ARTICLE IN PRESS

Int. J. Human-Computer Studies 62 (2005) 784–808

www.elsevier.com/locate/ijhcs

An extension of Trust and TAM model with

TPB in the initial adoption of on-line tax:
An empirical study
Ing-Long Wu, Jian-Liang Chen
Department of Information Management, National Chung Chen University, 160, San-Hsing,
Ming-Hsiung, Chia-Yi, Taiwan

Received 23 August 2004; received in revised form 7 March 2005; accepted 22 March 2005

Communicated by P. Zhang

Abstract

While on-line tax is considered as a special type of e-service, the adoption rate of this service
in Taiwan is still relatively low. The initial adoption of on-line tax is the important driving
force to further influence the use and continued use of this service. The model of Trust and
technology acceptance model (TAM) in Gefen et al. (2003a, MIS Quarterly 27(1), 51–90) has
been well studied in on-line shopping and showed that understanding both the Internet
technology and trust issue is important in determining behavioral intention to use. Besides, the
diffusion of on-line tax could also be influenced by the potential antecedents such as
individuals, organizational members, and social system while the issue for innovative
technology is well discussed in Rogers (1995, The Diffusion of Innovation, fourth ed. Free
Press, New York). Theory of planned behavior (TPB) is the model widely used to discuss the
effect of these antecedents in behavioral intention. An extension of Trust and TAM model
with TPB would be in more comprehensive manner to understand behavioral intention to use
on-line tax. Furthermore, a large sample survey is used to empirically examine this framework.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: On-line tax; Trust and TAM model; Trust; TPB

Corresponding author. Tel.: +886 5 2720411x34620; fax: +886 5 2721501.

E-mail address: ilwu@mis.ccu.edu.tw (I.-L. Wu).

1071-5819/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhcs.2005.03.003
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 785

1. Introduction

Customer service is a series of activities designed for resolving purchasing

problems that customers encounter throughout the product life cycle to enhance
customer satisfaction. When customer service is supplied over the Internet,
sometimes automatically, it is referred to as e-service (Turban et al., 2002). In
general, e-service could include customer service as part of on-line shopping and
pure-play service offered in e-commerce. Initially, on-line consumers did not demand
high levels of customer services and the Internet service was fairly basic such as on-
line catalogue, on-line transaction, and order fulfillment. However, on noticing the
Internet bubble burst and the profit gained from e-commerce far away from
marketer expectations, business managers began to search the new potency of e-
commerce. They found that the key to success in the Internet era is mainly attributed
to the ability of providing customers with better service to attract and retain
customers, and eventually, building a long-term relationship with customers.
In contrast, while the functions of government is mainly to provide information
and delivery service to citizens and business partners, government with its customers
such as citizens and business organizations, in essence, can be considered as a special
type of service industry. This consideration drives us to impose e-commerce features
on supporting the operation of government. This is called e-government and a type
of pure-play service offered in e-government. In particular, on-line tax declaration is
an important function of e-government since it is highly related to the life of citizens.
Thus, the government in Taiwan is aggressively encouraging citizens to use this e-
service for their tax declaration. Currently, the survey data indicates that the usage
rate is still quite low regardless the constantly promotional effort. Among the
influential factors of the low usage rate, the key fundamental can be attributed to the
initial adoption (acceptance) of the innovative service by s since the initial adoption
of an e-service is the important driving force to further influence continued use of the
service (Kwon and Zmud, 1987).
For advocating users’ behavior toward the initial adoption of on-line tax, system
developers thus require first understanding their real needs and expectations in order
to offer more favorable services. In fact, an understanding of the users’ behavior
would be fundamentally beneficial to system design of an e-service since it could
effectively identify the barriers for designing reference in advance. However, e-
commerce is a less verifiable and controllable environment in which on-line service or
transaction is offered without physical face-to-face contact and simultaneous
exchange of services and money. The spatial and temporal separation of e-commerce
between customers and e-vendors as well as the unpredictability of the Internet
infrastructure generate an implicit uncertainty around the initial adoption of on-line
service (Pavlou, 2003). Accordingly, the initial adoption of on-line tax basically
involves the acceptance of both the Internet technology and on-line service
providers. As technology acceptance model (TAM) is mainly proposed for
technology-based perspective through two system features of perceived usefulness
(PU) and perceived ease of use (PEOU) (Davis et al., 1989), it is incomplete in the
context of on-line services.
 ARTICLE IN PRESS
786 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

A model, named Trust and TAM, has been previously presented in exploring the
acceptance of on-line shopping setting (Gefen et al., 2003a). This model integratively
placed use of on-line system into both system features such as ease of use and
usefulness and trust in e-vendors. This result indicated that these variables are good
predictors for behavior intention to use on-line shopping. However, a diffusion of
innovative technology is highly related to communication channels, individuals,
organizational members, and social system except for the technology itself (Rogers,
1995). Theory of planned behavior (TPB) is the model widely used in predicting and
explaining human behavior while also considering the roles of individual
organizational members and social system in this process (Ajzen, 1991). Accordingly,
the three influencers in this theory, i.e. attitude, subjective norm and perceived
behavioral control, can be interpreted as attitude for technology role, subjective
norm for organizational members and social system roles, and perceived behavioral
control for individual role.
As the focus of this study is on the on-line tax setting, which is considered as a type
of innovative technology, organizational and social systems such as peer or superior
influence and self-efficacy in computer or external resource constraint should play
the important role in determining the acceptance of on-line tax (Taylor and Todd,
1995). As a result, an extension of Trust and TAM model with TPB including
subjective norm and perceived behavioral control should be in a more comprehen-
sive manner to examine the acceptance of on-line tax. In this extension, trust is
placed as an important antecedent of attitude, subjective norm, and perceived
behavioral control. Hopefully, this will provide us more information to solve this
problem of low usage rate in using on-line tax.

2. Literature review

2.1. On-line tax declaration

As the Internet and its applications are increasingly becoming popular in business
organization and public institutions and governments are indeed a special type of
service industry, its applications in public agencies or e-government in Taiwan has
been greatly driven by current and previous administrations for providing citizens
and organizations with more convenient access to government information and
better services. Among them, on-line tax declaration is one of the top priorities in the
construction of e-government and begins for trial and experimental use around 2
years ago and is going for the third-year period. Taxpayers are still allowed to
declare their tax for the choice of either paper form or e-form. In order words, it is a
voluntary-based context for use of emerging technology. Until now, on-line tax is
still in the initial stage of its usage and the usage rate is still relatively low for keeping
in the interval of 10–15% while it was initially launched in the year 2000. There is no
indication in a stable growth of its usage in the near future. On the basis of the
dilemma in the use of on-line tax, the challenges may lie in convincing taxpayers of
communicating with on-line tax in an efficient, effective, and safe manner. This study
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 787

tries to understand, analyse, and solve this problem from the perspective of the initial
adoption of virtual service. This may explain some of the major reasons for a low
rate in system usage.

2.2. Relevant models in IT adoption

TAM is an adaptation of the theory of reasoned action (TRA) by Fishbein and

Ajzen (1975) and mainly designed for modeling user acceptance of information
technology (Davis et al., 1989). This model hypothesizes that system use is directly
determined by behavioral intention to use, which is in turn influenced by user’s
attitude toward using the system and PU of the system. Attitude and PU are also
affected by PEOU. PU, reflecting a person’s salient belief in using the technology,
will be helpful in improving performance. PEOU, explaining a person’s salient
beliefs in using the technology, will be free of any effort (Taylor and Todd, 1995).
The appeal of this model lies in both specific and parsimonious as well as an
indication of high prediction power of technology usage. These determinants are also
easy to understand for system developers and can be specifically considered during
system requirement analysis and other system development stages. These factors are
common in technology-usage settings and can be applied widely to solve the
acceptance problem (Taylor and Todd, 1995).
TPB underlying the effort of TRA has been proven successful in predicting and
explaining human behavior across various information technologies (Ajzen, 1991,
2002). According to TPB, a person’s actual behavior in performing certain action is
directly influenced by his or her behavioral intention and in turn, jointly determined
by attitude, subjective norm and perceived behavioral control toward performing the
behavior. Behavioral intention is a measure of the strength of one’s willingness to try
and exert while performing certain behavior. Attitude (A) explains the feeling of a
person’s favorable or unfavorable assessment regarding the behavior in question.
Furthermore, a favorable or unfavorable attitude is a direct influence to the strength
of behavioral beliefs about the likely salient consequences. Accordingly, attitude (A)
is equated with attitudinal belief (abi) linking the behavior to a certain outcome
weighted by an evaluation of the desirability of that outcome (ei) in question, i.e.
A ¼ Sabi ei . Subjective norm (SN) expresses the perceived organizational or social
pressure of a person while intending to perform the behavior in question. In other
word, subjective norm is relative to normative beliefs about the expectations of other
persons. It can be depicted as individual’s normative belief (nbi) concerning a
particular referent weighted by motivation to comply with that referent (mci) in
question, i.e. SN ¼ Snbi mci .
Perceived behavioral control (PBC) reflects a person’s perception of ease or
difficulty toward implementing the behavior in interest. It concerns the beliefs about
presence of control factors that may facilitate or hinder to perform the behavior.
Thus, control beliefs about resources and opportunities are the underlying
determinant of perceived behavioral control and it can be depicted as control
beliefs (cbi) weighted by perceived power of the control factor (pi) in question, i.e.
PBC ¼ Scbi pi . In sum, grounded on the effort of TRA, TPB is proposed to eliminate
 ARTICLE IN PRESS
788 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

the limitations of the original model in dealing with the behavior over which people
have incomplete volitional control (Ajzen, 1991). In essence, TPB differs from TRA
in its addition of the component of perceived behavior control.
However, TPB does not further elaborate the relationship between the belief
structures (i.e. Sabi ei , Snbi mci , Scbi pi ) and the antecedents (attitude, subjective
norm, perceived behavior control) of intention. TPB simply combines each of the
belief structures into one unidimensional belief construct and as a result, the belief
structures, in fact, representing a variety of underlying dimensions, may not be
consistently related to the antecedents of intention. Moreover, the underlying
dimensions of the beliefs structures are, in essence, different for various application
settings and this combination makes TPB difficult to be generalizable across various
settings. By decomposing the belief structures of TPB (Decomposed TPB), their
relationships should become clearer, more understandable for practical purpose
(Taylor and Todd, 1995).
Attitudinal belief structure is decomposed into three dimensions: ease of use, PU,
and compatibility. Normative belief structure is decomposed into two dimensions:
peer and superior influences. Control belief structure is decomposed into three
dimensions: individual self-efficacy, resource facilitating conditions, and technology
facilitating conditions. After that, while comparing Decomposed TPB with TAM,
TAM is, in fact, a part of Decomposed TPB and consequently, Decomposed TPB
should provide a more complete understanding of IT adoption relative to the more
parsimonious TAM (Taylor and Todd, 1995). Based on the above logic, it is better
off to extend Trust and TAM model with TPB or Decomposed TPB to widely
consider the potential underlying determinants, system features, individuals,
organizational members and social system, for better predicting the intention
toward the initial adoption of on-line tax.

2.3. Trust

The functionality and contribution of trust can be apparently identified from the
economic framework of social exchange (Kelley and Thibaut, 1978; Kelley, 1979).
Within social exchange, business transactions are usually carried out without explicit
contract or control mechanism against opportunistic behavior so that the parties
involved in these activities are not able to attain complete legal protection and
expose themselves in a complicated social environment with mass uncertainty. To
insure better rewards from the economic activities, people make efforts to reduce this
social complexity and avoid risk from being exploited (Wrightsman, 1972). Trust is
basically seen as a common mechanism for reducing social complexity and perceived
risk of transaction through increasing the expectation of a positive outcome and
perceived certainty regarding the expected behavior of trustee (Luhmann, 1979;
Grabner-Kraeuter, 2002; Gefen, 2004). In particular for on-line business, without
reducing social complexity and risk resulting from the undesirable opportunistic
behavior of e-vendor, only short-term transactions would be possible (Kim et al.,
2004; Pavlou and Gefen, 2004). Accordingly, trust is an important determinant in e-
commerce including public services.
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 789

Moreover, trust was further explained more clearly in terms of a number of trust
antecedents: knowledge-based trust, cognition-based trust, calculative-based trust,
institution-based trust, and personality-based trust (Zucker, 1986; Gefen et al.,
2003a). Knowledge-based trust is built on familiarity with other parties. Familiarity
builds trust because it reduces social uncertainty through increased understanding of
what is happening in the present (Luhmann, 1979). Cognition-based trust examines
how trust is developed from first impression rather than through experience of
personal interactions. According to this research stream, cognition-based trust is
formed through categorization process and illusion of control (Brewer and Silver,
1978; Meyerson et al., 1996). Calculative-based trust can be developed by people’s
rational assessment of the costs and benefits of another party while cheating or
cooperating in the relationship. Trust in this view is derived from an economic
analysis occurring in ongoing relationship, namely that it is not worthy for the other
party to engage in opportunistic behavior (Coleman, 1990; Lewicki and Bunker,
1995; Doney et al., 1998). Institution-based trust refers to an individual’s perception
of an institutional context, which mainly concerns security from guarantees, safety
nets, or other impersonal structures inherent in the specific context (Shapiro, 1987;
McKnight et al., 1998). Personality-based trust or propensity trust explains the
tendency to believe or not to believe in others and further trust them. This type of
trust is based on a belief that the others are typically well meaning and reliable
(Wrightsman, 1972; McKnight et al., 2002).
Among the five types of trust antecedents, cognition-based and personality-based
trusts are more relevant to the formation of the initial trust, since people inherently
has cognitive resource limitation for often recognizing subjects by the first
impression and personality is an important determinant in the initial stage of a
relationship building. Initial trust refers to trust in an unfamiliar trustee while the
actors do not yet have credible, meaningful information about or affective bounds
with each other. While people gain experience and familiarity with the trustee in the
later stage, continued trust by people will be more influenced by experiential personal
interaction (McKnight et al., 1998). In sum, as on-line tax is a type of e-service
between government agency and citizens, and their transactions are primarily
through virtual channel without face-to-face contact, perceived uncertainty and risk
associated with on-line tax are the major concern of the citizens in using this new
technology. Trust will be the important potential influencer to examine the initial
adoption of on-line tax.

2.4. Trust and TAM relationship

The connections between trust and TAM have been widely discussed in literature
in that the relationships between PU, PEOU, and trust are hypothesized in many on-
line-based business settings (Gefen et al., 2003a, b; Pavlou, 2003; Saeed et al., 2003;
Gefen, 2004). In particular, a model of Trust and TAM was well defined in on-line
shopping setting (Gefen et al., 2003a). This model explicitly indicated their
relationship as trust is an antecedent of PU, PEOU is an antecedent of trust, and
trust has a direct influence on behavioral intention to use. Trust is one of the
 ARTICLE IN PRESS
790 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

determinants of PU, especially in an on-line environment, because part of the

guarantee that consumers will sense the expected usefulness from the web site is
based on the sellers behind the web site. Moreover, trust is recognized to have
positive effect on PU since trust allows consumers to become vulnerable to e-vendor
to ensure that they gain the expected useful interaction and service (Pavlou, 2003).
While consumers initially trust their e-vendors and have an idea that adopting on-
line service is beneficial to their job performance, they will believe the on-line service
is useful (Gefen et al., 2003a).
On the other hand, PEOU is hypothesized to have positive influence on trust
because PEOU can help promote customers’ favorable impression on e-vendors in
the initial adoption of on-line service and further, cause customers to be willing to
made investment and commitment in buyer-seller relationship (Ganesan, 1994;
Gefen et al., 2003a). In general, while following the definition of social cognitive
theory, PEOU can be argued to positively influence a person’s favorable outcome
expectation toward the acceptance of an innovative technology (Bandura, 1986).
This is because cognition-based trust, as discussed previously, is mainly built on the
first impression of a person toward certain behavior and extensively, PEOU in terms
of on-line service can be considered the first feeling or expectation established for
further continued on-line transaction. In sum, while on-line tax is considered a
special type of e-service, the Trust and TAM model is partly fitted to this on-line tax
setting while there are additional variables, as discussed below, to be included in the
particular context.

2.5. Trust and TPB relationship

The relationship between trust and TPB can be examined in a variety of aspects in
which trust is hypothesized as the common antecedent of attitude, perceive
behavioral control, and subjective norm. For attitude construct, trust in e-vendor
is viewed as a salient behavioral belief that directly affects customer’s attitude toward
the purchase behavior. While an e-vendor is trustworthy, it is more possible that the
consumer will gain benefits and avoid possible risks from adopting on-line service
(McKnight and Chervany 2002; Pavlou, 2003). As cost-benefit paradigm greatly
influences people’s attitudinal beliefs and outcome judgments, trust can be a direct
influencer that determines people’s attitude toward behavior (Bandura, 1986; Davis
et al., 1989). Besides, research has shown that trust definitely increases the
confidentiality of business relationship and determines the quality of transaction
between buyers and sellers as well as people’s outcome expectation on many
commerce activities (Luhmann, 1979; Lewis and Weigert, 1985; Hosmer, 1995).
According to social cognitive theory, outcome expectation refers to people’s
estimation of a given behavior yielding a particular outcome, which is closely related
to people’s attitude toward behavior (Bandura, 1986). Therefore, trust is apparently
an important antecedent of attitude toward the on-line transaction behavior.
For perceived behavioral control construct, trust can increase perceived
behavioral control over on-line transactions since the virtual interactions between
customers and e-vendors become more expectable (Pavlou, 2002). Explicitly, trust
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 791

influences perceived behavioral control through control factors of self-efficacy and

facilitating favorable conditions. According to the psychological reports, self-efficacy
in personal relationships is constructed from self-confidence and mutual trust in
friendships (Matsushima and Shiomi, 2003). Hence, mutual trust in the relationship
between customers and e-vendors should increase customer self-efficacy and in turn,
increase perceived behavioral control. On the other hand, trust can be a perceptual
resource that facilitates customers to gain control over on-line transactions. While
customers trust an e-vendor that behaves in accordance with their expectation, the
trust beliefs are likely to increase customer’s perceived behavioral control over on-
line transactions (Pavlou, 2002).
For subjective norm construct, researchers have found that mutual trust and
mutual influence between users and IS units are highly correlated to each other based
on a study concerning the performance of information system group (Nelson and
Cooprider, 1996). Furthermore, Decomposed TPB revealed that there are peer and
superior influences on users for determining subjective norm toward IS usage
(Taylor and Todd, 1995). Derivatively, it can be predicted that trust in peers and
superiors about their beliefs of IS usage should play a role in determining subjective
norm. Similarly, trust in e-vendors about their reputation, brand name, and service
may positively influence subjective norm over the behavior of on-line transactions.
Besides, they may indicate certain relationship between trust in peers and superiors
and trust in vendors. As the opinions from the referents of peers and superiors are
positive for certain e-vendors in the market, trust in peers and superiors in this
situation can enhance user beliefs in trusting these e-vendors and in turn, subjective
norm toward the behavior of on-line transactions. Therefore, whatever types of trust
are with direct and indirect influences on subjective norm, they are all the important
antecedents of subjective norm in on-line service.

3. Research model

While on-line tax is considered as a special type of e-service, the initial adoption in
on-line tax, in essence, concerns both the roles of the Internet technology and e-
vendor in providing service. The Trust and TAM model in Gefen et al. (2003a) has
been well studied in on-line shopping setting and showed that understanding both
the Internet technology and trust issue is critical in determining behavioral intention
to use on-line shopping, as discussed in Section 2.3. Besides, the diffusion of on-line
tax could also be influenced by the potential antecedents such as individuals,
organizational members, and social system while the issue for innovative technology
is well discussed in Rogers (1995). An extension of Trust and TAM model with TPB
would be in more comprehensive manner to understand the acceptance behavior
toward on-line tax and hopefully, this extension would provide us with higher
explanatory power to examine this problem and effectively improve the low usage
rate. This extension model in on-line tax is indicated in Fig. 1. Accordingly, the
hypotheses are presented as below.
 ARTICLE IN PRESS
792 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

Hypotheses 1, 2, 5, 6, and 10 are proposed based on TAM as discussed in Section

2.1 while Hypotheses 3 and 4 are initiated underlying TPB as described in Section
2.1. More importantly, Hypotheses 7–9 are the unique features from Trust and TAM
model, which are derived from the detailed discussion in the first, second, and third
paragraphs of Section 2.4, respectively. Hypotheses 11 and 12 are mainly developed
based on Trust and TAM model in Section 2.3, i.e. PEOU indicated as a direct
prediction to trust and trust to PU. Furthermore, these hypotheses were further
verified for their validity by empirical data.
Hypothesis 1. PU has positive effect on intention to use on-line tax.
Hypothesis 2. Attitude has positive impact on intention to use on-line tax.
Hypothesis 3. Perceived behavior control positively influences intention to use on-
line tax.
Hypothesis 4. Subjective norm has positive effect on intention to use on-line tax.
Hypothesis 5. PU has positive impact on attitude to use on-line tax.
Hypothesis 6. PEOU positively influences attitude to use on-line tax.
Hypothesis 7. Trust has positive effect on attitude to use on-line tax.
Hypothesis 8. Trust has positive impact on perceived behavior control to use on-line
tax.
Hypothesis 9. Trust positively influences subjective norm to use on-line tax.
Hypothesis 10. PEOU has positive impact on PU to use on-line tax.

PU TAM

H10 H5 H1

H6 H2
H11 PEOU Attitude Intention
H7
H12 H3

H8
Trust PBC H4

H9

SN TPB

Fig. 1. Research model.

 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 793

Hypothesis 11. Trust has positive effect on PU to use on-line tax.

Hypothesis 12. PEOU positively influences trust in using on-line tax.

4. Research design

A large sample survey of on-line tax declaration was employed to empirically test
this research model. The instrument and respondent sample are designed as below.

4.1. Instrument development

The instrument is designed to include a four-part questionnaire as presented in

Appendix A. The first part is nominal scales and the remainders are seven-point
Likert scales.

4.1.1. Basic information

This part of questionnaire is used to collect basic information about respondent
characteristics including gender, age, education, occupation, and experience (one-
time users for the first year, or continued users for more than 1-year experience) in
on-line income tax declaration.

4.1.2. TAM
This part of questionnaire is constructed based on the constructs of PU and
PEOU in TAM model and is adapted from the measurement defined by Venkatesh
and Davis (1996, 2000), containing four items for both constructs.

4.1.3. TPB
This part of questionnaire is developed based on the constructs of attitude,
perceived behavior control, subjective norm, and intention to use. Attitude is
adapted from the measurement defined by Bhattacherjee (2000), including four
items. Perceived behavior control was adapted from the measurement defined
by Taylor and Todd (1995) and Bhattacherjee (2000), including three items.
Subjective norm is adapted from the measurement defined by Taylor and Todd
(1995) and Bhattacherjee (2000), including three items. Intention to use is adapted
by the measurement defined by Venkatesh and Davis (1996, 2000), including
three items.

4.1.4. Trust
Trust items are composed to reflect trust beliefs of citizens in using on-line tax.
This part of questionnaire is thus adapted from the study of Gefen et al. (2003a).
Because the measurement in Gefen et al. is originally developed for on-line business
and its focus is on customer–seller relationship, therefore, a couple of measuring
items concerning market, opportunistic, and honest issues, which are irrelevant to
 ARTICLE IN PRESS
794 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

the on-line tax setting, are dropped from the list. After the screen and shortening
process, this part comprises three items.

4.2. Sample organizations and respondents

In order to collect on-line tax declaration users’ information, researchers first

required getting permission from the Tax Bureau to express the need for academic
research purpose. Basically, the personal information of the users in on-line income
tax declaration is confidential under the law of privacy right and forbidden to
distribute it. However, under certain circumstances, the Tax Bureau can permit to
provide certain types of the personal information for academic research purpose
while at the same time without violating the law of privacy right. The application
procedure for this service is described as below.
While the application gets approval, the Tax Bureau will help e-mail invitation
letters to the users in the e-service with an elicitation message for the purpose of
understanding their experience in the initial adoption of on-line income tax
declaration. The invitation letter also indicates a web site for the users to instantly
hyperlink to an on-line questionnaire. The users are free to participate in this
invitation. After that, 8000 users were randomly selected from the population sample
and accordingly, invitation letters were sent out by e-mail. Furthermore, in order to
improve survey return, follow-up procedure was carried out with another invitation
letter for non-responding users after 3 weeks.

4.3. Sample demographics

Of the 8000 on-line questionnaires distributed, 1383 users were replied, with
incomplete response and not the one-time users (the continued users) deleted,
resulting in a sample size of 1032 users for an overall response rate of 12.9%. Sample
demographics are depicted in Table 1. The seemingly low response rate raises the
concern about non-response bias. A test for non-response bias was conducted using
two responding subsamples: early and late respondents. These two groups were
correlated on the sample characteristics of gender, age, education, occupation, and
experience. The result indicates that there is no significant systematic non-response

Table 1
Sample demographics

Gender Age Education level Occupation

Female 20.1% o20 0.3% High school 8.9% Finance 7.5%

Male 79.9% 20–29 10.9% College 60.4% Institution 22.9%
30–39 44.8% Graduate 26.5% Information 20.3%
40–49 30.2% Doctorate 4.2% Service 15.2%
450 13.8% Manufacturing 11.9%
Others 22.2%
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 795

bias in the respondent sample, suggesting that the respondent sample was a random
subset of the sample frame.

5. Analysis and findings

5.1. Analysis of the measurement model

First, content validities should be relatively acceptable since the various parts of
questionnaire were all adapted from the literature and have been reviewed carefully
by practitioners. Next, confirmatory factor analysis in AMOS software was used to
analyse construct validities, basically the analytical procedure including three
stages as described below. First, a measurement model should be assessed for
goodness-of-fit. The literature suggested that, for a good model fit, chi-square/
degrees of freedom (w2 =df) should be less than 3, adjusted goodness-of-fit index
(AGFI) should be larger then 0.8, goodness-of-fit index (GFI), normed fit index
(NFI), and comparative fit index (CFI) should all be greater than 0.9, and root mean
square error (RMSE) should be less than 0.10 (Henry and Stone, 1994). Second,
convergent validity is assessed by three criteria. Item loading (l) is at least 0.7 and
significant, composite construct reliability is a minimum of 0.8, and average variance
extracted (AVE) for a construct is larger than 0.5 (Fornell and Larcker, 1981).
Finally, discriminant validity is assessed by the measure that the AVE of each
construct should be larger than its square correlation with other constructs (Fornell
and Larcker, 1981).
The indices for the measurement model indicate a good fit with w2 =df (991.1/
231 ¼ 4.29), AGFI (0.90), GFI (0.93), NFI (0.97), CFI (0.98), and RMSE (0.056).
The results of reliability as well as convergent and discriminant validities for
this model are reported in Table 2. The item loading (l) for these constructs
ranges from 0.78 to 0.98 and is also significant at 0.01 level, construct reliability

Table 2
Construct reliability, convergent validity and discriminant validity

Construct Item loading Construct reliability Factor correlations

AVE ATT PEOU INT PBC PU SN TST

ATT 0.80–0.90 0.92 0.75 —

PEOU 0.93–0.97 0.96 0.87 0.54 —
INT 0.97–0.98 0.98 0.95 0.82 0.50 —
PBC 0.92–0.94 0.95 0.85 0.75 0.65 0.73 —
PU 0.84–0.92 0.93 0.77 0.67 0.48 0.59 0.52 —
SN 0.78–0.98 0.86 0.67 0.24 0.16 0.24 0.17 0.22 —
TST 0.84–0.98 0.92 0.79 0.63 0.44 0.57 0.55 0.45 0.24 —

Attitude (ATT), Perceived ease of use (PEOU), Intention (INT), Perceived belief control (PBC), Perceived
usefulness (PU), Subjective norm (SN), Trust (TST).
 ARTICLE IN PRESS
796 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

ranges from 0.86 to 0.98, and AVE ranges from 0.67 to 0.95. Appendix B also
reports the covariance matrix generated by AMOS. Moreover, the AVE of each
construct is all above its square correlation with other constructs. Thus, this
measurement model indicates a high degree of reliability as well as convergent and
discriminant validities.

5.2. Analysis of the structural model

The technique of structured equation modeling was used to examine the causal
structure of the proposed model in this study. The evaluation of this research model
can be carried out in three steps. First, a GFI for the structural model was examined
as the same GFIs applied in assessing the measurement model. Second, the
standardized path coefficients and their statistical significance for the hypotheses in
this model were estimated. Finally, as a measure of the entire structural equation, an
overall coefficient of determination ðR2 Þ was calculated, similar to that found in
multiple regression analysis. The testing results of GFIs are all under the acceptable
levels with, w2 =df (1049.2/236 ¼ 4.45), AGFI (0.90), GFI (0.92), NFI (0.97), CFI
(0.97), and RMSE (0.06). Furthermore, the standardized path coefficients are all
significant at 0.01 level except for the paths from PU to intention and subjective
norm to intention. As a result, Hypothesis 1 and 4 are not supported while the other
hypotheses are all supported. In general, trust indicates important relationships with
the three antecedents of intention to use in TPB while the relationships in Trust and
TAM model are maintained in on-line tax. The detailed discussion of the results will
be presented by the order of the antecedents of intention to use, attitude, perceived
behavioral control, and subjective norm as well as the relationships among trust,
PEOU, and PU in Trust and TAM model (Fig. 2).

PU
R2 = 0.31
0.08
0.35* 0.34*

0.21* Attitude 0.55* Intention

0.30* PEOU
R2 = 0.59 R2 = 0.69

0.44* 0.40* 0.27*

Trust 0.33* PBC

R2 = 0.19 R2 = 0.27 0.08

0.24*

SN
R2 = 0.08

Fig. 2. Standardized solution of the structural model. Number on path: standardized coefficient, R2 :
coefficient of determination, *: po0:01.
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 797

Intention to use on-line tax in this research is jointly predicted by PU (b ¼ 0:08,

Standardized path coefficient), attitude (b ¼ 0:55), perceived behavior control
(b ¼ 0:27), and subjective norm (b ¼ 0:05) and these variables totally explain 69%
of the variance on intention to use (R2 ¼ 0:69, Coefficient of determination). While
comparing the presented results with previous TPB-based studies in IS acceptance,
the explanatory power of the current research model for behavioral intention to use
is higher than Taylor and Todd (1995) with R2 ¼ 0:60, Bhattacherjee (2000) with
R2 ¼ 0:52, and Chau and Hu (2001) with R2 ¼ 0:42. Among these relationships,
attitude toward the behavior and perceived behavior control are two major
influencers on individual’s behavioral intention to use on-line tax. Moreover,
attitude indicates more importance than perceived behavior control in determining
behavioral intention to use on-line tax. The result quite conforms to the findings
reported with business-based setting in prior research. Nevertheless, PU and
subjective norm do not produce significant impacts on behavioral intention to use in
this research.
For the result in PU, previous empirical studies on TAM and extended TAM have
shown inconsistence for either with significant influence (Moore and Benbasat, 1991;
Chau, 1996) or with insignificant influence on behavioral intention to use (Chen
et al., 2002). Indeed, it, in essence, implies an indirect influence of PU on behavioral
intention to use via the mediator, attitude toward using on-line tax. A plausible
reason for this may be explained as below. The on-line tax context in this study is
focused on the stage of the initial adoption and voluntary use in tax declaration.
In other words, users in the on-line tax are still in a trial and experimental
manner. Users’ positive PU in using on-line tax may not immediately lead to a
behavioral intention to use, rather than firstly form a favorable attitude/belief to use
on-line tax. The favorable attitude/belief to use on-line tax is just like a time cushion
before directly taking behavioral intention to use on-line tax. This implies that
potential users would need to take a period of time to carefully change their
psychological state to adopting on-line tax. Consequently, the attitude toward
adopting on-line tax demonstrates a larger influential power on behavioral intention
to use (b ¼ 0:55).
For subjective norm, the result is similar to the finding reported in Taylor and
Todd (1995) and Chau and Hu (2001), but differs from the conclusion in
Bhattacherjee (2000) for exploring the adoption of e-service with the case of
electronic brokerage. The latter one indicated that subjective norm could influence
intention to use as strong as attitude does. However, Venkatesh and Davis (2000)
gave a more complete report in that subjective norm could significantly determine
intention to use in a mandatory-usage context, but its impact would become less
significant while users are in a voluntary-usage context as the case of on-line tax in
this study. In particular, while on-line tax in this study is placing at the initial
adoption stage, there are lack of enough references from prior adopters such as
friends, peers and superiors (perceived social pressure). From the perspective, on-line
tax in this study quite differs from the case of e-service in Bhattacherjee’s study.
Accordingly, it is reasonable to expect that the effect of subjective norm on intention
to use on-line tax should indicate insignificance.
 ARTICLE IN PRESS
798 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

Attitude is predicted by PU (b ¼ 0:34), PEOU (b ¼ 0:21), and trust (b ¼ 0:40)

with jointly 59% of the total variance explained. In that, the effect of trust on
attitude is greater than PU and PEOU. This implies an important fact for researchers
that traditional TAM may not completely explain the attitude/behavior toward the
acceptance of on-line tax. The result also partially validates the conclusion of Trust
and TAM model by Gefen et al. (2003a) since the influential relationship is in terms
of trust and behavioral intention to use in the Trust and TAM model. In general,
trust should be necessarily included in TAM for effectively understanding the
acceptance of e-service. Moreover, trust (b ¼ 0:33) explains 27% of the total
variance in determining perceived behavioral control and is considered as an
important antecedent of perceived behavioral control in on-line tax. In other words,
while citizens trust the on-line tax provider that behaves to improve self-efficacy in
computer or external resource constraint such as the Internet infrastructure for
citizens, the trust beliefs will be able to increase citizen’s perceived behavioral control
in performing the behavior.
On the other hand, trust (b ¼ 0:24) significantly influences subjective norm while
explaining only 8% of the total variance in subjective norm. The reason for this is
two-fold. First, this indicates that while users establish the initial trust in on-line tax,
it will help enhance the users’ normative beliefs about the expectations of referents
such as friends, peers, and superiors who concern the initial adoption of the on-line
tax. The connection between user’s trust and perceived social pressure to perform
on-line tax behavior seems to be expectable as the underlying definition in this
model. Next, the reason for 8% of the total variance explained might be because
there are a number of potential influencers to subjective norm remaining to be
identified for accounting for the rest of the total variance explained. In sum, trust,
generally, is closely linked to the three antecedents of behavioral intention to use in
TPB in the on-line tax setting. This validates the necessity to extend Trust and TAM
model with TPB in this study in order to have larger explanatory power in the initial
adoption of on-line tax (R2 ¼ 0:69 as indicated above).
Finally, trust (b ¼ 0:30) and PEOU (b ¼ 0:35) both significantly influence PU and
jointly explain 31% of the total variance in PU. The former is similar to the findings
reported in the literature such as Trust and TAM model in Gefen et al. (2003a)
and this model discussed in Pavlou (2003). The latter regularly corroborates
most prior research on TAM in both on-line and general information techno-
logies. Furthermore, PEOU (b ¼ 0:44), as discussed earlier in the literature,
significantly affects trust and explains 19% of the total variance in trust. This result
also conforms to Trust and TAM model in Gefen et al. (2003a) in on-line shopping
setting.

6. General discussions

There are many issues influencing user’s decision in the initial adoption of on-line
service. While considering both the Internet and e-vendor issues in the acceptance of
on-line service, Trust and TAM model, as discussed in Gefen et al. (2003a), is well
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 799

defined for its validity in exploring on-line service setting. Extensively, on-line tax is
considered as a special type of e-service and the diffusion of this service
might concern the roles of individual state, organizational members, and social
system except for the factors in TAM. An extension of Trust and TAM model
with TPB aims at increasing the predictive power of behavioral intention to use on-
line tax. Empirical data show that trust is considered as an important antecedent
of the three determinants of intention to use, attitude with b ¼ 0:40, perceived
behavioral control with b ¼ 0:33, and subjective norm with b ¼ 0:24, and in
turn, jointly contributes a high explanatory power with R2 ¼ 0:69 to behavioral
intention to use on-line tax. While compared to other models with trust and TAM in
the literature, this extension with TPB empirically demonstrates substantial
improvement in the explanatory power of behavioral intention to use on-line tax.
This result indeed provides more insight for understanding the low usage rate in on-
line tax.
Next, although PU (b ¼ 0:34) and PEOU (b ¼ 0:21) in TAM more likely
representing technology-based antecedents both significantly influence attitude
toward the behavior; however, trust (b ¼ 0:40) indicating trust-based antecedent
demonstrates more positive impact on the attitude. The results indicate a fact that
initial users tend to rely more on trust in non-technology features than on PEOU and
usefulness in technology-based features to form their attitude toward the behavior.
As a result, they jointly determine 59% of the total variance in the attitude.
Moreover, trust (b ¼ 0:30) and PEOU (b ¼ 0:35) both have positive impact on PU.
As we knew from previous research, PU always showed it as an important
determinant of attitude in TAM and PEOU may often indicate its influence on
attitude through the mediator of PU. The reason can be explained by that PEOU has
been well recognized as a basic requirement for system design and deductively, its
impact on attitude toward adopting information technology has increasingly become
of less importance (Davis et al., 1989; Chau, 1996). This can also be found in this
study, b ¼ 0:35 for PEOU linking to PU versus b ¼ 0:21 for PEOU linking to
attitude. In addition, this study indicates that trust almost plays an equally
influencing role on PU as PEOU.

7. Conclusions

The purpose of this research is to propose an extension of Trust and TAM model
with TPB in a more comprehensive manner that jointly predicts user acceptance
(initial adoption) in on-line tax. A large sample survey from users of on-line tax was
employed to empirically examine this research model. There are several new findings
regarding the roles of Trust, TAM, and TPB in on-line tax as discussed previously.
These findings have important implications for both practitioners and researchers.
For practitioners, although on-line tax is mainly presented for usage by the
features of the Internet and communication technologies, however, this study
shows that recognizing both technological and trust-based issues are important
in increasing citizen’s behavioral intention to use this service. The TAM beliefs
 ARTICLE IN PRESS
800 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

(PU and PEOU) and trust are shown to be two sets of underlying antecedents in
determining behavioral intention to use, each contributing its significant influence
on behavioral intention to use through a number of mediators such as attitude,
perceived behavioral control, and subjective norm. This means that to effectively
attract citizens to use on-line tax, the design of on-line tax needs to carefully
pay attention to both aspects. Besides, as discussed previously, novice users tend to
rely more on trust in non-technology features than on PEOU and usefulness
in technology-based features to develop their attitude toward the behavior. In
other words, trust is more important in determining user’s attitude than PEOU
and usefulness in on-line tax. The major trust-based concerns may include
privacy protection, accuracy to declaration, and unauthorized access and
so on.
Fundamentally, while trust is empirically identified as an antecedent of PU and in
turn, an antecedent of attitude, this has some practical implications in enhancing the
attitude toward using on-line tax. On-line tax provider should first develop trust-
building mechanisms for citizens in order to attract novice users to accept on-line
tax. Examples of the mechanisms include statements of guarantees, increased
familiarity through advertising, long-term customer service, and offering
incentives to use. After that, PU of on-line tax emerges as an important issue in
attracting new users and should be carefully designed in terms of users’ requirements
to reflect PU of this service. Without an original consideration from trust aspect, a
well-designed on-line tax with significant PU will not well perform in attracting
novice users.
For researchers, past research on technology acceptance implicitly assumed that
the success of system use is mainly dependent on technological aspect and does not
consider the notion of uncertainty. However, the advent of the Internet has
introduced uncertainty and risk in system acceptance and use because people often
need to use the Internet to communicate, collaborate, and transact with individuals
and organizations without physical face-to-face interaction. Thus, uncertainty is
increasingly becoming the underlying determinant of the Internet-base system usage.
Traditionally, TAM mainly focuses on the aspect of system features and thus, is
insufficient in capturing the roles of individuals, organizational members, and
social system in the Internet-based system usage, in particular, on-line tax. TPB
with the antecedents of attitude, perceived behavioral control, and subjective norm
will be in a complementary manner to enhance the prediction capability of TAM.
This study extends Trust and TAM model with TPB in exploring on-line tax and
further, empirically demonstrates relatively satisfactory results for providing
more insight to this problem. This approach may be as a basis for similar research
in the area.
Furthermore, subsequent research can be founded on this work. This study has
focused on users who are inexperienced or the initial adoption in e-service. However,
prior research has suggested that determinants of behavioral intention change in
terms of users’ level of experience (McKnight et al., 1998; Karahanna et al., 1999).
Additional research, both longitudinal and cross-sectional, is needed to examine the
differences of this framework as users evolving from being aware of the e-vendor, to
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 801

having experience with the e-vendor, to being continued use of the e-vendor.
Despite the significant influence of trust on subjective norm, there is only 8% of
total variance explained in subjective norm. Thus, it is possible to identify
potential factors that could influence subjective norm to some extent. Future
research could be explored on the matter to better predict subjective norm and in
turn, behavioral intention to use. Other possible beliefs have been suggested in the
management and psychological areas, including loyalty, reliability, and openness
(Hosmer, 1995). More research with the alternative conceptualization of trust would
be useful in more understanding the role of trust in the initial adoption of on-line
service.
Finally, although this study has produced some interesting results, it may still have
some limitations. First, approximately 80% of the respondents are male in this
empirical study. Much research has shown that gender difference could cause
discrepancies in the effects of attitude, perceived behavioral control, and subjective
norm on user’s behavioral intention (Venkatesh and Morris, 2000; Armitage et al.,
2002). Although gender does not produce statistical significance on systematic
non-response bias in the sample respondents, the empirical findings may be little
biased for not reflecting the population distribution of gender. Next, there are
approximately 10–15% of taxpayers in adopting on-line tax. Obviously, the on-line
tax is still at the early stage of adoption. Definitely, this research is greatly necessary
for us to gain more insight on further promoting its widespread usage. This imposes
a limitation of generalizability to the population. However, the same respondents are
randomly selected from the sample frame and thus, in a position to be well
representative of the population. As a result, the empirical findings should be free for
the population problem and can be widely generalized for its practical use.
 802
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808
Appendix A. Questionnaire

ARTICLE IN PRESS
Part 1. Basic information
1. Gender: &Female &Male
2. Age: &Less than 20 years old &20–30 years old &30–40 years old
&40–50 years old &Larger than 50 years old
3. Education: &High school &College &Graduate school &Doctorate
4. Occupation: &Finance &Institution &Information &Service &Manufacturing
&Other
5. Experience in using on-line income tax declaration: &One-time user &Continued user

Part 2–4. Constituent constructs in hypothetic research model

Scale design for the following questionnaire:
1: Strongly disagree (SD) 2: Moderately disagree 3: Somewhat disagree
4: Neutral (N) 5: Somewhat agree 6: Moderately agree
7: Strongly agree (SA)

Note: OITD: abbreviation of on-line income tax declaration.

 SD N SA
Perceived usefulness (adapted from Venkatesh and Davis, 1996, 2000)
PU1 Using the OITD would improve my performance in income tax 1 2 3 4 5 6 7
declaration.
PU2 Using the OITD would improve my productivity in income tax 1 2 3 4 5 6 7

I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

declaration.
PU3 Using the OITD would enhance my effectiveness in income tax 1 2 3 4 5 6 7
declaration.
PU4 I find the OITD to be useful in income tax declaration. 1 2 3 4 5 6 7

ARTICLE IN PRESS
Ease of use (adapted from Venkatesh and Davis, 1996, 2000)
EOU1 My interaction with the OITD is clear and understandable. 1 2 3 4 5 6 7
EOU2 Interaction with the OITD does not require a lot of mental effort. 1 2 3 4 5 6 7
EOU3 It is easy to get the OITD to do what I want it to do. 1 2 3 4 5 6 7
EOU4 It is easy to use the OITD. 1 2 3 4 5 6 7

Attitude (adapted from Bhattacherjee, 2000)

ATT1 Using OITD for income tax declaration would be a good idea. 1 2 3 4 5 6 7
ATT2 Using OITD for income tax declaration would be a wise idea. 1 2 3 4 5 6 7
ATT3 I like the idea of using OITD for income tax declaration. 1 2 3 4 5 6 7
ATT4 Using OITD for income tax declaration would be a pleasant 1 2 3 4 5 6 7
experience.

Subjective norm (adapted from Taylor and Todd, 1995; Bhattacherjee, 2000)
SN1 People who are important to me would think that I should use 1 2 3 4 5 6 7
OITD.
SN2 People who influence me would think that I should use OITD. 1 2 3 4 5 6 7
SN3 People whose opinions are valued to me would prefer that I should 1 2 3 4 5 6 7
use OITD.

803
 804
Behavioral control (adapted from Taylor and Todd, 1995; Bhattacherjee, 2000)
PBC1 I would be able to use the OITD well for income tax declaration. 1 2 3 4 5 6 7
PBC2 Using OITD was entirely within my control. 1 2 3 4 5 6 7
PBC3 I had the resources, knowledge, and ability to use OITD. 1 2 3 4 5 6 7

I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

Intention to use (adapted from Venkatesh and Davis, 1996, 2000)
INT1 Assuming I have access to the OITD, I intend to use it. 1 2 3 4 5 6 7
INT2 Given that I have access to the OITD, I predict that I would use it. 1 2 3 4 5 6 7
INT3 If I have access to the OITD, I want to use it as much as possible. 1 2 3 4 5 6 7

ARTICLE IN PRESS
Trust (adapted from Gefen et al., 2003a)
TST1 Based on my perception with OITD, I know it is predictable for the 1 2 3 4 5 6 7
service.
TST2 Based on my perception with OITD, I believe it provides good 1 2 3 4 5 6 7
service.
TST3 Based on my perception with OITD, I believe it helps or cares citizens 1 2 3 4 5 6 7
in tax declaration.
Appendix B. Covariance matrix

I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

PU1 PU2 PU3 PU4 EOU1 EOU2 EOU3 EOU4 ATT1 ATT2 ATT3 ATT4 SN1 SN2 SN3 PBC1 PBC2 PBC3 INT1 INT2 INT3 TST1 TST2 TST3

PU1 0.943
PU2 0.892 1.392
PU3 0.819 0.962 1.056
PU4 0.701 0.73 0.748 0.804
EOU1 0.412 0.476 0.456 0.461 1.314
EOU2 0.451 0.516 0.49 0.488 1.213 1.332

ARTICLE IN PRESS
EOU3 0.44 0.527 0.49 0.474 1.204 1.228 1.339
EOU4 0.469 0.523 0.507 0.49 1.169 1.229 1.228 1.397
ATT1 0.483 0.516 0.519 0.489 0.432 0.469 0.472 0.487 0.845
ATT2 0.408 0.413 0.438 0.428 0.402 0.415 0.425 0.436 0.679 0.892
ATT3 0.449 0.472 0.487 0.468 0.417 0.441 0.463 0.475 0.662 0.611 0.805
ATT4 0.535 0.595 0.578 0.564 0.667 0.711 0.698 0.766 0.737 0.627 0.731 1.189
SN1 0.32 0.326 0.354 0.305 0.298 0.319 0.305 0.33 0.401 0.302 0.38 0.576 1.747
SN2 0.217 0.243 0.26 0.201 0.192 0.204 0.19 0.245 0.243 0.115 0.204 0.392 1.359 2.152
SN3 0.253 0.281 0.294 0.221 0.211 0.22 0.215 0.271 0.262 0.148 0.218 0.404 1.386 1.799 1.909
PBC1 0.416 0.46 0.446 0.44 0.629 0.648 0.683 0.666 0.562 0.511 0.557 0.745 0.338 0.202 0.217 0.959
PBC2 0.434 0.489 0.471 0.457 0.717 0.723 0.757 0.763 0.582 0.531 0.577 0.827 0.388 0.255 0.256 0.937 1.228
PBC3 0.383 0.431 0.417 0.416 0.635 0.619 0.663 0.641 0.534 0.497 0.541 0.689 0.269 0.111 0.13 0.836 0.929 0.969
INT1 0.513 0.556 0.546 0.521 0.535 0.561 0.578 0.582 0.656 0.582 0.655 0.794 0.438 0.288 0.323 0.701 0.734 0.679 1.046
INT2 0.506 0.548 0.546 0.513 0.514 0.547 0.572 0.569 0.666 0.581 0.659 0.791 0.442 0.293 0.322 0.696 0.718 0.659 1.002 1.05
INT3 0.512 0.554 0.545 0.525 0.513 0.551 0.568 0.56 0.672 0.582 0.663 0.789 0.427 0.288 0.309 0.69 0.703 0.653 0.977 0.997 1.029
TST1 0.439 0.503 0.497 0.431 0.542 0.552 0.559 0.604 0.549 0.467 0.51 0.73 0.454 0.414 0.407 0.539 0.671 0.514 0.621 0.632 0.612 1.764
TST2 0.441 0.495 0.484 0.445 0.527 0.543 0.555 0.584 0.593 0.529 0.551 0.745 0.433 0.354 0.354 0.571 0.679 0.547 0.641 0.65 0.634 1.286 1.366
TST3 0.439 0.494 0.476 0.434 0.509 0.535 0.544 0.573 0.601 0.525 0.552 0.754 0.442 0.346 0.353 0.578 0.678 0.554 0.645 0.66 0.643 1.243 1.291 1.378

805
 ARTICLE IN PRESS
806 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

References

Ajzen, I., 1991. The theory of planned behavior. Organizational Behavior and Human Decision Process
50, 179–211.
Ajzen, I., 2002. Perceived behavior control, self-efficacy, locus of control, and the theory of planned
behavior. Journal of Applied Social Psychology 32, 1–20.
Armitage, C.J., Norman, P., 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 7, 299–316.
Bandura, A., 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice Hall,
Englewood Cliffs, NJ.
Bhattacherjee, A., 2000. Acceptance of e-commerce services: the case of electronic brokerages. IEEE
Transactions on System, Man, and Cybernetics—Part A: Systems and Humans 20 (4), 411–420.
Brewer, M.B., Silver, M., 1978. In-group bias as a function of task characteristics. European Journal of
Social Psychology 8, 393–400.
Chau, Y.K., 1996. An empirical assessment of a modified technology acceptance model. Journal of
Management Information Systems 13 (2), 185–204.
Chau, Y.K., Hu, J.H., 2001. Information technology acceptance by individual professionals: a model
comparison approach. Decision Sciences 32 (4), 699–719.
Chen, L.-D., Gillenson, M.L., Sherrell, D.L., 2002. Enticing online consumers: an extended technology
acceptance perspective. Information & Management 39, 705–719.
Coleman, J.S., 1990. Foundation of Social Theory. Harvard University Press, Cambridge, MA.
Davis, F.D., Bagozzi, R.P., Warshaw, P.R., 1989. User acceptance of computer technology: a comparison
of two theoretical models. Management Science 35, 982–1002.
Doney, P.M., Cannon, J.P., Mullen, M.R., 1998. Understanding the influence of national culture on the
development of trust. Academy of Management Review 23 (3), 601–620.
Fishbein, M., Ajzen, I., 1975. Belief, Attitude, Intention, and Behavior: An Introduction to Theory and
Research. Addison-Wesley, Reading, MA.
Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with unobservable variables and
measurement errors. Journal of Marketing Research 18, 39–50.
Ganesan, S., 1994. Determinants of long-Term orientation in buyer-seller relationships. Journal of
Marketing 58, 1–19.
Gefen, D., 2004. What makes an ERP implementation relationship worthwhile: linking trust mechanisms
and ERP usefulness. Journal of Management Information Systems 21 (1), 263–288.
Gefen, D., Karahanna, E., Straub, D., 2003a. Trust and TAM in online shopping: an integrated model.
MIS Quarterly 27 (1), 51–90.
Gefen, D., Karahanna, E., Straub, D., 2003b. Inexperience and experience with online stores: the
importance of TAM and Trust. IEEE Transactions on Engineering Management 50 (3), 307–321.
Grabner-Kraeuter, S., 2002. The role of consumers’ trust in online-shopping. Journal of Business Ethics
39, 43–50.
Henry, J.W., Stone, R.W., 1994. A structural equation model of end-user satisfaction with a computer-
based medical information systems. Information Resources Management Journal 7 (3), 21–33.
Hosmer, L.T., 1995. Trust: the connecting link between organizational theory and philosophical ethics.
Academy of Management Review 20 (2), 379–403.
Karahanna, E., Straub, D.W., Chervany, N.L., 1999. Information technology adoption across time: a
cross-sectional comparison of pre-adoption and post-adoption beliefs. MIS Quarterly 23 (2), 183–213.
Kelley, H.H., 1979. Personal Relationships: Their Structure and Processes. Lawrence Erlbaum Associates,
Mahwah, NJ.
Kelley, H.H., Thibaut, J.W., 1978. Interpersonal Relations: A Theory of Interdependence. Wiley, New York.
Kim, H.-W., Xu, Y., Koh, J., 2004. A comparison of online trust building factors between potential
customers and repeat customers. Journal of the Association for Information Systems 5 (10), 392–420.
Kwon, T.H., Zmud, R.W., 1987. Unifying the fragmented models of information systems implementation.
In: Boland, R.J., Hirschheim, R.A. (Eds.), Critical Issues in Information Systems Research. Wiley,
New York, pp. 227–251.
 ARTICLE IN PRESS
I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808 807

Lewicki, R.J., Bunker, B.B., 1995. Trust in relationships: a model of trust development and decline.
In: Bunkers, B.B., Rubin, J.Z. (Eds.), Conflict, Cooperation and Justice. Jossey Bass, San Francisco,
pp. 133–173.
Lewis, J.D., Weigert, A., 1985. Trust as a social reality. Social Forces 63 (4), 967–985.
Luhmann, N., 1979. Trust and Power. Wiley, Chichester, England.
Matsushima, R., Shiomi, K., 2003. Developing a scale of self-efficacy in personal relationships for
adolescents. Psychological Reports 92 (1).
McKnight, D.H., Chervany, N.L., 2002. What trust means in e-commerce customer relationships:
an interdisciplinary conceptual typology. International Journal of Electronic Commerce 6 (2),
35–72.
McKnight, D.H., Cummings, L.L., Chervany, N.L., 1998. Initial trust formation in new organizational
relationships. Academy of Management Review 23 (3), 472–490.
McKnight, D.H., Choudhury, V., Kacmar, C., 2002. Developing and validating trust measures for e-
Commerce: an integrative typology. Information Systems Research 13 (3), 334–359.
Meyerson, D., Weick, K.E., Kramer, R.M., 1996. Swift trust and temporary groups. In: Kramer, R.M.,
Tyler, T.R. (Eds.), Trust in Organizations: Frontiers of Theory and Research. Sage Publications,
Thousand Oaks, pp. 166–195.
Moore, G.C., Benbasat, I., 1991. Development of an instrument to measure the perception of adopting an
information technology innovation. Information System Research 2 (3), 192–222.
Nelson, K.M., Cooprider, J.G., 1996. The contribution of shared knowledge to IS group performance.
MIS Quarterly 20 (4), 409–432.
Pavlou, P.A., 2002. What drives electronic commerce? A theory of planned behavior perspective. Best
Paper Proceedings of the Academy of Management Conference, Denver, CO, pp. 9–14.
Pavlou, P.A., 2003. Consumer acceptance of electronic commerce—integrating trust and risk with the
technology acceptance model. International Journal of Electronic Commerce 7 (3), 69–103.
Pavlou, P.A., Gefen, D., 2004. Building effective online marketplaces with institution-based trust.
Information Systems Research 15 (1), 37–59.
Rogers, E.M., 1995. The Diffusion of Innovation, fourth ed. Free Press, New York.
Saeed, K.A., Hwang, Y., Yi, M.Y., 2003. Toward an integrative framework for online consumer behavior
research: a meta-analysis approach. Journal of End User Computing 15 (4), 1–26.
Shapiro, S.P., 1987. The social control of impersonal trust. American Journal of Sociology 93, 623–658.
Taylor, S., Todd, P.A., 1995. Understanding information technology usage: a test of competing models.
Information System Research 6 (2), 144–176.
Turban, E., King, D., Lee, J., Warkentin, M., Chung, H.M., 2002. Electronic Commerce: A Managerial
Perspective. Prentice-Hall, Upper Saddle River, NJ.
Venkatesh, V., Davis, F.D., 1996. A model of the antecedents of perceived ease of use: development and
test. Decision Sciences 27 (3), 451–481.
Venkatesh, V., Davis, F.D., 2000. A theoretical extension of the technology acceptance model: four
longitudinal field studies. Management Science 46 (2), 186–204.
Venkatesh, V., Morris, M.G., 2000. A longitudinal field investigation of gender difference in individual
technology adoption decision-making process. Organizational Behavior and Human Decision Process
83 (1), 33–60.
Wrightsman, L.S., 1972. Interpersonal trust and attitudes toward human nature. In: Zand, D.E. (Ed.),
Trust and Managerial Problem Solving. Administrative Science Quarterly 17, 229–239.
Zucker, L.G., 1986. Production of trust: institutional source of economic structure, 1840–1920. In: Staw,
B.M., Cummings, L.L. (Eds.), Research in Organizational Behavior. JAI Press, Greenwich, CT,
pp. 53–111.

Further reading

Teo, T.S.H., 2002. Attitudes toward online shopping and the Internet. Behaviour & Information
Technology 21 (4), 259–271.
 ARTICLE IN PRESS
808 I.-L. Wu, J.-L. Chen / Int. J. Human-Computer Studies 62 (2005) 784–808

Ing-Long Wu is a professor and chair in the Department of Information Management at National Chung
Cheng University. He gained a Bachelor in Industrial Management from National Cheng-Kung
University, an M.S. in Computer Science from Montclair State University, and a Ph.D. in Management
from Rutgers, the State University of New Jersey. He has published a number of papers in Information &
Management, Decision Support Systems, Behavior and Information Technology, Psychometrika, Applied
Psychological Measurement, and Journal of Educational and Behavioral Statistics. His current research
interests are in the areas of e-commerce, customer relationship management, supply chain management,
strategic information systems, and business process reengineering.