( 3) risk perception, and ( 4) trust, as well as determinants of online shopping
discussed in the popular media. In an appendix to the report itself, we have included a tabulation of key facts
regarding the 65 empirical studies we reviewed and cited herein.
7.1 Perceived characteristics of the web as a shopping channel
As a sales channel, the internet deals with both information system and marketing activities. Therefore, both
online features ( such as information) and offline features ( such as delivery) of an online store influence
consumers’ e- shopping behavior ( Ahn et al., 2004). Further, e- stores, traditional stores, and other shopping
channels compete with each other to survive. Consumers’ adoptions of e- shopping are highly dependent on its
superiority to other shopping channels and its perceived drawbacks.
Qualities of e- shopping service
Qualities of e- shopping service include system quality, information quality, service quality, product quality,
delivery quality, and post- purchase quality ( Ahn et al., 2004; Lee, 2002). System quality “ describes the
measures of web sites as information processing systems and taps engineering- oriented performance
characteristics such as operational efficiency and appearance” ( Ahn et al. 2004, p. 407). Ahn et al. ( 2004) found
that an indicator of system quality, which was derived from measurements on multiple traits such as design,
navigation, response time, and so on, had a positive impact on consumers’ e- shopping intention. Similarly,
Belanger et al. ( 2000) and McKnight et al. ( 2002) concluded that the impact of site quality on e- shopping
behavior is positive. Further, some specific measurements of system quality were found to positively influence
consumers’ e- shopping intention and actual use; for example, ease of navigation ( Childers et al., 2001, Yoon
2002) and better design ( Liang and Lai, 2002; Ranganathan and Ganapathy, 2002). Generally, better system
quality tends to increase the likelihood of purchasing online.
Traditionally, information quality means “ the quality of reports that the system produces” ( Ahn et al. 2004, p.
407). In the context of e- shopping, information quality refers to the content and content quality that the e- sites
provide. High- quality information is assumed to help consumers to compare products and make better choices.
26
Empirically, information quality was found to be positively associated with consumers’ e- shopping intention
( Ahn et al., 2004; Ranganathan and Ganapathy, 2002; Shih, 2004). Conversely, if products are hard to find or
little product information is available for an e- store, consumers are less likely to choose it ( Raijas, 2002).
Service quality can be defined as a global judgment about the superiority or excellence of the service provided.
Its typical dimensions include tangibles, reliability, responsiveness, assurance, and empathy ( Berry et al., 1988).
“ Tangibles refer to physical facilities, equipment, and appearance of personnel. Reliability is defined as the
ability to perform the promised service dependably and accurately. Responsiveness refers to the willingness to
help customers and provide prompt service. Assurance refers to knowledge and courtesy of employees and their
ability to inspire trust and confidence. The fifth dimension, Empathy, is defined as individualized attention given
to visitors” ( Li and Lee, 2001). Presumably, high- quality service can improve the ease of use of an internet
service, and hence help consumers complete their e- shopping transactions with minimal intellectual input. Ahn
et al. ( 2004), Chen and Tan ( 2004), and Shih ( 2004) found that service quality is positively associated with
consumers’ intention to use e- shopping. Vijayasarathy and Jones ( 2002) concluded that providing good
post- selection information increases consumers’ intention to shop online.
Product quality refers to the actual functionality of the product. The variety and quality of products in an e- store
have the potential to influence consumers’ patronage. Empirically, product quality was found to positively affect
consumers’ e- shopping intention ( Ahn et al., 2004). By contrast, if the quality of a product cannot be verified in
an e- store, consumers are less likely to choose this e- store ( Raijas, 2002). Further, product variety was found to
be positively associated with consumers’ e- shopping intention and adoption ( e. g., Cho, 2004; Sin and Tse, 2002).
Reliable and timely delivery is one of the fundamental objectives for e- shoppers. Reliable and timely delivery
increases satisfaction and hence likelihood of adoption ( e. g., Ahn et al., 2004, Koyuncu and Bhattacharya, 2004).
Conversely, consumers having concerns over delivery are more likely to abort an online transaction ( Cho, 2004).
E- shopping reduces consumers’ physical travel to traditional stores, and thus those who prefer to avoid
performing the product picking and delivery tasks themselves are more likely to purchase online ( Raijas, 2002).
Post- purchase service quality also influences e- shopping behavior. Previous studies explored the influences of
money- back guarantees and ease of return. Van den Poel and Leunis ( 1999) concluded that providing a
money- back guarantee for consumers increases the likelihood of purchasing online. Consumers concerned over
return difficulties are more likely to abort an online transaction ( Cho, 2004).
As consequences of specific perceived qualities of e- shopping service, perceived usefulness and perceived ease
of use were extensively found to positively influence consumers’ e- shopping intention and actual use ( e. g., Chen
et al., 2002; Childers et al., 2001). These findings are consistent with the fundamental assumptions of TAM ( see
Chapter 3). Conversely, complexity reduces the probability of shopping online ( e. g., Huang, 2000; Verhoef and
Langerak, 2001).
27
Relative advantages of e- shopping
Previous studies have identified various advantages of e- shopping. Some studies presented these advantages in a
general way, such as relative advantage ( Verhoef and Langerak, 2001) and perceived consequences ( Limayem et
al., 2000). Other studies employed more specific measurements: quickness ( Goldsmith and Goldsmith, 2002;
Koyuncu and Bhattacharya, 2004), time saving, ease of ordering ( Raijas, 2000), effort saving ( Cho, 2004;
Verhoef and Langerak, 2001), product value in terms of price and quality ( Vijayasarathy and Jones, 2000),
decreased transaction costs ( Liang and Huang, 1998; Teo and Yu, 2005), price reduction or financial benefits
( e. g., Eastin, 2002; van den Poel and Leunis, 1999). Some advantages fall under the qualities discussed above,
such as increased product variety and value. Generally, e- shopping advantages were found to have a positive
impact on consumers’ e- shopping intention and actual use. Further, if relative advantages of e- shopping fit
consumers’ shopping orientations or personalities ( discussed in Section 7.3), they are more likely to value such
benefits and purchase online.
Combining several elements both of the qualities of e- shopping and its relative advantages, Mathwick et al.
( 2001) surveyed 213 internet and 302 catalog customers of a women’s apparel and housewares retailer with
respect to six types of characteristics: visual appeal, entertainment value, escapism, intrinsic enjoyment, effi-ciency,
and economic value of the vendor’s web site and catalog ( see Figure 5 of Section 2.6, and accompanying
discussion). Those perceptions were then linked to measures of preference and future patronage intentions.
Perceived risk of and confidence in e- shopping
Perceived risk refers to the uncertainty and negative outcomes of performing a behavior. The perceived risk of
e- shopping has been extensively researched. These studies measured the perception of risks either in a general
way ( e. g., Jarvenpaa et al., 2000; Liu and Wei, 2003) or in a specific way. Specific risks include financial ( e. g.,
Bhatnagar et al., 2000), product performance, time/ convenience ( e. g., Forsythe and Shi, 2003), payment ( e. g.,
Koyuncu and Bhattacharya, 2004), transaction security ( e. g., Liao and Cheung, 2001), technology, vendor,
consumer ( e. g., Lim, 2003), privacy ( e. g., Lunn and Suman, 2000), and web deception ( e. g., Grazioli and
Jarvenpaa, 2000). Further, the uncertainty measurements derived from TCT ( such as performance uncertainty,
behavioral uncertainty, environmental uncertainty) are also indicators of risks. In general, perceived risk
negatively influences consumers’ intention and actual use of e- shopping. Conversely, consumers are more likely
to patronize the e- stores having sound security and privacy features ( e. g., Liu et al., 2004; Miyazaki and
Fernandez, 2001).
Perceived confidence in the web as a shopping channel is positively associated with consumers’ e- shopping
intention and actual use. Perceived confidence was measured in two respects: confidence and self- efficacy. The
dimensions of confidence cover confidence in the ability to buy online, diversity of e- shopping companies, and
information availability ( Goldsmith and Goldsmith, 2002). Self- efficacy refers to the belief in one’s ability to
successfully perform a behavior; or the level of task difficulty an individual believes is attainable with her
perceived skill level ( Bandura, 1997). In the e- shopping environment, self- efficacy measures consumers’
confidence in participating in e- shopping activities, including the online shopping process and actual purchase
( Choi and Geistfeld, 2004; Eastin, 2002).
28
Trust
Trust is essentially another way of formulating some of the risk factors discussed above, specifically those
related to “ individuals’ uncertainty regarding the motives, intentions, and prospective actions of others on whom
they depend” ( Kramer, 1999, p. 571). If risk is present, consumers need trust before conducting a transaction;
and the higher the perceived risk, the more the need for trust ( Grazioli and Jarvenpaa, 2000). As found in many
studies, lack of trust in the online transactions and the web vendors is an important obstacle in the market
penetration of e- shopping ( e. g., Corbitt et al., 2003; Liu et al., 2004). Previous studies defined and
operationalized trust in a number of ways. Some studies conceptualized trust as an overall construct: the trustor
is willing to be vulnerable to the action of the trustee ( e. g., Gefen, 2000; van der Heijden, et al., 2003) or a
general belief that the trustee can be trusted ( e. g., Grazioli and Jarvenpaa, 2000). Some researchers combine
these two kinds of definitions to establish their own trust constructs ( e. g., McKnight et al., 2002). Aside from
overall trust, studies have also investigated the effects of specific trust beliefs on the intention to purchase online.
Specifically, integrity, predictability, and familiarity have been found to increase consumers’ e- shopping
intention ( e. g., Gefen and Straub, 2004; van der Heijden and Verhagen, 2004). Web awareness was also found to
positively impact consumers’ e- shopping intention ( Foucault and Scheufele, 2002; Yoon, 2002).
7.2 Vendor and product characteristics
Vendors and consumers constitute two basic actors of a product exchange system. Vendors have the potential to
influence consumers’ e- shopping behavior ( Swaminathan et al., 1999); for example, a well- known vendor name
may help relieve consumers with respect to perceived risk and establish trust with them, and hence motivate
consumers to shop in this e- store. As discussed in Chapter 5, it is important to identify what products are more
suitable to purchase online, since product characteristics will affect consumers’ choices among e- stores,
traditional stores, and other shopping options.
Vendor characteristics
Liao and Cheung ( 2001) evaluated the impacts of the overall quality of e- vendors on e- shopping behavior and
found positive relationships between them. Perceived vendor reputation and size were found to positively affect
consumers’ e- shopping intention with respect to those specific vendors ( Jarvenpaa et al., 2000; McKnight et al.,
2002).
Product characteristics
Phau and Poon ( 2000) concluded that products and services that have a low cost, are frequently purchased,
intangible, and/ or are relatively high on differentiation ( that is, for which a vendor is able to create a sustainable
competitive advantage through product and service differentiation) are more amenable to be purchased over the
internet. Vijayasarathy ( 2002) also found that a product’s tangibility had a significant influence on consumers’
intentions toward e- shopping. However, cost and the interaction between cost and tangibility were not
significantly associated with consumers’ intentions. Further, a product’s asset specificity, the lack of
transferability of the assets from one retailer to another, is positively associated with transaction costs, which, in
turn, have a negative impact on e- shopping intention ( Liang and Huang, 1998).
29
7.3 Online consumer characteristics
Among the two basic actors of product exchange, the retailers’ role is apparent in Sections 7.1 and 7.2; here we
note that consumers’ purchasing behavior is expected to be influenced by their own personal characteristics.
Consumers with different characteristics may react to e- shopping in different ways; for example, those
unfamiliar with computers and the internet will not think e- shopping is an active alternative. Empirically,
previous studies have found that e- shopping behavior is affected by a variety of personal characteristics, such as
their shopping orientations, personality, experiences, demographics, and social and psychological characteristics.
The impact of these characteristics can either be modeled in linear compensatory form together with the traits
discussed in Sections 7.1 and 7.2, or as segmentation variables affecting the weights given to other attributes.
Shopping orientations
Shopping orientations refer to a consumer’s general predispositions toward shopping activities. They are
conceptualized as specific dimensions of lifestyle and operationalized based on activities, interests and opinions
regarding shopping behavior ( Li et al., 1999). Various studies have explored the links between consumers’
shopping orientations and their intention ( or actual choice) to shop online. First, convenience- oriented
consumers are more likely to purchase online ( e. g., Childers et al., 2001; Bhatnagar et al., 2000). Convenience
mainly refers to two dimensions of shopping activities: when to shop and where to shop. E- shopping is
convenient compared to traditional stores since it is free from temporal and spatial constraints. Further, since
e- shopping increases search efficiency and transaction efficiency, time- conscious consumers are more likely to
buy via the internet ( e. g., Sim and Koi, 2002; Sin and Tse, 2002). Consumers who treat e- shopping as an
enjoyable or fun activity are also more likely to purchase online ( e. g., Goldsmith and Goldsmith, 2002; van der
Heijden and Verhagen, 2004). Enjoyment indicates the extent to which consumers perceive e- shopping to
provide an intrinsic utility.. Moreover, consumers who think e- shopping is compatible with their lifestyle and
shopping habits are more likely to engage in e- shopping ( e. g., Chen et al., 2002; Verhoef and Langerak, 2001).
E- shopping compatibility is adapted from innovation diffusion theory, and refers to the degree to which
e- shopping is viewed as being consistent with individuals’ shopping requirements ( Verhoef and Langerak, 2001).
Price- conscious consumers are more likely to purchase online when e- shopping sites offer a lower price or
financial benefits ( e. g., Sim and Koi, 2002; Koyuncu and Bhattacharya, 2004). Finally, brand- oriented
consumers are more likely to purchase online since well- known brands help relieve consumers with respect to
perceived risk ( Lunn and Suman, 2002; van den Poel and Leunis, 1999)
Personality
Consumers’ personality characteristics influence their e- shopping behavior. Innovativeness was found to be
positively associated with the adoption of e- shopping ( e. g., Blake et al. 2003; Donthu and Garcia, 1999;
Limayem et al., 2000). Innovativeness refers to “ the degree to which an individual is receptive to new ideas”
( Chang et al., 2005, p. 553). Similar dimensions of innovativeness include novelty ( Huang, 2000), trying
something new ( Raijas, 2002) and being first to use new technology ( Bellman et al., 1999). Some studies further
defined domain- specific innovativeness – how innovative a consumer is regarding a product category ( e. g.,
Goldsmith, 2002). Donthu and Garcia ( 1999) concluded that consumers who are impulsive or variety- seeking
30
are more likely to purchase online, and those who tend to minimize risk ( risk aversion) are less likely to adopt
e- shopping. By contrast, consumers who are willing to take risks are more likely to buy online ( Sim and Koi,
2002). These results are consistent with the perception that e- shopping is riskier than traditional shopping.
Gefen ( 2000) found that the disposition to trust ( a general inclination to display faith in humanity and to adopt a
trusting stance toward others) is positively related to the adoption of e- shopping. Also, McKnight et al. ( 2002)
found that the willingness to depend on a web vendor has a positive association with the intention to purchase
online.
Social and psychological characteristics
Consumers’ relevant social and psychological characteristics are mainly derived from psychological theories
such as TPB and TAM. Previous findings are consistent with these theories. Favorable attitudes toward
e- shopping are positively associated with consumers’ e- shopping intention and actual use ( e. g., Ahn et al., 2004;
O’Cass and Fenech, 2002). In addition, favorable attitudes toward the internet ( Grazioli and Jarvenpaa, 2000;
Sim and Koi, 2002), toward catalog retailing ( Cho, 2004), and toward marketing and advertising ( Donthu and
Garcia, 1999) were also found to positively affect the adoption of e- shopping. Similarly, favorable behavioral
intention to use has positive impacts on the actual usage of e- shopping ( e. g., Chen et al. 2002; and Limayem et al.,
2000). Favorable intention to use the internet for information gathering was found to positively influence
consumers’ e- shopping intention ( Shim et al., 2001).
Another important social/ psychological variable is perceived behavioral control, which refers to an individual’s
perception of her ability to perform a behavior. Perceived behavioral control was found to have a positive impact
on consumers’ e- shopping intention and actual use ( Limayem et al, 2000; Shim et al., 2001). Similarly, the
subjective norm of an individual ( that is, perceptions of the feasibility of a behavior as evaluated by referent
people) positively influences her intention to purchase online, which, in turn, has a positive effect on the actual
choice of e- shopping ( Choi and Geistfeld, 2004; Limayem et al., 2000). Blake et al. ( 2003) found strong
evidence that the greater the prevalence of internet shopping among one’s social network, the greater one’s own
frequency of e- shopping.
Computer/ internet experience
Consumers’ computer/ internet experience refers to their knowledge of the computer and the internet, as well as
the frequency, duration, length, and nature of their usage ( such as working online, checking email, reading online
news, and information search). Generally, internet usage is positively associated with the intention to use
e- shopping and with actual use ( e. g., Bellman et al., 1999; Forsythe and Shi, 2003). Education and information
technology training ( Liao and Cheung, 2001) and internet knowledge ( Goldsmith and Goldsmith, 2002; Li et al.,
1999) were also found to positively impact consumers’ e- shopping intention and adoption.
In- home shopping experience
In- home shopping refers to shopping activities conducted by mail, telephone, catalog, TV, and internet. In
general, consumers’ in- home shopping experience increases the likelihood of purchasing online. Specifically,
Eastin ( 2004) found that prior usage of the telephone for product acquisition positively affects consumers’
adoption of e- shopping. Similarly, Lunn and Suman ( 2002) found that previous purchases by mail and phone are
31
positively associated with purchasing online. And previous purchases from catalogs were found to have positive
influences on the spending and adoption of e- shopping ( Lohse et al., 2000; Cho, 2004). By contrast, Bellman et
al. ( 1999) found a negative association between ordering by mail from print catalogs and online spending. Their
explanation is that regular mail is too slow for those who spend more money online. Kaufman- Scarborough and
Lindquist ( 2002) found that e- shoppers tend to experience more non- store ( catalog and TV) shopping activities.
Similarly, prior internet shopping experience was found to have a positive association with e- shopping intention
and adoption ( e. g., Shim et al., 2000; Cho, 2004). However, consumers who experienced negative consequences
of e- shopping are less likely to purchase online ( Lunn and Suman, 2002).
Socio- demographics
Consumers’ socio- demographic characteristics influence their e- shopping behavior. The findings on
demographic variables are mixed. For example, Bhatnagar et al. ( 2000) found that men are more likely to
purchase some products ( such as hardware, software, and home electronics) online, but less likely to buy others
( such as apparel and food) via internet. Generally, however, consumers who are male, higher educated, and have
higher incomes are more likely to purchase online. The studies that measured whether consumers owned a credit
card found those who did to be more likely to shop online ( Sim and Koi, 2002; van Slyke et al., 2002). The speed
of the internet connection which consumers used for e- shopping was also found to be positively associated with
actual use of e- shopping ( e. g., Lunn and Suman, 2002).
7.4 Summary
Previous studies have identified various determinants of consumers’ e- shopping behavior. These determinants
mainly cover three essential elements: characteristics of e- shopping as a shopping channel, consumer
characteristics, and vendor and product characteristics. Among these characteristics, the former two have been
examined extensively in previous research, confirming their importance in understanding e- shopping behavior.
Also, consumers’ e- shopping intention and actual use differ by product types.
It should be kept in mind that only significant influences were summarized here. Although some variables
showed consistent effects over some studies, their impacts may be mixed if we consider the insignificant effects
in other studies. Interested readers can refer to Chang et al. ( 2005) for a more detailed discussion on both
significant and insignificant variables.
32
Table 6. Summary of the Effect of Antecedents on the Intention and Usage of Online Shopping
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
33
Table 6 ( continued)
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
34
Table 6 ( continued)
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
35
Table 7. Summary of the Findings of the Effect of Antecedents of Attitude toward Online Shopping
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
36
Table 8. Summary of the Findings of the Effect of Antecedents of Risk Perception
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
37
Table 9. Summary of the Findings of the Effect of Antecedents of Trust
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
38
Table 10. Summary of the Determinants of Online Shopping Discussed in Popular Press
Reprinted from Chang et al. ( 2005); the citations to which the numbers above refer can be found therein.
39
8. Summary
The rapid growth of e- commerce is imposing profound impacts on modern society. On the supply side, the
emergence of e- commerce is greatly changing the operation behavior of some retailers, and is increasing product
internationalization due to its geographically unlimited nature. On the demand side, the pervasiveness of
e- commerce affects how, where, and when consumers shop, and indirectly influences the way in which we live
our lives. However, the development of e- commerce is still in an early stage, and why consumers choose ( or do
not choose) online purchasing is far from being completely understood. To better evaluate and anticipate those
profound impacts of e- commerce, therefore, it is important to further refine our understanding of consumers’
e- shopping behavior.
A number of studies have investigated e- shopping behavior, and reviewing them is valuable for further
improving our understanding. This report aims to summarize previous e- shopping research in a systematic way.
In this review, we are interested primarily in the potential benefits and costs that the internet offers for the
business- to- consumer segment of e- commerce in the transaction ( purchase) channel. An overview of the 65
empirical studies analyzed in this report is provided in the Appendix.
Most previous studies fall into one or more of several theoretical frameworks, including the theory of reasoned
action, the theory of planned behavior, the technology acceptance model, transaction cost theory, innovation
diffusion theory, and others. Among them, social psychological theories ( the theory of reasoned action, the
theory of planned behavior, the technology acceptance model) were widely applied. As shown in the
applications of different theories, e- shopping behavior is not a simple decision process, and thus an integration of
various theories is necessary to deal with its complexities. We suggest synthesizing these theories through the
development of a comprehensive list of benefits and costs, using each of the key constructs of the pertinent
theories as a guide to identifying the nature of those benefits and costs.
The dependent variables mainly include e- shopping intention and actual e- shopping behavior ( a few studies used
attitudes toward e- shopping). E- shopping intention was measured by various dimensions. Among them, the
directly- stated intention to purchase online is the most frequently used measure. Although some studies used a
unidimensional measure, most adopted a latent construct to assess consumers’ e- shopping intentions. Actual
e- shopping behavior mainly includes three dimensions: adoption, spending, and frequency. Most studies
examined one or more of these three dimensions directly, while a few studies constructed a latent variable to
measure actual e- shopping behavior. When both behavioral intention and actual behavior are included in model
development, attention should be paid to the time precedence between intention and behavior.
With respect to sampling, a choice- based sampling approach is probably preferable given that online shopping
activity accounts for a minor proportion of all consumers, and a far smaller proportion of total retail sales. In
previous studies, most chose internet/ computer/ email users or students as their subjects. Generally, a student
sample is a natural choice for some particular products such as books. However, parameter estimates developed
from a student sample lack generalizability to a larger population because of its homogeneity. By contrast, a
more general sample of internet/ computer/ email users is more applicable for e- shopping behavior research.
40
The characteristics of products strongly influence the degree to which they are suitable for selling online.
Mixing product categories in e- shopping behavior research tends to yield vague or inconsistent results. It is
therefore necessary to explicitly consider product characteristics when exploring consumers’ e- shopping
behavior. However, relatively little effort has been invested into product classification in the context of
e- shopping. Although Nelson’s dichotomized system ( search and experience goods) and Peterson et al.’ s
three- dimensional system ( cost, tangibility, and differentiability) provide useful guides for product type
classification, each has some shortcomings. Therefore, more research should focus on the construction of
product classification systems.
Different methodologies have been applied in previous research. Generally, descriptive analysis is used to
describe consumers’ e- shopping behavior; correlational analysis goes beyond descriptive analysis and attempts
to analyze how two variables are related; and multivariate analysis is mainly used to explain consumers’
behavior using many variables considered together. Therefore, although descriptive and correlational analyses
are important steps in helping to construct multivariate analyses, multivariate studies provide more information
than these other two types of analyses. Multivariate analysis is ideal to study e- shopping behavior in depth.
Among multivariate analysis techniques, multiple regression, structural equations modeling, and discrete choice
models were most frequently used.
Previous studies have identified various determinants of consumers’ e- shopping behavior. These determinants
mainly cover three essential elements: characteristics of e- shopping as a shopping channel, consumer
characteristics, and vendor and product characteristics. Among these characteristics, the former two have been
examined extensively in previous research, confirming their importance in understanding e- shopping behavior.
Specifically, the dimensions of channel characteristics of e- shopping include e- shopping service quality, relative
advantages, perceived risk of and confidence in e- shopping, and trust. Consumer characteristics include their
shopping orientations, personality, social and psychological characteristics, computer/ internet experience,
in- home shopping experience, and socio- demographics.
41
References
Ahn, Tony, Seewon Ryu, and Ingoo Han ( 2004). The impact of the online and offline features on the user
acceptance of internet shopping malls. Electronic Commerce Research and Applications, 3 ( 4), 405- 420.
Ajzen, Icek ( 1991). The theory of planned behavior. Organizational Behavior and Human Decision Process, 50
( 2), 179- 211.
Ajzen, Icek and Martin Fishbein ( 1980). Understanding Attitudes and Predicting Social Behavior. NJ:
Prentice- Hall.
American Marketing Association ( 1948). Report of the definitions committee. Journal of Marketing, 13 ( 2),
202- 217.
Bagozzi, Richard P. ( 1974). Marketing as an organized behavior system of exchange. Journal of Marketing, 38
( 4), 77- 81.
Bandura, Albert ( 1997). Self- efficacy: The Exercise of Control. NY: W. H. Freeman and Company.
Belanger, France, Janine S. Hiller, and Wanda J. Smith ( 2002). Trustworthiness in electronic commerce: The role
of privacy, security, and site attributes. Journal of Strategic Information Systems, 11, 245- 270.
Bellman, Steven, Gerald L. Lohse, and Eric J. Johnson ( 2000). Predictors of online buying behavior.
Communications of the ACM, 42 ( 12), 32- 38.
Ben- Akiva, Moshe and Steven R. Lerman ( 1985). Discrete Choice Analysis: Theory and Application to Travel
Demand. Cambridge, MA: The MIT Press.
Berry, Leonard L., A. Parasuraman, and Valarie A. Zeithaml ( 1988). The service- quality puzzle. Business
Horizons, 31 ( 5), 35- 43.
Bhatnagar, Amit and Sanjoy Ghose ( 2004). Segmenting consumers based on the benefits and risks of internet
shopping. Journal of Business Research, 57, 1352- 1360.
Bhatnagar, Amit, Sanjog Misra, and Raghav H. Rao ( 2000). On risk, convenience and internet shopping behavior.
Communications of the ACM, 43 ( 11), 98- 105.
Blake, Brian F., Kimberly A. Neuendorf, and Colin M. Valdiserri ( 2003). Innovativeness and variety of internet
shopping. Internet Research: Electronic Networking Applications and Policy, 13 ( 3), 156- 169.
42
Brengman, Malaika, Maggie Geuens, Bert Weijters, Scott M. Smith, and Willam R. Swinyard ( 2005).
Segmenting internet shoppers based on their web- usage- related lifestyle: A Cross- cultural validation. Journal of
Business Research, 58, 79- 88.
CACI Information Solutions, Market Planning Group ( 2000). Who’s Buying Online? UK 2000. Available at
http:// www. datadepot. co. uk/ downloads/ UK2000. pdf, accessed on April 24, 2005.
Census Bureau, USA ( 2004). Quarterly Retail E- Commerce Sales: 3rd Quarter 2004. Available at
http:// www. census. gov/ mrts/ www/ data/ html/ 3q2004. html, accessed on December 9, 2004.
Chang, Man Kit, Waiman Cheung, and Vincent S. Lai ( 2005). Literature derived reference models for the
adoption of online shopping. Information & Management, 42. 543- 559.
Chen, Lei- da, Mark L. Gillenson, and Daniel L. Sherrell ( 2002). Enticing online consumers: An extended
technology acceptance perspective. Information & Management, 39, 705- 719.
Chen, Lei- da and Justin Tan ( 2004). Technology adaptation in e- commerce: Key determinants of virtual stores
acceptance. European Management Journal, 22 ( 1), 74- 86.
Childers, Terry L., Christopher L. Carr, Joann Peck, and Stephen Carson ( 2001). Hedonic and utilitarian
motivations for online retail shopping behavior. Journal of Retailing, 77, 511- 535.
Cho, Jinsook ( 2004). Likelihood to abort an online transaction: Influences from cognitive evaluations, attitudes,
and behavioral variables. Information & Management, 41, 827- 838.
Choi, Jayoung and Loren V. Geistfeld ( 2004). A cross- cultural investigation of consumer e- shopping adoption.
Journal of Economic Psychology, 25 ( 6), 821- 838.
Cole, Jeffrey I., Michael Suman, Phoebe Schramm, et al. ( 2003). The UCLA Internet Report: Surveying the
Digital Future – Year Three. Available at http:// ccp. ucla. edu/ pdf/ UCLA- Internet- Report- Year- Three. pdf,
accessed on December 9, 2004.
Copeland, Melvin T. ( 1924). Principles of Merchandising. Chicago: A. Shaw Company.
Corbitt, Brian J., Theerasak Thanasankit, and Han Yi ( 2003). Trust and e- commerce: A study of consumer
perceptions. Electronic Commerce Research and Applications, 2, 203- 215.
Davis, Fred D. ( 1993). User acceptance of information technology: System characteristics, user perceptions,