Exploring consumers initial intention and continuance intention

to use online shopping: An extension of the

Expectation-Confirmation Model
Ming-Chi Lee
Dept. of Computer Science and Information Engineering, National Pingtung Institute of
Commerce, Taiwan, R.O.C
Abstract

This paper investigates how consumers perceive and adopt or resist online shopping from

initial intention to continued intention. We proposed a theoretical model which synthesizes

the expectation confirmation model (ECM), technology acceptance model (TAM), theory of

planned behavior (TPB), and perceived risk to identify the determinants which influence

consumers’ online shopping decision processes. The proposed model is then validated

empirically using a longitudinal method with a two-stage survey. The results demonstrate that

satisfaction has the most significant effect on consumers’ continuance intention, followed by

perceived usefulness as a significant but weaker predictor, and also that the initial intention to

use online shopping is adversely affected mainly by perceived risk. This study is significant

for two reasons. First, it synthesizes the expectation confirmation model, technology

acceptance model, theory of planned behavior in a complementary approach and, second, it

help bride the existing gap among resistance, initial acceptance and continuance intention to

use online shopping.

Keywords: Expectation-confirmation model (ECM), Theory of planned behavior (TPB),

Technology acceptance model (TAM), Online shopping.

1
1. Introduction

With the rapid growth of the Internet, online shopping has become one of the most

profitable e-commerces on the Internet over the last decade (Vijayasarathy 2004). However,

online shopping is still confronted with many obstacles. First, considerable Internet users are

still reluctant to use online shopping (Forsythe and Shi, 2003; Soopramanien and Robertson,

2007). Second, the acceptance-discontinuance anomaly phenomenon (i.e., users discontinue

using online shopping after initially accepting it) frequently occurs (Bhattacherjee 2001).

For the former issue, several authors (Forsythe and Shi, 2003; Hsu et al., 2006) have

attributed consumers’ reluctance to purchase online to apparent risks (e.g. credit card stolen

risk, identity theft risk) but these risks and their potential impact on consumers’ continuance

online purchase behavior are unclear. For the latter issue, prior studies have given more

attention to examining factors that drive users to “initially adopt online shopping” (i.e., use

online shopping for the first time), rather than the factors that influence uses to “continue to

use online shopping” after they have adopted it. Although initial acceptance of online

shopping is an important first step toward achieving success of online stores, the eventual

success of online stores still needs customers to repurchase things online. Therefore, with a

better understanding of the consumers’ online purchase decision processes from resistance

and initial adoption to continued intention to use online shopping, practitioners should be able

to make adjustments in their marketing plans that should help to provide them with a

sustainable advantage over their competition.

In order to fill this knowledge gap, this study proposed a theoretical model which

synthesizes the expectation confirmation model (ECM), technology acceptance model (TAM),

theory of planned behavior (TPB), and perceived risk to understand consumers’ online

shopping behavior from initial intention to continuance intention. We combine these

theoretical models and perspectives for the following three reasons. First, although previous

2
research has found ECM to be a robust model for continued IT adoption (Bhattacherjee,

2001), it employs only three variables to explain behavioral intention, namely satisfaction,

confirmation, and post-adoption expectations. However, a user’s behavioral intention toward

adopting IT will also be affected by other factors, such as the opinions of important

individuals (subjective norms) (Fishbein and Ajzen, 1975). Furthermore, even if users have a

strong intention to perform a behavior, they may feel that they lack the necessary resources

and skills (perceived behavioral control) (Ajzen, 1991), and the use of TPB addresses this

gap.

Second, while TPB captures the roles of individuals, organizational members, and social

influences on behavioral intention, it does not inform us what attitudinal beliefs would affect

a user’s attitude toward online shopping (Hsu et al., 2006). According to Taylor’s research

(Taylor and Todd, 1995a), TAM provides two attitudinal beliefs, namely perceived ease of

use and perceived usefulness as two major antecedents of attitude, which make up precedent

factors of attitude for TPB. Moreover, since each theory has distinct roots and is based on a

different set of antecedent variables, we contend that they independently provide a partial

understanding of users’ cognitive processes related to IT usage. It is therefore possible that,

when combined, these theories may collectively provide an improved and more

comprehensive understanding of the cognitive processes and behaviors related to IT usage

than when each theory considered alone.

Third, customers have shown reluctance to complete online purchases, primarily due to

risk concerns (Gefen et al., 2003). This is because there is a lot at stake for consumers as they

entering into a business relationship with distant, faceless online purchases. While perceived

risk has been examined in the online shopping behavior, little has been known about its

influence on the consumers’ continuance intention to shop online, which may help

understand the acceptance-discontinuance anomaly phenomenon. Therefore, we will also use

3
perceived risk to examine customer’s initial acceptance and continued usage of online

shopping.

While prior research has examined ECM, TAM, and TPB independently in explaining

information technology (IT) usage, to the best of our knowledge, no study has yet

theoretically combined these three models. The primary contributions of this study are its

examination of the integration of ECM, TAM, TPB and perceived risk in explaining initial

acceptance and continuance intention toward online shopping and an empirical evaluation of

which factors are critical to affecting this intention. The findings from this paper may

therefore help bridge the existing gap between initial acceptance and continuance streams of

online shopping research. The results of this work would be useful for online store

practitioners in formulating strategies aimed at increasing online purchases.

The rest of the paper is organized as follows. The next section presents our research

model and hypotheses, while Section 3 proposes the measurement method and scales. We

present the research results in Section 4, followed by discussion in Section 5. Finally, the

implications and conclusions of this work are presented in Sections 6 and 7.

2. Theoretical background and literature review

2.1. Expectation-confirmation model (ECM)

Expectation-confirmation theory (ECT) is widely used in the consumer behavior

literature to study consumer satisfaction, post-purchase behavior, and service marketing in

general (Anderson and Sullivan, 1993; Oliver, 1980, 1993). In order to understand users’

continued information technology (IT) usage behavior, Bhattacherjee (2001) adapt the ECT

to propose the expectation confirmation model (ECM), which emphasizes the importance of

post-usage expectation rather than pre-usage expectation. Therefore, instead of pre-usage

expectation, post-usage perceived usefulness, a surrogate for the post-usage expectation is

included. Furthermore, because ECM assumes the influence of perceived performance is

4
already explained by confirmation variable, it does not include the performance variable.

The ECM posits that an individual’s intention to continue IT usage is dependent on three

variables: the user’s level of satisfaction with the IT; the extent of user’s confirmation of

expectations; and post-usage expectations, in the form of perceived usefulness. Fig. 1

presents the ECM.

Post-Adoption Expectations
(Perceived Usefulness)

Confirmation Satisfaction Continued IT

Usage Intention

Fig. 1 Expectation-Confirmation Model

There are five main hypotheses in the ECM. First, users’ satisfaction with IT has a positive

effect on their intention to continue using the IT. Studies in marketing have discovered that

the major reason for a consumer’s decision to repurchase products or patronize services is

their level of satisfaction (e.g. (Bearden and Teel, 1983; Oliver, 1993; Szymanski and Henard,

2001)). Owing to the similarity between re-purchasing products/services in a consumer

context and the continued usage of IT products/services, the ECM posits an equivalent

relationship in the latter context. In turn, user’s satisfaction with IT is determined by the

user’s confirmation of expectations and their perceived usefulness of IT (which is one type of

post-adoption expectation). The confirmation of expectations suggests that users obtained

expected benefits through their usage experiences with the IT, and thus leads to a positive

effect on users’ satisfaction. On the other hand, based on the expectancy-confirmation

5
paradigm, users’ post-usage perceived usefulness of IT has a positive effect on their

satisfaction with IT by working as a baseline for reference against confirmation judgments.

This relationship is supported by the adaptation level theory, which proposes that users

perceive stimuli only in relation to an adapted level.

2.2. The technology acceptance model (TAM)

Davis (1989) first introduced the TAM as a theoretical extension of the theory of

reasoned action (TRA) (Fishbein and Ajzen, 1975) and found that it could better explain

user’s acceptance. TAM proposes that two particular beliefs, perceived usefulness and

perceived ease of use, are the primary drivers for technology acceptance. Perceived

usefulness is defined as ‘‘the degree to which a person believes that using a particular system

would enhance his/ her job performance’’, and perceived ease of use is defined as ‘‘the

degree to which a person believes that using a particular system would be free of physical and

mental effort’’ (Davis, 1989). Further, perceived usefulness and perceived ease of use both

affect a person’s attitude toward using the system, and consistent with TRA, these attitudes

toward using the system determine behavioral intentions, which in turn lead to actual system

use. TAM has been extensively applied to user acceptance research of various types of

technologies including e-mail, word processor, world wide web, enterprise resources

planning (ERP) systems, and e-commerce (Davis, 1989; Gefen and Straub, 1997; Lu et al.,

2009).

2.3. Theory of planned behavior (TPB)

Both the theory of planned behavior (TPB) and TAM were developed based on the

theory of reasoned action (TRA) (Ajzen and Fishbein, 1980; Fishbein and Ajzen, 1975),

which argues that both behavioral attitude and subjective norm affect behavioral intention,

which in turn affects the actual behavior. TPB adds to TRA a third factor – perceived

behavioral control – that affects behavioral intention and actual behavior (Ajzen, 1991).

6
Many studies have replicated and investigated these three constructs and found that they are

valid in explaining individual intention to use various forms of IT (Liao, 1999; Venkatesh,

2000).

2-4 Perceived risks

Since the 1960s, perceived risk theory has been used to explain consumers’ behavior.

Considerable research has examined the impact of risk on traditional consumer decision

making (Lin, 2008). Peter and Ryan (1976) defined perceived risk as a kind of subjective

expected loss, and Featherman and Savlou (Featherman and Pavlou, 2003) also defined

perceived risk as the possible loss when pursuing a desired result. Cunningham(Cunningham,

1967) noted that perceived risk consisted of the size of the potential loss (i.e. that which is at

stake) if the results of the act were not favorable and the individual’s subjective feelings of

certainty that the results will not be favorable. Most of scholars claimed that consumers’

perceived risk is a kind of a multi-dimensional construct. Six components or types of

perceived risk have been identified: financial, performance, social, physical, privacy, and

time-loss (Jacoby and Kaplan, 1972; Kaplan et al., 1974; Roselius, 1971) (see Table 1).

3. Research model and hypothesis development

3-1 Research model

This study synthesizes the ECM, TAM, TPB and perceived risk to hypothesize a theoretical

model to understand consumers’ online shopping behavior from initial intention to

continuance intention. Figure 2 illustrates the research model. It asserts that the continued

intention to use online shopping is a function of: satisfaction post-usage perceived usefulness

and perceived risk. The model further indicates satisfaction mediates the impact of

confirmation and post-usage perceived usefulness on the continued intention conduct online

7
shopping. The proposed constructs and hypotheses are supported by prior studies in

information systems literature.

H2
Confirmation ECM

H5 Satisfaction H1

H3
Perceived
usefulness H4 Continued
Post-usage Intention
Perceived
usefulness

Pre-usage
Perceived H7 TAM
usefulness

H6 H14
H9 Initial
Attitude Intention
H10
H13
Perceived H8
ease of use H11

Perceived
risk
H12
Subjective
norms
TPB

Perceived
behavior
control

Fig. 2. Research model and hypotheses.

3-2 Hypothesis development

3-2-1 Hypotheses about ECM

Prior marketing studies have found that the higher the users’ expectations, the higher are

their satisfaction (Oliver and DeSarbo, 1988). Moreover, the IT adoption literature has

consistently found that post-usage perceived usefulness is the most important determinant of

8
users’ adoption intentions (e.g., (Davis et al., 1989; Taylor and Todd, 1995b; Venkatesh,

2000)). As a result, the ECM posits users’ post-usage perceived usefulness of IT has a

positive effect on their intention to continue IT usage. Lastly, the ECM posits that the users’

confirmation of expectations will have a positive effect on the post-usage perceived

usefulness of IT. Perceived usefulness of IT could thus be adjusted by confirmation

experience, particularly when the users’ initial perceived usefulness is not concrete due to the

uncertainty over what to expect from using the IT (Bhattacherjee, 2001). Because online

shopping is a kind of information technology on the Internet, we derived the following

hypotheses from the ECM:

H1. Users’ satisfaction with online shopping is positively related to their continuance
intention to shop online.

H2. Users’ confirmation of expectations is positively related to their satisfaction toward

online shopping.

H3. Users’ post-usage perceived usefulness of online shopping is positively related to their
satisfaction toward online shopping.

H4. Users’ post-usage perceived usefulness of online shopping is positively related to their
continuance intention to shop online.

H5. Users’ confirmation of expectations is positively related to their post-usage perceived

usefulness of online shopping.

3-2-2 Hypotheses about TAM

Consumers need to see online shopping as a useful information technology that can

improve their purchase efficiency, providing compelling advantages such as broader product

lines, faster transaction, and greater flexibility. Moreover, online shopping users need feel

that the online shopping is easy to use. Both pre-usage perceived usefulness and perceived

ease of use are beliefs that, according to TRA, will affect a user’s attitude and initial intention.

Thus, we have:

9
H6. Pre-usage perceived usefulness is positively related to behavioral attitude toward online

shopping.

H7. Pre-usage perceived usefulness is positively related to the initial intention toward online

shopping.

H8. Perceived ease of use is positively related to behavioral attitude toward online shopping.

In addition, TAM states that perceived usefulness will have a direct effect on user’s

behavioral intention, and perceived ease of use affects behavioral intention indirectly through

pre-usage perceived usefulness (Davis, 1989). That is, pre-usage perceived usefulness

mediates the effect of perceived ease of use on behavioral intention. Many empirical studies

have supported this argument (Venkatesh and Davis, 2000; Wu and Chen, 2005). Thus,

H9: Perceived ease of use is positively related to pre-usage perceived usefulness of online

shopping.

3-2-3 Hypotheses about TPB

Attitude (A) refers to ‘‘the degree of a person’s favorable or unfavorable evaluation or

appraisal of the behavior in question’’ (Fishbein and Ajzen, 1975). According to the TPB,

attitude impacts users’ behavioral intention, which in turn influences their actual behavior.

When individuals form positive attitude towards online shopping, they will have a stronger

intention toward adopting it, and thus they are more likely to use it.

H10. Behavioral attitude toward online shopping is positively related to the initial intention to

shop online.

Subjective norm refers to ‘‘the perceived social pressure to perform or not to perform

the behavior’’ (Ajzen, 1991). In other words, subjective norm is related to the normative

beliefs about the expectation from other people. Many Internet users choose to shop online

because their friends shop online too, and they recommend it to them. Hence, we propose:

H11. Subjective norm is positively related to the initial intention toward online shopping.

10
 Perceived behavioral control refers to ‘‘people’s perception of ease or difficulty in

performing the behavior of interest’’. It is associated with beliefs about the presence of

control factors that may facilitate or hinder the performance of the behavior in question

(Ajzen, 1991, 2002; Liao et al., 2006). In this case, although online shopping is a useful tool

to improve shopping efficiency, users still need to have the basic Internet skills to use it. Thus,

we posit that:

H12. Perceived behavioral control is positively related to the initial intention toward online

shopping.

3-2-4 Hypotheses about perceived risk

The theory of perceived risk has been applied to explain consumer’s behavior in decision

making since the 1960s. Mitchell (1999) suggested that perceived risk is powerful at

explaining consumers’ behavior because consumers are more often motivated to avoid

mistakes than to maximize utility in purchasing. According to Gefen (2003), an increase in

the risk perceived by consumers could reduce their intention to buy through a particular

website. Similar logic should hold true for perceived risk toward continuance intention. Thus,

we propose that:

H13. Perceived risk is negatively related to the initial intention to use online shopping.

H14. Perceived risk is negatively related to the continuance intention to use online shopping.

4. Research methodology

4.1 Questionnaire development

We used a questionnaire survey with two parts to test our theoretical model. The first

part had questions measuring the constructs in the research model, while the second part had

demographic questions about the participants. Each item corresponding to the constructs was

measured using a seven-point Likert scale, with answer choices ranging from “disagree
11
strongly” (1) to “agree strongly” (7), and most of these items were adapted from the extant

literature. The scale items for perceived usefulness and perceived ease of use were adapted

from Davis(1989) and Bhattacherjee (2001), while the scale items for subjective norm,

perceived behavioral control, and attitude were adapted from the Taylor and Todd (1995a). In

addition, the items measuring perceived risk was adapted from Lee (2009), and the

continuance intention, satisfaction, confirmation were adapted from Bhattacherjee (2001).

After the questionnaire was drafted, it was first sent to three academic experts on online

shopping for their review, and then revised according to their comments and suggestions to

make the wording of the items more precise.

4-2 Pilot test

The questionnaire was pilot-tested by convenient sampling. There were 150 responses, of

which 123 were complete, giving a valid response rate of 82%, and the results of the pilot test

were evaluated by using Cronbach’s reliability and factor analysis. The reliability coefficient

was first calculated for the items of each construct, and the standard lower bound for

Cronbach’s alpha set at 0.7 (Hairs et al., 1998), with items that did not significantly

contribute to the reliability being eliminated. A factor analysis was then performed to

examine whether the items produced the anticipated number of factors and whether the

individual items were loaded on their appropriate factors. All items had high loadings on their

related factors and low cross-loadings on other factors, showing good convergent and

discriminate validities.

4.3 Sample plan and data collection

The participants had prior experience of using the Internet but had no experience of

online shopping. They aged from 21 to 60, and their educational backgrounds graduated from

junior school graduates to Ph.D. holders. Finally, 62% of the respondents were male. Table 1

presents further demographic details on the sample group.

12
 These e-marketing courses last for three months. During this period, all the participants

received a two-stage survey. In the first stage survey, they were given a two-hour training

session by three research assistants to help understand the purchasing procedure of the online

shopping. Then, they were asked to just visit shopping stores but were not allowed to

purchase at that time. After this training, we conducted the questionnaire to measure subjects’

perceptions of pre-purchase beliefs for online shopping such as perceived risk, perceived

usefulness, attitude, subjective norm, perceived behavioral control, and initial intention. At

the end of first survey, 322 questionnaires were collected.

In the second stage survey, subjects who had participated in the first-stage survey were

encouraged to purchase online. At the end of the course, we conducted questionnaires to

measure subjects’ post-purchase beliefs for online shopping, such as feelings of confirmation,

satisfaction, and continuance intention. The returned questionnaires with incomplete or

invalid answers were eliminated, and a total of 302 valid responses were received. The results

showed that all the subjects accomplished the shopping procedure completely, namely

browsing, gathering product information, making product and price comparisons, making

purchase decision, and completing a transaction by offering payment and address information.

In addition, 288 of the participants shopped online more than once in the period studies.

5. Results

In analyzing the collected data, we followed the two-step procedure suggested by

Anderson and Gerbing (1988). First, we examined the measurement model to measure

convergent and discriminant validity. We then examined the structural model to investigate

the strength and direction of the relationships among the theoretical constructs.

5.1 Analysis of the measurement model

Convergent validity was assessed based on the criteria that the indicator’s estimated

13
coefficient was significant on its posited underlying construct factor. We evaluated the

measurement scales using the three criteria suggested by Fornell and Lacker (1981), namely:

(1) All indicator factor loadings (  ) should be significant and exceed 0.5

(2) Construct reliabilities should exceed 0.8

(3) Average variance extracted (AVE) by each construct should exceed the variance due to

measurement error for the construct (e.g., AVE should exceed 0.5)

The Cronbach’s alpha scores, shown in Table 2, indicated that each construct exhibited

strong internal reliability, which all the standard factor loading (  ) values in confirmatory

factor analysis of the measurement model exceeded 0.5 and were significant at p=0.001. In

addition, the composite reliabilities of constructs ranged from 0.81 to 0.93, and the AVE,

ranging from 0.57 to 0.83, was greater than the variance due to measurement error. Therefore,

all three conditions for convergent validity were met.

---Insert Table 2 about here---

Discriminant validity assesses the extent to which a concept and its indicators differs

from another concept and its indicators (Bagozzi et al., 1991). According to Fornell and

Larcker (1981), the correlations between items in any two constructs should be lower than the

square root of the average variance shared by items within a construct. As shown in Table 3,

the square root of the variance shared between the construct and its items was greater than the

correlations between the construct and any other construct in the model, satisfying Fornell

and Larckers’ (1981) criteria for discriminant validity. All diagonal values exceeded the

inter-construct correlations, and thus the results confirmed that our instrument had

satisfactory construct validity.

14
 Confirmation 0.31**

Satisfaction
0.33* 0.53**
0.22***

0.37***
Continued
Post-usage Intention
Perceived
usefulness

Perceived -0.11
Pre-usage risk
Perceived 0.21*
usefulness -0.41*
0.11* *

0.41*
Attitude Initial
0.19* Intention
Perceived 0.31*
ease of use
0.22*

0.25**
Subjective
norm

Perceived
behavior
control

Fig. 3. Results of structural modeling analysis.

5.2 Analysis of the structural model

We assessed the overall goodness of fit using the chi-square test. The chi-square test assesses

the adequacy of a hypothesized model in terms of its ability to reflect the variance and

covariance of the data. Due to its tendency to be sensitive to sample size, other fit indices

(namely GFI, AGFI, CFI, NFI, and RFI) were considered in conjunction with the chi-square.

For the statistical significance of parameter estimates, t values were used. The results of

structural equation modeling obtained for the proposed conceptual model revealed a ratio of

chi-square to the degree of freedom (  2 / df ) of 2.04 ( p < 0.001), goodness-of-fit index (GFI)

15
of 0.91, adjusted goodness-of-fit index (AGFI) of 0.85, comparative fit index (CFI) of 0.95,

normed fit index (NFI) of 0.95, relative fit index (RFI) of 0.94, and root mean square error of

approximation (RMSEA) of 0.05. Generally, fit statistics greater than or equal to 0.9 for GFI,

NFI, RFI, and CFI indicate a good model fit (Hairs et al., 1998). Furthermore, RMSEA

values ranging from 0.05 to 0.08 are acceptable (Hairs et al., 1998), indicating that our model

fit was acceptable. The other fit indices, except AGFI, indicated that our proposed model

obtained an adequate model fit. The low AGFI values may have been due to the small sample

size used.

5.3 Hypotheses testing

The fourteen hypotheses presented above were tested collectively using the structural

equation modeling (SEM) approach (Bagozzi et al., 1991), also performed using AMOS 6.

The path significance of each hypothesized association in the research model and variance

explained (R2 value) by each path were examined, and Fig. 3 and Table 4 show the

standardized path coefficients and path significances. All fourteen hypothesized associations

were strongly significant at p<0.05, except for one link between perceived risk and

continuance intention. The continuance intention to use online shopping in this study was

jointly predicted by satisfaction (   0.53, p  0.01 ), perceived usefulness

(   0.37 , p  0.001 ) and perceived rik (   0.12 , p  0.12 ) and these variables together

explained 80% of the variance of intention to use ( R 2  0.80, coefficient of determination).

In addition to its direct effects, post-usage perceived usefulness also had a significant indirect

effect on continuance intention (   0.12 , p  0.05 ) via the satisfaction, explaining 18% of

the variance in the dependent variables. Satisfaction, in turn, was predicted by confirmation

(   0.31 , p  0.01 ) and post-usage perceived usefulness (   0.22 , p  0.001 ).

Confirmation also had a small indirect effect (   0.06 , p  0.05 ) on satisfaction, via the

16
post-usage perceived usefulness construct (see Table 4).

Initial intention was predicted by perceived risk (   0.41 , p  0.01 ), pre-usage

perceived usefulness (   0.21 , p  0.05 ), subjective norm (   0.22 , p  0.05 ) and

perceived behavioral control (   0.25 , p  0.01 ), and together these variables explained

67% of the total variance. To further assess the significance of indirect effects of predictor

variables on intentions to use online shopping, a decomposition of the effects analysis was

conducted (see Table 5). A discussion of these effects for online shopping continuance is

presented in next Section.

---Insert Table 4 about here---

---Insert Table 5 about here---

6. Discussion

The results of this study provide support for the research model and for the hypotheses

regarding the directional linkage among the model’s variables. The overall explanatory power

of our research model had an R-square of 75% for initial intention to use online shopping and

an R-square of 80% for continued intention, suggesting that the extended ECM model is

capable of explaining a relatively high proportion of variation of intention to use online

shopping. Several significant results could be derived from our research framework, and

these are presented below.

We examined the effects of satisfaction, perceived usefulness, and perceived risk on the

continuance intention with regard to online shopping. The results show that satisfaction is the

strongest predictor of users’ continuance intention (   0.53 , p<0.01), followed by

post-usage perceived usefulness (   0.37 , p<0.05) as a significant but weaker predictor, and

perceived risk (   0.11 , p=0.14) as an insignificant predictor. Satisfaction explains 43% of

17
the variance of continuance intention relative to the other predictors, which jointly explained

33% of the variance. Users who are dissatisfied with online shopping may stop using it,

despite having positive perceptions with regard to other elements of the experience. In other

words, dissatisfaction is the necessary condition for online shopping discontinuance.

Therefore, satisfaction may be the key to explaining the online shopping

acceptance-discontinuance anomaly, a little-understood phenomenon in the literature.

Although both the post-usage and pre-usage perceived usefulness were identified in this

study as important determinants of initial intention and continuance intention toward online

shopping, the post-usage perceived usefulness has a stronger effect than pre-usage perceived

usefulness. This is consistent with Bhattacherjee (2001), which found there are often huge

changes in consumers’ expectations following their consumption experience, and these

changes then have an impact on subsequent cognitive processes. The pre-usage perceived

usefulness is a cognitive belief, formed potentially via second-hand information from referent

others, the media, advertising or other sources. These influence sources may be biased, and

hence user attitude potentially may be inaccurate, unrealistic, and uncertain. In contrast,

post-usage perceived usefulness is grounded in users’ first-hand experience, and is therefore

more realistic, unbiased, and less susceptible to change. In view of this, online store

practitioners should adopt a two-fold strategy in order to increase the initial adoption and

continued usage of online shopping: specifically, they should inform new (potential) users of

the potential benefits of online shopping use, and educate old (continued) users on how to use

online shopping more effectively so as to maximize their perceived usefulness with such

methods.

It is worth noting that the perceived risk has not significant influence on the continuance

intention. One possible explanation for this is that once users decide to continue to repurchase

online, they are satisfied with and /or perceive the usefulness of such purchases, and so are

18
willing to accept the related risks. In contrast, perceived risk has a significantly negative

impact on the initial intention to use online shopping. This implies that before adopting

online shopping, concerns about perceived risk are foremost in the minds of online

consumers, and thus may be the prime reason for not undertaking such purchases.

Consequently, practitioners need to develop for risk-reducing strategies that might assist in

inspiring higher confidence in potential customers.

The results of this study also support the idea that the user’s confirmation and post-usage

perceived usefulness are key determinants of satisfaction. In addition, we find that

confirmation has the primary effect on satisfaction. Combined with the earlier finding that

satisfaction is the primary antecedent of continuance intention, these two results confirm the

saliency of the expectation-confirmation paradigm in understanding users’ continuance

intention toward online shopping.

While all the TAM and ECM variables are significant in this study, they implicitly assume

that behavior is volitional. However, online shopping users face several novel constraints,

such as the impersonal nature of the online environment, certain necessary resources and

skills (perceived behavioral control), and the influence of the opinions of other important

persons (subjective norms). These issues call for the inclusion of TPB in the online shopping

adoption model, and both subjective norms and perceived behavioral control are verified as

having a significant influence on continuance intention. This finding implies that when users

find that people around them have adopted online shopping services, they will be more

willing to use them. Online shopping providers may thus use a positive word-of-mouth

strategy to enhance the awareness of their services and promote their benefits. They may also

need to consider how to bring more positive experiences to their existing customers to retain

their acceptance, rather than relying on the mass media to deliver reinforcing message

(Bhattacherjee, 2000).

19
7. Implications

7-1 Implications for Academics

In terms of theory building, this study attempts to develop a new theory by grounding a

new variable in an integration of two schools of the nomological structure model of theory of

reasoned action (TRA) as well as a derivation of expectation disconfirmation theory (EDT)

and applying them into a new context. It is important to note that the new variable, perceived

risk, is compatible with the TAM, TPB and ECM variables. This approach is likely to ensure

a stable theory development. Hence, the proposed model makes an important contribution to

the emerging literature on online shopping.

The present study has two implications for future online shopping research. First, the

empirical results show that the unified model supports all the hypotheses and has good

explanatory power, implying that the integration of ECM, TAM, and TPB provides a model

with a theoretical basis to explain online shopping. This approach may provide an initial

blueprint for the further integration of other theoretical acceptance models. For example,

information technology (IT) research (Venkatesh et al., 2003) has already yielded many

competing models such as innovation diffusion theory (IDT), social cognitive theory (SCT),

expectation disconfirmation theory (EDT) and theory of reasoned action (TRA), each with

different sets of acceptance determinants. It is anticipated that this study may encourage other

research that integrates these competing models into unified ones.

Second, while the results show that the factors in TAM, TPB and ECM all have

significantly direct or indirect effects on continuance intention to use online shopping,

satisfaction has the strongest effect. Since confirmation and expectations are critical

antecedents to satisfaction, future research may explore what factors influence these variables

and how they can be manipulated in order to improve eventual user experience with online

20
shopping, and hence its subsequent continuance.

7-2 Implications for practitioners

The results of this study shed light on some important issues related to consumers’ initial

adoption and continuance intention toward online shopping. First, while the intention to use

online shopping is significantly influenced by many factors, it is adversely affected mainly by

perceived risk. This implies that controlling the risk of online shopping is particularly

important for managers as they decide how to allocate resources to retain and expand their

current customer base. However, building a risk-free online shopping transaction

environment is a difficult thing. Therefore, online stores need to search for risk-reducing

strategies that might assist in inspiring high confidence in potential customers. This study

suggests that they should consider focusing on the prevention of fraud, identity theft, and

financial loss. For example, building secure firewalls to avoid intrusion, developing methods

for strengthening encryption, and authenticating websites are all measures that should be

undertaken. In addition, this study suggests that online shopping companies could develop

trust-building mechanisms to attract customers, such as statements of guarantee, increased

familiarity through advertising, and long-term customer service. It is worth noting that

because online shopping is a less verifiable and controllable environment, online shopping

customers usually have difficulty in asking for compensation when dissatisfaction for

products/services occurred. Thus, we suggest that online shopping companies should provide

customers with legal digital receipts or a guarantee for every transaction in order to increase

confidence in such services.

8. Limitations and suggestions future research

Several limitations and suggestions for future research are worth noting. First, some

threats to internal validity may need to be taken into account with a longitudinal study such as

21
this one, including history, maturation, testing, mortality and regression (Bhattacherjee and

Premkumar, 2004). To minimize the impact of jeopardizing factors, we adopt several

proactive steps, as suggested by Bhattacherjee and Premkumar (2004). The time period

between pre-usage and usage stage is three months, and this period is short enough to avoid

the effects of history and maturation, but long enough to get rid of effect of testing. In

addition, our high response rates between both stages demonstrate a low mortality effect.

Second, because we collected the data for the independent and dependent variables from the

same respondents, concerns about common method bias could arise (Woszczynski and

Whitman, 2004), and we thus conducted Harmon’s one-factor test (Podsakoff et al., 2003) to

assess the risk of this. We entered all the variables into a factor analysis. These factors

emerged with the first factor, accounting for 23.4% of the variance in the variables. Because

more than a single factor emerged from the factor analysis and no general factor accounted

for the majority of the variance in those variables, we saw no evidence to suggest the

presence of common method variance bias. Third, about 62% of the respondents were male in

this empirical study, and thus the gender distribution was not symmetrical. Therefore, the

results of the current empirical study might tend to model the specific behavior of men, rather

than general behavior of all users. Much evidence has shown that gender differences can

cause discrepancies in the effects of satisfaction, perceived behavioral control, and subjective

norms on a user’s behavioral intention (Armitage et al., 2002; Liao et al., 2006). Other

references also show that the effects of ease of use and perceived usefulness on behavioral

intention can be moderated by gender difference (Gefen and Straub, 1997). Accordingly,

further research may be needed to examine the moderating effect of gender difference on the

behavioral performance of online customers.

22
References
Ajzen, I. (1991), "The Theory of Planned Behavior", Organizational Behavior and Human
Decision Processes, Vol. 50 No. 2, pp. 179-211.
Ajzen, I. (2002), "Perceived behavior control, self-efficacy, locus of control, and the theory
of planned behavior", Journal of Applied Social Psychology, Vol. 32 No. 1, pp. 1-20.
Ajzen, I. and Fishbein, M. (1980), Understanding attitudes and predicting social behavior.,
Englewood Cliffs, NJ: Prentice-Hall.
Anderson, E.W. and Sullivan, M.W. (1993), "The antecedents and consequences of
customer satisfaction for firms," Marketing Science, Vol. 12 No. 2, pp. 125-143.
Anderson, J.C. and Gerbing, D.W. (1988), "Structural equation modeling in practice: A
review and recommended two step approach", Psychological Bulletin, Vol. 103 No. 3,
pp. 411-423.
Armitage, C.J., Norman, P. and Conner, M. (2002), "Can the theory of planned behavior
mediate the effects of age, gender and multidimensional health locus of control?",
British Journal of Health Psychology, Vol. 7 No. 2, pp. 299-316.
Bagozzi, R.P., Yi, Y. and Phillips, L.W. (1991), "Assessing Construct Validity in
Organizational Research", Administrative Science Quarterly, Vol. 36 No. 3, pp.
421-430.
Bearden, W.O. and Teel, J.E. (1983), "Selected determinants of consumer satisfaction and
complaint reports", Journal of Marketing Research, Vol. 20 No. 1, pp. 21-28.
Bhattacherjee, A. (2000), "Acceptance of Internet applications services: the case of electronic
brokerages.", IEEE Transactions on Systems, Man, and Cybernetics --- Part A:
Systems and Humans, Vol. 30 No. 4, pp. 411-420.
Bhattacherjee, A. (2001), "Understanding information systems continuance: An expectation
confirmation model", MIS Quarterly, Vol. 25 No. 3, pp. 351-370.
Bhattacherjee, A. and Premkumar, G. (2004), "Understanding changes in belief and attitude
toward information technology usage: a theoretical model and longitudinal test", MIS
Quaterly, Vol. 28 No. 2, pp. 229-254.
Cunningham, S.M. (Ed.). 1967. The major dimensions of perceived risk. In: Cox, D.F. (Ed.),
Risk Taking and Information Handling in Consumer Behaviour.: Harvard
University Press, Boston, MA.
Davis, F.D. (1989), "Perceived usefulness, perceived ease of use, and user acceptance of
information technology.", MIS Quarterly, Vol. 13 No. 3, pp. 319-340.
Davis, F.D., Bagozzi, R.P. and Warshaw, P.R. (1989), "User acceptance of computer
technology: a comparison of two theoretical models", Management Science, Vol. 35
No. 8, pp. 982-1003.
Entrust (2008), "Pshing attack", http://www.entrust.com/phishing/index.htm, Vol. No., pp.
Featherman, M.S. and Pavlou, P.A. (2003), "Predicting e-services adoption: a perceived
risk facets perspective", International Journal of Human-Computer Studies, Vol. 59
No. 4, pp. 451-474.
Fishbein, M. and Ajzen, I. (1975), Belief, attitude, intention, and behavior: An introduction
to theory and research., MA: Addison-Wesley.
Fornell, C. and Larcker, D.F. (1981), "Evaluating Structural Equation Models with
Unobservable Variables and Measurement Error", JMR, Journal of Marketing
Research, Vol. 18 No. 1, pp. 39-47.
Forsythe, S.M. and Shi, B. (2003), "Consumer patronage and risk perceptions in Internet
shopping", Journal of Business Research, Vol. 56 No. 2, pp. 867-875.
Gefen, D., Karahanna, E. and Straub, D.W. (2003), "Trust and TAM in online shopping:
An integrated model", MIS Quaterly, Vol. 27 No. 1, pp. 51-90.

23
Gefen, D. and Straub, D.W. (1997), "Gender differences in the perception and use of email:
An extension to the technology acceptance model", MIS Quarterly, Vol. 21 No. 4, pp.
389-400.
Hairs, J.F., Anderson, R.E. and Tatham, R.L. (1998), Multivariate Data Analysis, 5th ed.
New York:Macmillian.
Hsu, M.H., Yen, C.H., Chiu, C.M. and Chang, C.M. (2006), "A longitudinal investigation
of continued online shopping behavior: An extension of the theory of planned
behavior. ", International Journal Human-Computer Studies, Vol. 64 No. 6, pp.
889-904.
Jacobs, P. (1997), "Privacy: what you need to know", InfoWorld, Vol. 19 No. 44, pp.
111-112.
Jacoby, J. and Kaplan, L., B (1972), the components of perceived risk, Association for
Consumer Research, Chicago.
Kaplan, L.B., Szybille, G.J. and et al. (1974), "Components of perceived risk in product
purchase: a cross validation", Journal of Applied Psychology, Vol. 59 No. 3, pp.
278-291.
Kuisma, T., Laukkanen, T. and Hiltunen, M. (2007), "Mapping the reasons for resistance to
Internet banking: A means-end approach", International Journal of Information
Management, Vol. 27 No. 2, pp. 75-85.
Lee, M.-C. (2009), "Factors influencing the adoption of internet banking: An integration of
TAM and TPB with perceived risk and perceived benefit", Electronic Commerce
Research and Applications, Vol. 8 No. 3, pp. 130-141.
Liao, C., Chen, J.-L. and Yen, D.C. (2006), "Theory of planning behavior (TPB) and
customer satisfaction in the continued use of e-service: An integrated model",
Computers in Human Behavior, Vol. 23 No. 6, pp. 2804-2822.
Liao, S. (1999), "The adoption of virtual banking: an empirical study", International Journal
of Information Management, Vol. 19 No. 1, pp. 63-74.
Lin, W.-B. (2008), "Investigation on the model of consumers' perceived risk --- integrated
viewpoint", Expert Systems with Applications, Vol. 34 No. 1, pp. 977-988.
Littler, D. and Melanthiou, D. (2006), "Consumer perceptions of risk and uncertainty and
the implications for behaviour towards innovative retail services: The case of Internet
Banking", Journal of Retailing and Consumer Services, Vol. 13 No. 6, pp. 431-443.
Lu, Y., Zhou, T. and Wang, B. (2009), "Exploring Chinese users' acceptance of instant
messaging using the theory of planned behavior, the technology acceptance model,
and the flow theory", Computers in Human Behavior, Vol. 25 No. 1, pp. 29-39.
Mitchell, V.W. (1999), "Consumer perceived risk:conceptualizations and models", European
Journal of Marketing, Vol. 33 No. 2, pp. 163-195.
Oliver, R.L. (1980), "A cognitive model for the antecedents and consequences of
satisfaction," Journal of Marketing Research, Vol. 17 No., pp. 460-469.
Oliver, R.L. (1993), "Cognitive, Affective, and Attribute Bases of the Satisfaction Response",
The Journal of Consumer Research, Vol. 20 No. 3, pp. 418-430.
Oliver, R.L. and DeSarbo, W.S. (1988), "Response Determinants in Satisfaction
Judgments", The Journal of Consumer Research,, Vol. 16 No. 4, pp. 495-507.
Peter, J.P. and Ryan, M.J. (1976), "An investigation of perceived risk at the brand level",
Journal of Marketing Research, Vol. 13 No., pp. 184-188.
Podsakoff, P.M., MacKenzie, S.B. and Lee, J.Y. (2003), "Common Method Biases in
Behavioral Research: A Critical Review of the Literature and Recommended
Remedies", Journal of Applied Psychology, Vol. 88 No. 5, pp. 879-903.
Reavley, N. (2005), "Securing online banking", Card Technology Today, Vol. 17 No. 10, pp.
12-13.

24
Roselius, T. (1971), "Consumer Rankings of Risk Reduction Methods", Journal of Marketing
Vol. 35 No. 1, pp. 56-61.
Soopramanien, D.G.R. and Robertson, A. (2007), "Adoption and usage of online shopping:
An empirical analysis of the characteristics of ''buyers'' ''browsers" adn "non-internet
shoppers"", Journal of Retailing and Consumer Services, Vol. 14 No. 1, pp. 73-82.
Szymanski, D.M. and Henard, D.H. (2001), "Customer satisfaction: a meta-analysis of the
empirical evidence", Journal of the Academy of Marketing Science, Vol. 29 No. 1, pp.
16-35.
Taylor, S. and Todd, P.A. (1995a), "Assessing IT usage: the role of prior experience", MIS
Quarterly, Vol. 19 No. 4, pp. 561-570.
Taylor, S. and Todd, P.A. (1995b), "Understanding information technology usage: a test of
competing models", Information Systems Research, Vol. 6 No. 2, pp. 144-176.
Venkatesh, V. (2000), "Determinants of perceived ease of use: integrating control, intrinsic
motivation, and emotion into the technology acceptance model.", Information Systems
Research, Vol. 11 No. 4, pp. 342-365.
Venkatesh, V. and Davis, F.D. (2000), "A theoretical extension of the technology
acceptance model: four longitudinal field studies", Management Science, Vol. 46 No.
2, pp. 186-204.
Venkatesh, V., Morris, M.G., Davis, G.B. and Davis, M.D. (2003), "User acceptance of
information technology: Toward a unified view", MIS Quarterly, Vol. 27 No. 3, pp.
425-478.
Woszczynski, A.B. and Whitman, M.E. (Eds.). 2004. The problem of common method
variance in IS research: The Handbook of Information Systems Research. Idea
Group Inc., Hershey, PA, pp.66-77.
Wu, I.L. and Chen, J.L. (2005), "An extension of Trust and TAM model with TPB in the
initial adoption of online tax: An empirical study", International Journal of
Human-Computer Studies, Vol. 62 No. 6, pp. 784-808.

25
Appendix A. The questionnaire

Constructs Questionnaire items

Pre-usage perceived usefulness (adapted from Davis (1989))
If I were to adopt online shopping, it would enhance my efficiency.
If I were to adopt online shopping, the effect of purchasing goods or services
would improve.
If I were to adopt online shopping, it would enable me to improve the ability of a
commercial transaction.
Post-usage perceived usefulness (adapted from Bhattacherjee (2001))
Using online shopping help me purchase products/services efficiently.
I find online shopping useful in my life.
In my opinion, using online shopping increase my effectiveness in buying
products/services.
Perceived ease of use (adapted from Davis (1989))
I’ll find using online shopping easy for me.
Learning how to use online shopping will be easy for me.
It’ll be easy to find products that I want to buy using online shopping.
Attitude (adapted from Davis (1989))
Using online shopping is a good idea.
I feel good about using online shopping.
I like using online shopping.
Intention (adapted from Davis (1989))
I intend to purchase online.
It will be worth to shop online.
It will be desirable to use online shopping.
Perceived risk (adapted from Lee (2009))
I think that online shopping is risky because it may lead to financial loss for me.
I think that online shopping is risky because it may incur fraud or the hacker
invades.
I think that online shopping is risky because it may cause identity theft.
I think that online shopping is risky because the products/services delivered may
fail to meet to my expectations.
I think that online shopping is risky because it may lead to a time loss for me.
Subjective norm (adapted from Taylor & Todd (1995b))
People important to me support my use of online shopping.
People who influence me think that I should use online shopping.
People whose opinions I value prefer that I should use online shopping
Perceived behavioral control (adapted from Taylor & Todd (1995b))
Using online shopping was entirely within my control.
I had the resources, knowledge, and ability to use online shopping.
I would be able to use online shopping well.
Confirmation (adapted from Bhattacherjee (Bhattacherjee, 2001))
My experience with using online shopping was better than I expected.
The service level provided by online shopping was better than I expected.
The online shopping can meet demands in excess of what I required for the
products/services.

Satisfaction (adapted from Bhattacherjee (2001))

I am satisfied with the performance of online shopping.
I am pleased with the experience of using online shopping.
My decision to use online shopping was a wise one.
Continuance Intention (adapted from Bhattacherjee (2001))
I intend to continue using online shopping on a regular basis in the future.
I will frequently use online shopping in the future.
If I could, I would like to discontinue my use of online shopping (reverse coded).

26
 Table 2: Construct reliability and convergent validity

Composite Average Cronbach’s

Factor
Construct/ Item t-value reliability variance alpha
loading
Indicators (CR) extracted
Pre-usage PU1 0.855 19.110
perceived PU2 0.826 18.339 0.91 0.7335 0.91
usefulness PU3 0.827 19.516

Intention INT1 0.844 13.433

INT2 0.833 15.232 0.91 0.7243 0.91
INT3 0.832 15.223

Perceived ease of PEOU1 0.846 19.518

use PEOU2 0.843 18.978 0.90 0.7734 0.90
PEOU3 0.865 19.267

AT1 0.848 19.168

Attitude AT2 0.851 18.838 0.91 0.7242 0.91
AT3 0.869 18.764

Continued CI1 0.857 12.531

intention CI2 0.845 11.312 0.89 0.7882 0.89
CI3 0.874 9.334

Subjective norm SN1 0.817 18.537

SN2 0.862 18.941 0.89 0.7334 0.88
SN3 0.860 18.857

Perceived PBC1 0.937 19.801

behavioral PBC2 0.861 19.139 0.91 0.7736 0.89
control PBC3 0.870 19.419

IT1 0.927 20.503

satisfaction IT2 0.907 21.100 0.93 0.8257 0.92
IT3 0.902 20.906

CF1 0.769 9.863

confirmation CF2 0.815 10.583 0.81 0.6078 0.80
CF3 0.773 11.676

Post-usage FR1 0.876 17.789

perceived FR2 0.846 18.179 0.82 0.6908 0.81
usefulness FR3 0.844 17.654

PR1 0.873 16.705

Perceived risks PR2 0.839 16.916 0.82 0.6906 0.81
PR3 0.833 13.443
PR4 0.817 12.876

27
 Table 3: Correlation Matrices and Discriminant Validity

Construct Pre-usage Post-usage Perceived Subjective Perceived Attitude Initial Continued Confirm Satisfaction Perceived
perceived perceived ease of Norm Behavior intention Intention -ation risk
usefulness usefulness use control
Pre-usage 0.833
Perceived
usefulness
Post-usage 0.33 0.855
perceived
usefulness
Perceived ease 0.53 0.43 0.867
of use
Subjective norm 0.21 0.41 0.32 0.851
Perceived 0.24 0.34 0.34 0.35 0.856
behavioral
control
Attitude 0.26 0.63 0.51 0.54 0.39 0.85
Initial 0.31 0.22 0.42 0.51 0.31 0.42 0.84
intention
Continuance 0.27 0.69 0.55 0.67 0.51 0.62 0.53 0.90
intention
Confirmation 0.28 0.52 0.43 0.12 0.22 0.54 0.23 0.51 0.88
Satisfaction 0.43 0.61 0.23 0.20 0.20 0.58 0.33 0.24 0.23 0.77
Perceived risk -0.35 -0.23 -0.24 -0.23 -0.26 -0.62 -0.31 -0.24 -0.21 -0.02 0.85

Note: All correlations significant at p< 0.05 except where noted. Diagonal elements are square roots of average variance extracted

28
 Table 4: Summary of hypotheses tests

Hypotheses  p-value Results

H1 satisfaction → continuance intention 0.53 0.008** Yes
H2 confirmation → satisfaction 0.31 0.007** Yes
H3 perceived usefulness → satisfaction 0.22 *** Yes
H4 post-usage perceived usefulness →continuance intention 0.37 *** Yes
H5 confirmation → post-perceived usefulness 0.33 0.04* Yes
H6 pre-usage perceived usefulness → attitude 0.11 0.02* Yes
H7 pre-usage perceived usefulness → initial intention 0.21 0.04* Yes
H8 perceived ease of use → attitude 0.31 0.03* Yes
H9 perceived ease of use → pre-usage perceived usefulness 0.41 0.04* Yes
H10 attitude → initial intention 0.19 0.02* Yes
H11 subjective norm → initial intention 0.22 0.04* Yes
H12 perceived behavior control → initial intention 0.25 0.002** Yes
H13 perceived risk → initial intention -0.41 0.002** Yes
H14 perceived risk → continuance intention -0.11 0.14 No

a. Standardized estimates are shown.

b. * p  0 .05 , * * p  0 .01, * * * p  0 .001

29
 Table 5. Direct, indirect and total effects --- Estimates

Intention Satisfaction Continuance Intention

C
r Direct Indirect Total Direct Indirect Total Direct Indirect Total
i effects effects effects effects effects effects effects effects effects
t
Post-usage 0.183*** - 0.22*** 0.22*** 0.37*** 0.12* 0.49*
perceived
usefulness
Confirmation 0.31** 0.07* 0.48** 0.16* 0.16*
Satisfaction 0.53** 0.06**
Perceived -0.41**
risk
Perceived 0.15* 0.15*
ease of use
Pre-usage 0.21* 0.02* 0.23*
perceived
usefulness
Attitude 0.19* 0.19*
Subjective 0.22* 0.22* 0.11* 0.11*
norm
Perceived 0.25* 0.25* 0.103** 0.103**
behavioral
control

* significant at p < 0.05.

** significant at p < 0.01.
*** significant at p < 0.001

30