Download - s2 sembarang
-
8/11/2019 s2 sembarang
1/19
e-Government Adoption Model (GAM): Differing service maturity levels
Mahmud Akhter Shareefa, Vinod Kumar b, Uma Kumar b, Yogesh K. Dwivedi c,a DeGroote School of Business, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4M4, Canadab Sprott School of Business, Carleton University, 1125 Colonel By Drive, Ottawa, ON, K1S 5B6, Canadac School of Business and Economics, Swansea University, Singleton Park, Swansea, SA2 8PP, UK
a b s t r a c ta r t i c l e i n f o
Available online 20 October 2010
Keywords:
e-Government (e-Gov)
Information and communication technology
(ICT)
Adoption
Citizens
Service maturity levels
This research has as its objective the discovery of the critical factors that enable citizens to adopt e-
Government (e-Gov) at different stages of service maturity. To accomplish the objective, this research has
explained the related concepts and theories and developed a research framework grounded on a strong
theoretical and literature review background. The empirical study was conducted in Canada, which is a leader
in providing mature e-Gov services. From our results, we have observed two ontological differences from the
present literature in the adoption behavior of e-Gov where organizational and nancial perspectives have
distinct implications over parsimonious technology adoption behavior. First, technology adoption model
(TAM),diffusion of innovation theory(DOI),and theoryof planned behavior (TPB) cannotcapture andspecify
the complete essence of e-Gov adoption behavior of citizens. Second, e-Gov adoption behavior also differs
based on service maturity levels, i.e., when functional characteristics of organizational, technological,
economical, and social perspectives of e-Gov differ. Our ndings indicate the critical factors that enable
citizens to adopt e-Gov at different stages of service maturity. Public administrators and policy-makers have
potential implications from the ndings of the adoption behavior of e-Gov at different maturity levels.
Crown Copyright 2010 Published by Elsevier Inc. All rights reserved.
1. Introduction
As a new and rapidly growing eld, the concepts and theories of e-
Government (e-Gov) are still in a premature stage. Researchers from
different disciplines address this phenomenal theme from their
respective speculations and conceptualize it in a scattered fashion
(Heeks & Bailur, 2007). e-Gov has several aspects, including social,
technical, economic, political, and public administrative. However,
most dominating concepts of e-Gov arise from the technical
perspective and a combination of the socio-economic and public
administrative perspectives. Nevertheless, all the denitions are
headed towards a single notion and encompass a generic and unique
mission of e-Govpresenting government systems using information
and communication technology (ICT) to serve citizens better (Al-
Mashari, 2007; Evans & Yen, 2006; Gil-Garcia & Martinez-Moyano,
2007; Reddick, 2006; Shareef, Kumar, Kumar, & Dwivedi, 2009;
Sprecher, 2000).
Though different countries' e-Gov implementationsextensively differ
in setting common missions and objectives, all of them contain the
similar fundamental essence of e-Gov value: it should be citizen focused.
Therefore, it may be signicant to observe that the most important toolfor implementation of e-Gov is the willingness of citizens to adopt it
(Evans & Yen, 2006; Shareef et al., 2009). While there is evidence for
substantial growth, development, and diffusion of e-Gov universally, it is
not clear whether citizens of all developed and developing countries are
ready to embrace those services (Carter & Blanger, 2005). The
acceptance, diffusion, and success of e-Gov initiatives are contingent
upon citizens' willingness to adopt these services.
Reviewing the existing literature on e-Gov adoption by citizens
and business organizations (Al-Adawi, Yousafzai, & Pallister, 2005;
Chen & Thurmaier, 2005; Ebrahim & Irani, 2005; Gilbert, Balestrini, &
Littleboy, 2004; Klievink & Janssen, 2009; Kumar, Mukerji, Butt, &
Persaud, 2007; Phang, Sutanto, Li, & Kankanhalli, 2005; Reddick, 2004;
Sakowicz, 2007; Schedler & Summermatter, 2007; Shareef et al., 2009;
Tung & Rieck, 2005; Wang & Liao, 2008), we can infer that the
adoption models offered so far in the academic literature are mainly
conceptual. Extensive empirical studies among the actual users to
validate and generalize the models are absent. Most of those who
have attempted to validate their models did not rigorously review the
literature and integrate discourses from technical, social, organiza-
tional, political, and cultural perspectives to develop their ontological
and epistemological paradigms of model validation doctrine. As
identied byHeeks and Bailur (2007)through an extensive literature
review of e-Gov, methodologically these models are not grounded on
a strong theoretical framework. While developing those models of
adoption, the generalization aspect is heavily ignored (Heeks & Bailur,
2007).
Government Information Quarterly 28 (2011) 17 35
Corresponding author. Fax: +44 1792 295626.
E-mail addresses:[email protected](M.A. Shareef),[email protected]
(Y.K. Dwivedi).
0740-624X/$ see front matter. Crown Copyright 2010 Published by Elsevier Inc. All rights reserved.
doi:10.1016/j.giq.2010.05.006
Contents lists available at ScienceDirect
Government Information Quarterly
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / g o v i n f
http://dx.doi.org/10.1016/j.giq.2010.05.006http://dx.doi.org/10.1016/j.giq.2010.05.006http://dx.doi.org/10.1016/j.giq.2010.05.006mailto:[email protected]:[email protected]://dx.doi.org/10.1016/j.giq.2010.05.006http://www.sciencedirect.com/science/journal/0740624Xhttp://www.sciencedirect.com/science/journal/0740624Xhttp://dx.doi.org/10.1016/j.giq.2010.05.006mailto:[email protected]:[email protected]://dx.doi.org/10.1016/j.giq.2010.05.006 -
8/11/2019 s2 sembarang
2/19
Despite the potentially signicant impacts of e-Gov systems on
public administrations, organizations, individuals, and society, so far
only a few systematic and thorough studies have been undertaken on
the subject to comprehensively integrate overall factors related to the
successful implementation of e-Gov (Jaeger, 2003; Kraemer & King,
2003). However, citizens' behavior in terms of adopting a new
technology-driven system is a very complex and robust subject. It is
expected that extensive research will focus on criteria necessary for
citizens to adopt technology that will enable successful implementa-tion of e-Gov. Understanding and estimating the effect of citizens'
adopting criteria, which leads to successful implementation of e-Gov,
would have important managerial implications. Therefore, this
research attempts to investigate the users' requirements for the
adoption of e-Gov and sets the rst objective:
1. To identify and conceptualize the critical factors that affect citizens
to adopt e-Government.
Implementation and successive upgrading of the e-Gov system
follow certain paths, levels of maturity, stages, or phases. Different
countries implementing e-Gov in their ICT framework certainly have
different missions and objectives; however, the gradual development
of an e-Gov system in any country follows some unique levels of
service maturity for evolution. Each of the service levels represents a
different service pattern, different levels of technological sophistica-
tion, different stakeholder orientation, different types of interaction,
different security requirements, and different reengineering processes
(Holden, Norris, & Fletcher, 2003; Moon, 2002; Dorner, 2009). It can
also be inferred that these levels describe the development of
maturity of service in a sequential manner.
Based on the conceptualization of service development stagesof e-
Gov by different researchers (Accenture, 2003; Andersen & Henriksen,
2006; Evans & Yen, 2006; Fang, 2002; Klievink & Janssen, 2009; Layne
& Lee, 2001), we dene levels of service maturity of e-Gov as the
pattern of service that a government develops, successively enhances
interactivity, and delivers for stakeholders' acceptance and usage with
upgrading of technological sophistication and functional characteristics.
Since this research is engaged in developing adoption concepts ofe-Gov by citizens at different levels of service maturity, we will put
more attention into the rst two functionally different levels of
maturity of service development of e-Gov: the static stage and the
interaction stage (Blanger & Carter, 2005; Chandler & Emanuels,
2002; Howard, 2001). The reason behind this is that these two levels
are widely developed in most of the countries. The third level is
described as transaction stage.However,most of thecountries arestill
struggling to attain this e-Gov service level, so this stage is not
considered for developing any comprehensive model. The next stages
of service maturity, such as vertical integration and horizontal
integration, are also not very important for this research, as these
stages are not fully achieved by most of the countries so far. Most
countries have failed to realize horizontal stage of e-Gov universally
across all public services in their countries.From the end users' perspectives, the two stages of services have
signicant differences in characteristics and functionality (Gottschalk,
2009). In the publishing or static stage, stakeholders can only view and
collect government information or download some forms and publica-
tions. This is one-way communication. Here the user cannot commu-
nicate with the government service system through this interface and
the government authority does not respond to the user electronically
(Accenture, 2005). In the next maturitylevel of servicethe interaction
phasetwo-way communication is established. Through the govern-
ment web page, at this stage, stakeholders can contact service providers
to resolve any issues in different electronic ways, such as sending e-
mails, using chat-room, etc. (Accenture, 2005).
However, differentiating and dening these two stages as gradual
service maturity of e-Gov do not mean that citizens useor adopt these
stages of e-Gov sequentially, i.e., rst static level and then interact
level. They can simply skip any beginning level and start adopting e-
Gov from the next matured level. It can be predicted that the various
development levels of e-Gov might differ in pursuing the intention to
adopt e-Gov for its successful implementation. Static and interaction
levels especially offer different modes of service with different levels
of association of technology. As a result, adoption criteria for different
stages by citizens might have signicant implications. However, no
literature so far has investigated these criteria while exploringadoption models for e-Gov. To investigate the users' requirements
for the adoption of e-Gov at different levels of service maturity (not
like Layne and Lee, 2001 who look at organizational growth), this
research paper sets its second objective as:
2. Arethese critical factorsthat affectcitizens' adoptionof e-Government
different at different levels of service maturity?
2. Design perspective
While investigating and revealing the theoretical perspectives of
independent variables of e-Gov adoption, this research explores two
major elds as theoretical background. The rst attempt is engaged in
reviewing the surrounding areas of literature addressing e-Gov
adoption, implementation, characteristics, and related issues. The
second attempt is the extensive cultivation of theories related to
technology adoption, public administration and organization, psy-
chology, sociology, political science, culture, and marketing. Several
researchers who have done a comprehensive literature synthesis on
different e-Gov issuesincluding e-Gov adoption by end users
asserted that e-Gov implementation and adoption concepts have a
signicant lack of theoretical synopsis (Heeks & Bailur, 2007; Titah &
Barki, 2005; Wang & Liao, 2008).
Different researchers (Al-Shehry, Rogerson, Fairweather, & Prior,
2006; Chen & Thurmaier, 2005; Kumar et al., 2007) emphasized that
e-Gov adoption is more than a technological matter as it is inuenced
by many factors, including organizational, human, economic, social,
and cultural issues. These perspectives provide important specula-
tions for analyzing the e-Gov structure that reects governmentnatureand its responsibility in society (Carter& Blanger,2004;Moon
& Norris, 2005). In addition, the adoption of e-Gov systems requires
analyzing the changes in social values over time (Ebrahim & Irani,
2005).Steyaert (2004)adopted a marketing perspective to analyze e-
Gov performance. He proposed an e-Commerce (EC)-based perfor-
mance model to evaluate e-Gov performance in terms of citizen
satisfaction.Parent, Vandebeek, and Gemino (2005) and Warkentin,
Gefen, Pavlou, and Rose (2002)investigated the effect of trust on the
adoption of e-Gov. Gilbert et al. (2004)proposed theintegrationof the
service quality, technology, and behavioral aspects of the e-Gov
adoption framework.Shareef, Kumar, and Kumar (2007) and Shareef
et al. (2009) investigated technological, behavioral, economic, and
service quality aspects of e-Gov adoption criteria in formulating and
validating a framework of e-Gov adoption. Titah and Barki (2005) andPhang et al. (2005) reviewed the adoption literature of e-Gov
extensively and they suggested that technological, organizational,
social, cultural, behavioral, and economic aspects should be consid-
ered in a comprehensive framework of e-Gov adoption. Therefore,
from our literature review, we perceive that technological, behavioral,
social, cultural, organizational, economic, political, and marketing
aspects might provide important insights while investigating explan-
atory variables for e-Gov adoption.
To delineate the theoretical paradigms, this research looks at the
core characteristics of e-Gov. We nd that e-Gov is affected by several
issues like, organizational reformation, cultural revolution, habitual
change, technology adoption, information and service modication,
speed of service, accessibility, availability, more participation, trans-
parency, cost effectiveness, democratization, and globalization (Evans
18 M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
3/19
& Yen, 2006; Kim, Kim, & Lee, 2009; Robin, Andrew, & Sasha, 2009;
Titah & Barki, 2005; Turner & Desloges, 2002; Wang & Liao, 2008).
Since e-Gov is a revolutionary reformation of organizational structure
and characteristics, its adoption might have close ties with organiza-
tional attributes. e-Gov offers enormous benets to its end users,
which include economic incentives and service improvement.
Therefore, marketing and economic behavior reects citizens'
preferences in adopting e-Gov. Transaction cost analysis (TCA) also
sheds light on these perspectives of e-Gov adoption criteria. From thecore principle of TCA, the motivation for behavioral intention to
interact with different organizational structures is signicantly
inuenced by economic parameters (Shelanski & Klein, 1995).
According to the theory of planned behavior (TPB) ( Ajzen, 1991)
and the theory of reasoned action (TRA) (Ajzen & Fishbein, 1980),
social and cultural values affect beliefs and attitudes and the adoption
of e-Gov operated through ICT. Beliefs and attitudes about e-Gov lead
to formation of behavioral intention to learn, accept, and use e-Gov
systems. Therefore, behavioral or attitudinal aspects of citizens are
very important in stimulating an adoption framework of e-Gov. If we
translate the core doctrine of socio-technical theory, which explains
the effect of social and technological aspects on a system, we get
thorough insights into integrating the social, organizational, and
technological aspects of the e-Gov adoption (Damodaran, Nicholls, &
Henney, 2005).
However, these perspectives, where we have concentrated our
investigation of critical factors for the adoption framework of e-Gov,
are not mutually exclusive phenomena. These are interrelated issues.
Adoption perspectives of e-Gov by different stakeholders at different
levels of service maturity of e-Gov are intertwined with different
explanatory variables. Therefore, our investigation for identifying
critical factors of e-Gov will not track those perspectives of e-Gov
adoption factors separately; rather, we will look for interdependent
and comprehensive effects. We will connect technological, behavioral,
social, cultural, organizational, economic, political, and marketing
aspects of consumers to develop a comprehensive e-Gov adoption
model.
3. Theoretical framework
As we mentioned in the Introduction section, the existing
literature on e-Gov has failed to present a comprehensive framework
of e-Gov adoption and performance at different phases of service
maturity of e-Gov implementation. Therefore, although, we are
attempting to identify allof the constructs from our detailed literature
review in conjunction with the insight from different theories related
to technology adoption, diffusion, and behavioral, social, and cultural
characteristics, the adoption behavior of e-Gov is in a very premature
stage (Heeks & Bailur, 2007). Consequently, this study has potentially
an exploratory nature. It means, we are conducting this research not
to test any specied theory of e-Gov adoption, rather we are
conducting this research in the hope of developing a theory of the
adoption of e-Gov at different service maturity levels. So, as we areadvancing from theory development to statistical analysis, we should
continue to rene our exogenous variables and also hypotheses to
develop our nal paradigms of adopting e-Gov at different service
maturity levels. For an exploratory study, this renement of variables
and hypotheses is typical and also a part of the theory development
process (Stevens, 1996).
3.1. Explanatory variables
According to information management principles for open
government adoption, a prime factor for adoption is creating
awareness among the stakeholders. This means informing the
citizens about the transformation of public administration, imple-
mentation of innovation, basic paradigms of the new system,
application of ICT, objectives and mission of e-Gov development,
comprehensive information about relative advantages and disad-
vantages of e-Gov, and the overall credibility of the system. A long
history of government service shows that citizens and business
organizations are traditionally habituated to use brick and mortar
government services for information collection, interaction, and all
types of transactions thatare basically operatedofine. Thehistoryof
e-Gov evolution is very new. Basically, most of the countries are just
at the beginning efforts of implementing e-Gov. Therefore, stake-holders are still not very aware of this new innovation of the
government system. As we learned from the TPB and TRA, beliefs
about a system turn to the attitude of using the system. However,
awareness of the system is important at the beginning to develop
beliefs (Limayem, Hirt, & Cheung,2007).
Before developing an attitude to adopt e-Gov, stakeholders need to
be aware of its complete characteristics, including the background of
the system, functional behavior, strategic benets, the safety and legal
environment, etc. Awareness of e-Gov has several different aspects:
political, marketing, behavioral, and social. Whencitizens are aware of
the political agenda of e-Gov, social values related to the strategic
implementation of e-Gov, service quality, and the competitive
advantages of e-Govi.e., the marketing paradigm and attitudinal or
behavioral motivation of the EG systemthey might then have an
intention to adopt the e-Gov system. Several researchers asserted
awareness as the signicant independent variable to create the
attitude to use an e-Gov system (Eggers, 2004; Parent et al., 2005). So
we categorize this predicted variable for adoption as the Attitudeto
use, since awareness is the primary stimulus of creating attitude.
Depending on the previous arguments, this research proposes:
H1. Perceived awareness (PA) has a positive relation with the
Adoption of e-Gov.
From TPB, diffusion of innovation theory (DOI), and TCA, a user will
not arrive at an intention to use an EG system, which requires computer
knowledge to get a competitive advantage, unless the user has
competence from experience in the use of modern ICT. From
technological, behavioral, economic, and organizational perspectives,it is anticipated that failing to get hands-on experience of technology
will not create in the user an attitude favorable to adopting the system.
Also, in the absence of computer knowledge, a user cannot perceive the
economic advantages of e-Gov. The organizational structure of e-Gov is
computer- and internet-based, while from the viewpoint of end users,
traditional government services do not require computer knowledge.
Therefore,fromorganizational perspectives, computer self-efcacyisan
important predictor of whether a user will adopt an e-Gov system
instead of using traditional government services.
Several researchers exploring the barriers of adoption of e-
Commerce and e-Gov revealed that users' computer self-efcacy
and experience of theinternet, ICT, andcomputers create a perception
of security in theusers' attitude toward using onlinesystems (Moon &
Norris, 2005; Tung & Rieck, 2005; van Dijk, Peters, & Ebbers, 2008)that affects their intention to use. Wang (2002) investigated the
relation of technology availability and computer self-efcacy with
behavioral intentionto adopt an onlinetax ling system and observed
a positive relation. Computer self-efcacy is concerned not with the
skills one has but with judgments of what one can do with whatever
skills one possesses. Therefore, we categorize this explanatory
variable of e-Gov adoption as the Attitudeto use.
Including others, the most important and dominating barriers for
adopting e-Gov, particularly in developing countries are scarcity of
electricity, telephones, computers, the internet, and related acces-
sories, and government supports like call-center, resource-center, and
cyber-cafs. If we examine the capability theory (Nussbaum & Sen,
1993), it highlights that citizens' use of e-Gov systems is limited
unless they have the freedom of utility required to use modern ICT-
19M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
4/19
based e-Gov. If a country cannot make the skills and resources
required for using e-Gov available to all citizens equally, the country
cannot expect the same capability from all citizens to adopt that
system. Therefore, without reducing the digital divide, promoting
equality in resources of using e-Gov and making available all
components of an e-Gov system and knowledge, the adoption of
e-Gov will not be successful. The availability of resources required for
the use of e-Gov has behavioral, economic, cultural, social, and
technological aspects. Generally, where computers, internet, andmodern ICT are not available, the citizens are economically poor, less
educated, unaware of modern technology, socially and culturally
unfamiliar with modern technology, and lack the necessary skills to
use technology. As a result, they also do not believe that they will
receive benets by using an e-Gov system. Therefore, there is an
obvious relation between availability of resources and the adoption of
e-Gov (van Dijk et al., 2008). We argue that AOR creates a belief in
using an e-Gov system operated through ICT, which, in turn, creates
an attitude to use e-Gov. Therefore, we categorize the construct AOR
as the Attitudeto use e-Gov. Drawing a conclusion from the above
arguments, we propose here:
H2. Computer-self efcacy (CSE) has a positive relation with
Adoption of EG.
H3. Availability of resources (AOR) has a positive relation with
Adoption Of EG.
The technology acceptance model (TAM) proposes that perceived
ease of use (PEOU) and perceived usefulness (PU) determine the
attitude towardadoption of ICT. This behavioral attitude,in turn, leads
to the intention to use ICT and the nal acceptance of the system
(Bhattacherjee, 2001; Davis, Bagozzi, & Warshaw, 1989; Lucas &
Spitler, 1999; Moon, 2002; Venkatesh, 2000). e-Gov fails if the users
do not have the ability to use the technology to access useful
information and services, and eventually do not perceive e-Gov as
useful. This would lead to a non-acceptance of the system by citizens
(Shareef et al., 2007). Numerous scholarly articles (Evans & Yen, 2006;
Gil-Garcia & Martinez-Moyano, 2007; Shareef et al., 2009) revealedthat PEOU and PU are potential indicators of user acceptance,
adoption, and motivation to use web services. However, some
researchers (Carter & Blanger, 2004) did not nd any signicant
relation between the adoption of e-Gov and PU. A plausible
explanation lies in the logic that inclusion of relative advantage as a
predicted variable in the adoption model explains enough of PU in the
adoption construct. So, at this stage, we will provide further insight
into PU and relative advantage for inclusion in the adoption model as
the predictor variables.
According to the DOI theory (Rogers, 1995), the rate of diffusion is
affected by an innovation's relative advantage, complexity, compati-
bility, trialability, and observability. Literature reviews suggest that
among those ve constructs, relative advantage, compatibility, and
complexity are the most relevant constructs to determine the adoptioncharacteristics of technology innovation (Gilbert et al., 2004; Moore &
Benbasat, 1991; Rogers, 1995; Tornatzky & Klein, 1982). These authors,
especially Tornatzky and Klein (1982)in their meta-analysis of research
on the adoption of innovations, argue that trialability and observability
are not related constructs for technology adoption. Therefore, in this
study we are not considering these two constructs.
Complexity, comparable to TAM's PEOU construct, captures the
perception of some pre-use complexities that seem to have very close
relation to the perceptions of complexities in using modern ICT,
internet, and computers. Therefore, we comprehend the generic
essence of the construct complexity with PEOU and introduce
perceived ability to use (PATU) as the predictor of adoption to reect
other aspects of e-Gov above and beyond technology. The construct
PATU has technological and organizational perspectives. Due to
revolutionary reengineering of the traditional government system,
the perception of online organizational structure, which is apparently
new, is an important aspect of the perceived ability to use the system.
Technology is a very predictable aspect to get insight into PATU. We
argue that PATU reects the ability of citizens to use an e-Gov system
and categorize it as the Abilityto use for e-Gov adoption.
Compatibility construct has cultural, behavioral,and socialaspects. It
is dependent both on individual characteristics such as avoiding
personal interaction, and social in
uence. Several researchers indicatedthat specic characteristic of e-Gov that allow citizens to avoid personal
interaction might create the perception of compatibility among citizens
to adopt an e-Gov system (Gilbert et al., 2004). Shedding light on the
TPB, TRA, and capability theory, the compatibility of an e-Gov system
with adopters' beliefs, values, and attitudes reects the behavioral
aspect. From the socio-technical and complementary theories, beliefs
and attitudes of adopters of a new technology system also have social
and cultural aspects. Several researchers use this construct as a
signicant predictor of EG adoption (Carter & Blanger, 2004; Chen &
Thurmaier, 2005; Shareef et al., 2007). This research argues that PC
creates citizens'attitudesto usean e-Gov systemand thus we categorize
this construct as the Attitudeto use.
Relative advantage captures the gain from receiving services and
information through e-Gov systems in comparison with that from
traditional government ofces. This denition expands the limited
concept of PU of the TAM, which captures only the absolute benets
from job performance. However, by adopting e-Gov systems, a user
can gain a lot of relative and absolute benets ranging from
effectiveness, efciency, availability, accessibility from anywhere,
comfort in use, time savings, cost savings, and convenience. The
combined effects of the two constructs, therefore, basically capture
the essence of absolute and comparable functional benets of the e-
Gov system. If we integrate both views of PU and relative advantage,
we can introduce perceived functional benet (PFB) as the predicted
variable of e-Gov adoption. This construct has economic, organiza-
tional, marketing, behavioral, and social perspectives. If we look at the
TCA, we can understand that PFB also captures the essence of time
efciency and price savings. Several researchers assumed that this
time constraint characteristic is also typical in EG because it is rationalto assume that citizens might adopt e-Gov systems as they save time
to perform tasks relative to the functions of a traditional paper-based
government ofce (Carter & Blanger, 2005; Gilbert et al., 2004;
Wagner, Cheung, Lee, & Ip, 2003). Another construct, price savings,
which is a measure of e-Gov efciency in terms of reduction in service
rendering cost, is also an overlapping concept of PFB (Tung & Rieck,
2005). This analysis shows that PFB captures the essences of the
behavioral, economic, and marketing aspects of adoption. It also has
an organizational aspect, because institutional theory asserts that
actors will accept organizational change as long as they perceive it to
be benecialto them(Lawrence & Suddaby, 2006). We argue that PFB
imparts reasons to use an e-Gov system, instead of using a traditional
government system, and, thus, it is categorized in this research as the
Adherence to adopt the use of e-Gov. We investigate the threedimensions PATU, PC, and PFB here since we assume that these might
affect EG adoption. Therefore, we propose:
H4. Perceived ability to use (PATU) has a positive relation with the
Adoption of e-Gov.
H5. Perceived compatibility (PC) has a positive relation with the
Adoption of e-Gov.
H6. Perceived functional benet (PFB) has a positive relation with the
Adoption of e-Gov.
In addition, we also include a new construct, Image, as proposed
by Moore and Benbasat (1991). According to the DOI, Image
20 M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
5/19
inuences the acceptance and use of an innovation. Image refers to
citizens' perceptions that adopting e-Gov makes them superior to
others in the society. Interaction with e-Gov systems, instead of
using traditional government ofces, is perceived to give these
citizens superior status. Several researchers have, therefore, includ-
ed this construct in their proposed model of EG adoption (Gilbert et
al., 2004; Phang et al., 2005; Tung & Rieck, 2005). Since the adoption
of e-Gov might reect the adopters familiarity with modern
technology, higher level of education, competence in using compu-ters and the internet, and perception of modernism, these phenom-
ena impart some degree of social values and prestige to adopters.
Therefore, this research argues that Image has social, behavioral, and
also cultural aspects and depends on an individual's personal
behavioral ideology. We also argue that the construct perception of
Image is a proponent of reasoning to use e-Gov and, thus, it is
categorized in this research as the Adherence to use for e-Gov
adoption. Therefore, we propose:
H7. Perceived image (PI) of using e-Gov has a positive relation with
the Adoption of e-Gov.
Content, organization, and presentation of informationi.e., infor-
mation quality, which includes accuracy, current information, relevan-
cy, fulllment, linkage, completeness, integration, organization,
timelinesare potential contributors to creating a perception of
reliability that inuences citizens to accept e-Gov (Collier & Bienstock,
2006; Kim, Kim, & Lennon, 2006; Kumar et al., 2007; Parasuraman,
Zeithaml, & Malhotra, 2005; Sebastianelli, Tamimi, & Rajan, 2006). The
assurance and condence in using e-Gov that information quality gives
citizens is characterized in this research as the Assurance tousefor e-
Gov adoption. Gilbert et al. (2004) conducted a survey in theUK among
users of e-Gov thatshowed thatinformationquality is a strongpredictor
of EG adoption. Another extensive empirical study conducted in
Australia (AGIMO, 2003) revealed that there is an obvious expectation
that information from government will be provided in accordance to
fullling citizen needs rather than serving the convenience of
government agencies. In order to ensure the success of an IS, Delone
and McLean (1992, 2003) proposed the Information System SuccessModel (IS Model). The model asserted that information quality is the
determinant of system use and user satisfaction that eventually leads to
regularadoption (Wang & Liao, 2008; Wangpipatwong, Chutimaskul, &
Papasratorn,2005). Information quality asserts service requirements of
consumers and the economic benets of viewing and collecting
information from a website in lieu of verbal interaction from traditional
government ofces. Therefore, information quality has a marketing
aspect.Lin and Lu (2000)asserted the conjecture that the features and
accuracy of information posted on a website signicantly affect users'
behavioral attitude. Grounded on the aforementioned arguments, we
propose:
H8. Perceived information quality (PIQ) has a positive relation with
Adoption of EG.
Several e-Gov researchers address customer service as one of the
important explanatory variables to satisfy customers and, thus, to
ensure a recurring use of e-Gov (Lee & Rao, 2003; Shareef et al., 2007;
Wangpipatwong et al., 2005). Service quality is a strong predictor to
differentiate performance of different organizations. From the
behavioral point of view, recurring users of e-Gov will achieve beliefs
and thus attitudes to adopt e-Gov if they perceive higher customer
service in e-Gov. Traditional government service has a different
approach than e-Gov. If citizens perceive a higher level of customer
service in e-Gov than that offered in a traditional government ofce,
they will purse the adoption of e-Gov. Kumar et al. (2007)proposed
that service quality leads to satisfaction that ensures regular use of e-
Gov. Service quality can be considered from different dimensions, like
technical or output quality, functional or process quality (Czepiel,
Solomon, Surprenant & Gutman, 1985; Grnroos, 1984; Lehtinen &
Lehtinen, 1982), and direct customer service from employees
(Shareef, Kumar, & Kumar, 2008). Since technical or output quality
is already being assessed in PATU and PIQand functional quality of the
system is incorporated in PFB, we will only discuss here the customer
service response to service quality. The reliability and assurance of
servicewhich are intertwined with trust, security, privacy, and risk
concepts of customer service
will also be discussed in a differentsection.
Due to the absence of any physical presentation, the service
response of e-Gov has different aspects and properties. In e-Gov, the
service response is generally assumed to be a recovery quality item.
When there are problems or concerns, stakeholders always expect
that customer service will resolve the problem promptly with
complete sympathy. If citizens feel that they do not nd any customer
service in e-Gov when they require it, that they are being treated
unfairly, or that the customer-oriented service policy of government
websites is not credible, they are less likely to adopt e-Gov; rather
they will go to a physical government ofce to seek services. A study
in a developing country conducted byShareef et al. (2009)also found
that service response has a signicant effect on citizens'adoption of e-
Gov. We argue that perceived service response (PSR) stimulates
citizens' adaptability and satisfaction in using e-Gov and thus ensures
recurring use. Therefore, this research categorizes the PSR construct as
the Adaptability to use for e-Gov adoption. Based on the above
arguments, we propose here:
H9. Perceived service response (PSR) has a positive relation with
Adoption of EG.
In an online transaction, since different physical cues are absent,
virtual transaction requires some extra facilities to perform transac-
tions for individuals withdifferent ethnic backgrounds. This criterion
is especially very important for a country that comprises multicul-
tural and multilingual groups. Additionally, e-Gov also has a global
aspect. Considering these aspects, the multilingual option of e-Gov
might enhance the adoption of e-Gov. The service quality research ofEC also nds this factor to be an important cause for consumers to
adopt a certain e-retailer'swebsite (Collier & Bienstock, 2006; Kim et
al., 2006; Parasuraman et al., 2005; Wolnbarger & Gilly, 2003).
Nantel, Sncal, and Mekki-Berrada (2005) conducted an empirical
study to determine critical factors of online purchase. Their research
captured a denite relation between the use of native language of a
user in a website as the medium of instruction and information and
adoptionof thewebsiteby thatuser. If an individual can interact with
a website using his/her primary language, he/she might feel more
cultural connection and have a more positive attitude to use of that
website. Through an extensive content analysis of 93 websites from
local companies in China, India, Japan, and the USA, Singh, Zhao, and
Hu (2005) conrmedthis statement. For citizens with less education,
a single language option other than mother tongue for viewing,collecting, interacting, and transacting with e-Gov websites might
create a signicant barrier. This is an important cultural aspect
(Michon & Chebat, 2004). If we borrow the speculations from the
TAMand DOI, the relativeadvantage of e-Gov will triggerinclusion of
a multilingual option in e-Gov web pages. However, beliefs in
competence with an e-Gov interactionwhich further promotes
attitude towards the adoption of e-Gov according to the TPBcan be
exaggerated if a known language option prevails in e-Gov websites.
Therefore, the multilingual option has behavioral aspects (Foo, Hui,
Leong, & Liu, 2000). If we look at the capability theory, the
multilingual option in e-Gov might create equal and competitive
capability, which ultimately enhances the economic power of
minority stakeholders. Therefore, from an economic perspective, a
multilingual option in e-Gov can create a level playingeld formajor
21M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
6/19
stakeholder groups with a multilingual background. From the
marketing, technological, behavioral, and cultural perspectives, this
research argues that a multilingual option in e-Gov might enhance
the processing and understanding capability of e-Gov. Therefore,
since this criterion promotes the ability to use e-Gov by improving
service quality, we categorize this plausible explanatory variable of
e-Gov adoption as the Ability of stakeholders. This research thus
proposes:
H10. Multilingual option (MLO) has a positive relation with the
Adoption of e-Gov.
Many scholarly articles conducting research in e-Gov adoption
have shown that security, privacy, uncertainty, and risk are predom-
inant factors for adoption (Al-Adawi et al., 2005; Parent et al., 2005;
Shareef et al., 2007, 2008; Welch & Pandey, 2005).Blanger, Hiller,
and Smith (2002) found that pleasure, privacy, security, and web
features are matters related to the perceived trustworthiness of a
website. Research in e-Commerce and e-Gov found that uncertainty,
security, privacy, and risk are all antecedents of perceived trust ( Al-
Adawi et al., 2005; Balasubramanian, Konana, & Menon, 2003; Parent
et al., 2005; Soat, 2003).
Perceived security is crucial to users' condence regarding the
safety of a website. Based on previous research on security in e-
Commerce (Blanger & Carter, 2005; Schaupp & Blanger, 2005), this
current study visualizes perceived security as the protection of
customers from any type of nancial or non-nancial risk during
transactions on websites, such as any type of identity thefts including
abuse of credit card, overcharging, non-payment, etc. These security
factors are potential contributors in developing trust among citizens
as the authentication of e-Gov nancial transaction and protection of
disposed information in e-Gov. Therefore, we conjecture that
perceived security (PS) has a causal effect on perceived trust (PT).
Since transactions in e-Gov are basically virtual, no actual physical
transaction takes place during interaction and payment by the clients,
and the uncertainty construct of TCA can be a potential factor of non-
accepting virtual environment of e-Gov. This, in turn, is related to
perceived trust in e-Gov (Al-Adawi et al., 2005). In the virtualenvironment in particular, being able to place trust in a website is
critical to consumers' successful interaction with the system ( Gefen,
Karahanna, & Straub, 2003).Cox and Rich (1964)delineate perceived
risk as the perception of uncertainty in a particular interaction
situation that has a close relation to trust. Therefore, PT is dependent
on PU (uncertainty).
In e-Gov, citizens provide written information in technology
interface while interacting or receiving/paying through e-Gov web-
sites. As a result, users of e-Gov always feel a lack of privacy. Several
researchers (Angst & Agarwal, 2009; Shareef et al., 2008; Yoo
& Donthu, 2001), who conducted empirical studies regarding the
acceptance of the online environmente.g., e-Commerce and e-Gov
observed that perceived privacy is a major concern for internet
customers during interaction with websites. Customers are afraid thatwebsites can disclose, share, or misuse their personal information or
that hackers can intercept their secret information (Brown & Muchira,
2004; Ranganathan & Ganapathy, 2002). During interaction with
websites, customers may perceive there to be a privacy risk
(Parasuraman et al., 2005). So, perceived privacy (PP) is also related
to the condence of users on the web, which nally indicates
trustworthiness. Trusting the web can enhance the perceived
privacy feeling of the customers during interaction in e-Gov ( Kemp,
2000).
Warkentin et al. (2002)extensively discussed the impact of trust in
e-Gov adoption, and consequently proposed that institutional-based
trust, characteristics-based trust, and process-based trustwhich all
together capture the essence of perceived security, privacy, and less
uncertainty in e-Gov
will lead to the intention to adopte-Gov. Thomas
(1998)outlined the three aspects by which trust in the government is
produced. The rst aspect is related to behavioral attitude that is
supported by TPB. The second aspect of trust derives from institutional
credibility. The third component captures the trusting attitude on
typical outcomes of a process where the seller sells goods to buyers and
buyers in return will pay for that. Nye and Zelikow (1997)classied
these causal factors as social, cultural, economic, and political. From the
marketingaspect, if a customer does nothave trust in theinstitution and
process, he/she might not embrace that organization for interaction. So,we can conclude that PT in e-Gov has political, behavioral, social,
organizational, technological, marketing, and cultural perspectives.
Trust is an important factor in analyzing the adoption behavior of
consumers in virtual environment, because citizens have few tangible
and veriable cues regarding the service provider's credibility and
performance (Urban, Sultan, & Qualls, 2000). In light of the above
discussion, we argue that PT creates condence in the overall e-Gov
performance and, thus, categorize this construct as the Assuranceto
use e-Gov. Thus, we propose:
H11. Perceived trust (PT) has a positive relation with the Adoption of
e-Gov.
H11a. Perceived uncertainty (PU) (uncertainty) has a negative
relation with Trust in e-Gov.
H11b. Perceived security (PS) has a positive relation with Trust in
e-Gov.
H11c. Perceived privacy (PP) has a positive relation with Trust in
e-Gov.
The measuring scale items for all the exogenous constructs of the
proposed model, developed and operationalized based on existing
literature on e-Gov, e-Commerce, IS, marketing, and expert opinions,
are shown in Appendix A. Thequestionnaire was pretested by a group
of people comprised of two scholarly researchers from Sprott School
of Business, Carleton University, Canada who have expertise in
analyzing the online adoption behavior, and eight PhD studentsfrom the social science and natural science departments of Carleton
University who have extensive knowledge in using Canadian e-Gov to
verify the structure, constructs, and respective measurement items of
the questionnaire. For any exploratory research, if the surveyor
happens to organize a group comprised of a couple of people, the
participating pretesting method is better (Converse & Presser, 1986).
In our pretesting procedure, we followed the participating method. A
structured questionnaire was used to measure the independent and
dependent variables of the study with a 5-point Likert scale ranging
from 1 (strongly disagree/never) to 5 (strongly agree/always). The 5-
point Likert scale is used to increase the response rate and response
quality along with reducing the respondent's frustration level
(Babakus & Mangold, 1992).
3.2. Dependent variables
According to the marketing theory, adoption of a new product
begins with consumer awareness, leads to an attitude toward that
product, then further advances to an intention to use as a trial basis,
and nally ends as full acceptance and regular use with satisfaction
(Pavlou & Fygenson, 2006). A new system can replace an old system.
So, if an individual using traditional government systems perceives e-
Gov to be more advantageous from any perspectives, he or she might
adopt the online government systemse-Gov. This research encom-
passes the adoption of e-Gov as a continuous process starting from
awareness of the system, beliefs of the system benets, attitude
toward using it, intention to use, actual use, satisfaction, and recurring
use. Some researchers (Al-Shehry et al., 2006; Chen & Thurmaier,
22 M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
7/19
2005; Kumar et al., 2007; Moon & Norris, 2005; Schedler &
Summermatter, 2007; Shareef et al., 2007, 2009; Wangpipatwong
et al., 2005) have developed their adoption models and measuring
items for the adoption construct by considering complete acceptance
of the process.
Depending on the paradigms of the e-Gov adoption process, we
nd logical underpinnings on thepremise that the adoption process of
e-Gov involves the frequent and recurrent use of online services by
citizens not only for obtaining information but also for interactionwith government. Adoption construct has behavioral, organizational,
economic, technological, political, marketing, social, and cultural
perspectives. We have mentioned that this research is concerned to
explore the objective by investigating adoption criteria into two
different levels of service maturity: Static or Publishing stage and
Interaction stage. Therefore, we have differentiated the dependent
variable Adoptioninto two sub-groups:
Adoption 1: Decision to accept and use an EG systemto view, collect
information, and/or download forms for different government
services as the user requires with the positive perception of
receiving a competitive advantage. Adoption 2: Decision to accept and use an EG system to interact
with, and seek government services, and/or search for queries for
different government services as the user requires with the positiveperception of receiving a competitive advantage.
The endogenous/dependent variable Adoption was operationa-
lized in a way that ensures measurement of the causal effects of the
exogenous variables on the two levels of service maturity of e-Gov
and increases response rate. It is obvious that citizens can view and
interact with e-Gov for many tasks. Dening any specic task for
adoption in the proposed questionnaire might reduce the response
rate in terms of adoption. Therefore, this study formulates the
instruments of Adoption not for any specic tasks but for general
tasks to keep the questionnaire general for all respondents. A total of
six scale items were selected to measure those two dimensions of the
adoption construct (AGIMO, 2003; Gil-Garcia & Martinez-Moyano,
2007; Murru, 2003; Sakowicz, 2007; Turner & Desloges, 2002) (scaleitems are shown in Appendix B). Based on these arguments and
identication, a model of e-Gov adoption (GAM) to investigate the
plausible relations is proposed inAppendix C.
4. Methodology
The research methodologies we use in this research are those
typically used in empirical business research. Based on the sugges-
tions ofHeeks andBailur (2007) about e-Gov research, andtheoriesof
Campbell and Fiske (1959) and Bagozzi, Yi, and Philips (1991) about
reliability and validity of research, we designed our research
methodology. In this research, the respondents are the users of the
Canadian e-Gov system; anyone who has experience using Canadian
e-Gov system could participate in the survey. This study was
conducted in four large cities in Ontario, Canada. We selected the
venue for the following purposes:
1. Canada is one of the leading countries in terms of offering e-Gov
services. Canada's e-Gov implementation and offered services are
very mature, and have different services in the static, interaction,
and transaction stages. Therefore, in terms of the development
stage of e-Gov and mission, vision, and objectives, Canada can be
viewed as one the most focused countries for e-Gov development
(Cardin, Holmes, Leganza, Hanson, & McEnroe, 2006).
2. The adoption rate (29.8%) and maturity of servicesof EG in Ontario is
the highest in Canada according to the study byParent et al. (2005).
Since this research hasset its objectives in detectingadoption criteria
of citizens at different maturity levels of services offered by e-Gov,
Ontario is assumed to fulll the research objectives.
3. The selected four cities are the most populated and largest cities in
the respective regions of Ontario. These cities are also located
strategically in important position and are prominent in multicul-
tural assembly. Therefore, it is assumed that the sample should
have enough variability.
To test the model in the most realistic way possible, the study was
conducted through a survey (a self-administered questionnaire) of a
broad diversity of citizens at several communities. From our previousexperience, we assumed that the study would receive around a 10%
response rate. Since there are 11 primary exogenous variables/
constructs, the number of response shouldbe at least 220(20 samples
per independent variable) for regression and factor analysis (Stevens,
1996, p. 143). However, a sample size of a minimum of 200 is good for
structural equation modeling (SEM) (Kline, 2005, p. 110). Therefore,
the questionnaire was distributed among 2200 citizens (or residents)
in thepreviously mentionedfour citiesin Ontario, Canada, to meet the
targetand fulll the statistical specications. The specic way weused
to distribute the questionnaire was:
1. We maintained roughly the population ratio of the four cities,
distributing 100 questionnaires in Sudbury, 200 in London, 500 in
Ottawa, and 1400 in Toronto.
2. We divided all the cities into ve regions: east, west, north, south,
and center.
3. We then collected addresses from the Telephone White Pages of
each city; we included houses, condominiums, and apartments
located in the ve regions we identied. We also collected the
addresses of the residents living in the suburban areas in the east,
west, north, and south regions immediately outside the city.
4. We distributed the questionnaires by mail throughout the
suburban areas in the east, west, north, and south regions outside
each city. Onehalf of thetotal questionnaires allocated foreach city
were distributed in this way. The other half was distributed
physically to the houses, condominiums, and apartments in
different areas in the ve zones.
5. We distributed 50% of the questionnaires in houses and condo-
miniums and 50% in apartments.6. The survey was conducted over a three-month period.
We received a total of 241 questionnaires from the respondents.
Two returned questionnaires were blank. Therefore, the eligible
response number is 239. The response rate is around 11%. This is quite
satisfactory based on our previous knowledge and also considering
the length of the questionnaireeight pages, including a one-page
cover letter.
5. Statistical analysis
Several interrelated procedures were performed to organize, re-
arrange, and summarize the raw data and make it amenable for
analysis to get justied output. This section sequentially describesdata preparation and analysis techniques for statistical analysis of this
research.
5.1. Data reduction
Before performing any causeeffect relation, reliability, validity,
and normality tests, we rst conducted exploratory factor analyses
(EFA) on the preliminary 57 scale items measuring the latent
variables having direct causal relations to the adoption of e-Gov
excluding the measuring items of the constructs perceived uncer-
tainty (PU), perceived privacy (PP), and perceived security (PS),
which are not hypothesized to have direct relation with adoption in
our model. These three constructs are hypothesized to have causal
relations with perceived trust (PT), which is an exogenous variable
23M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
8/19
for the adoption model. We have also done EFA on the 10 measuring
items of PU, PP, and PS separately because these three constructs are
widely used as the exogenous variables for PT (Shareef et al., 2008).
For EFA, we have used principal component analysis as the
extraction method and varimax rotation as the rotation method.
We used both the breaks-in-eigenvalues criterion (N1) and scree
plot to determine the number of factors to retain ( Stevens, 1996,
pp. 389390).
After conducting a series of EFAof those 57 measuring items, of the11 exogenous variables and also examining the correlation matrix we
found that nine constructs with 37 measuring items can be retained.
However, to support this renement in measuring items, we also
looked at the correlation matrix, analyzed convergence through CFA,
and thoroughly investigated theoretical aspects of those modica-
tions. For PA, AOR, CSE, PI, PIQ, MLO, and PFB constructs,
corresponding measuring items were loaded consistently (though
some items were removed because of low loading or cross loading).
Three PC items, three PATU items, and one PIQ item were loaded on a
single factor. Though the items loaded under this factor are the
measuring items from different hypothesized exogenous variables,
after close examination of those seven items loaded under a single
factor, we observed that all these items reect certain personal beliefs
and perceptions of ability of using e-Gov systems. This personal belief
denotes both physical and psychological perception (resembles to PC)
of ability to usee-Govsystems.Therefore, these seven items have very
close functional alignment. We veried the correlation between these
items andfoundmoderateto strongcorrelations. This also justiesthe
convergence of these items under a single factor. We also veried the
convergence of those seven measuring items in CFA by testing the
appropriateness of a single factor or two factor model (Appendix D).
Based on the functional meaning, we argued that PATU can still cause
those seven scale items and thus can be named PATU. However, the
denition of PATU should be edited adding new psychological
dimension. We have done that in Appendix E where all the
explanatory variables are dened. Four items of PT andve items of
PSR loaded under a single factor in EFA. We can explain this behavior
getting insight from the measuring items. On the one hand, perceived
trust of citizens is related to the credibility of e-Gov; on the otherhand, customer service of e-Gov also helps to enhance the perception
of trust and credibility, particularly in the virtual environment, among
the users of e-Gov. Therefore, we retained the name perceived trust
(PT) for the combined measuring items loaded under a single factor.
We also veried the correlation between these items and found
moderate to strong correlations. This also justies the convergence of
these items under a single factor. However, we also veried the
convergence of those nine measuring items in CFA by testing the
appropriateness of a single factor or two factor model (Appendix D).
We also veried all nine factors with the measuringitems individually
by CFA and observed conrmation of EFA results (Appendix D).
However, since 1 item of the construct AOR(AOR3) was loadedin CFA
with a loading factor of less than 0.50, we removed that item. In
addition, since we could retain only two items for CSE with highinternal correlations (more than 0.90) and two items for MLO with
high correlations (more than 0.95) from EFA, we could not perform
CFA for these two variables (since negative degree of freedom exists).
So, we took the average scores of the respective measuring items for
CSE and MLO respectively. Therefore, we retained a total of 34
measuring items with the nine exogenous variables(Appendix E).We
also retained two factors with nine indicators from the EFA of the 10
measuring items of PU, PS, and PP. However, four items of PS and two
items of PP were loaded under the same factor. Although as an
exploratory study we have hypothesized PS and PP as two different
exogenous variables for PT, several researchers used PS and PP as a
single construct by the name PS, because both the constructs are
related to security ofnancial transactions, identity, and personal in-
formation (Gummerus,Liljander, Pura, & van Riel, 2004; Wolnbarger
& Gilly, 2003). Therefore, we have provided the name of this construct
as PS, however, its denition was revised (Appendix E). We also
veried the correlation between these items and found moderate to
strong correlations. This also justies the convergence of these items
under a single factor. However, we also veried the convergence of
those six measuring items in CFA by testing the appropriateness of
a single factor or two factor model (Appendix D). So, for PT as
endogenous variable, we have retained two exogenous variables
namely, PU (uncertainty) and PS.The reliability scores for the constructs were measured by a
coefcient alpha, which justied the reliability of the items in each
dimension and thus internal consistency among the items in each
dimension. The reliability scores for all the nal exogenous and
endogenous variables are ranged from 0.706 to 0.974, which suggest
an acceptable internal consistency among the items in each
dimension (Nunnally & Bernstein, 1994). The CFA results suggest
that the scale items are reective indicators of their corresponding
latent constructs, which indicates construct validity (Chau, 1997;
Segars & Grover, 1993). In this data analysis, the average variances
extracted (AVE) for each factor and its measures all exceeded 0.50;
thus, convergent validity is achieved (Fornell & Larcker, 1981). We
also veried the correlation matrix of the items under each factor. All
the items individually under each factor have moderate to strong
correlation coefcients. This result also justied convergent validity.
We also examined multicolinearity, normality, and outliers.
5.2. Model testing: causal relationship by path analysis
We have used LISREL for path analysis, a family of structural
equation modeling (SEM), to test the causal relationships (the
hypotheses) of the model. Since we have measured all of the
exogenous and endogenous variables through Likert scale 1 5, data
gathered from this empirical study is not perfectly continuous.
Therefore, structural measurement through SEM by maximum
likelihood (ML) is not appropriate for this type of data (Kline, 2005,
p. 219). For structural measurement through SEM, one of the
fundamental requirements is that latent variables should be contin-
uous (Kline, 2005). Therefore, we took the average of the indicators ofeach of the latent variables individually for 239 cases and conducted a
path analysis (Kline, 2005) to nd out causeeffect relationships
between exogenous and endogenous variables. In path analysis, all of
the latent variables are treated as observed variables and their scores
represent the average of the scores of their respective indicators. We
have used the maximum likelihood procedure of LISREL for the
purpose of analysis. Since, the measuring items of PC and PATU were
integrated in a singleconstruct and themeasuring items of PT andPSR
were also integrated in a single item, we have now nine hypotheses to
test (with certain modications in the composition and denition of
PATU and PT constructs related to two hypotheses) from our
proposed 11 hypotheses having direct relations with adoption. Only
two hypotheses having direct relations with adoption were removed
during statistical renementin the previous section. Forpath analysis,we have used the correlation matrix as the input data for all the 11
exogenous variables (nine exogenous variableshaving direct relations
with adoption, i.e., PA, AOR, CSE, PATU, PFB, PI, MLO, PIQ, and PT and
two exogenous variables having indirect relations with adoption
through PT,i.e., PU and PS)and one endogenous variable (for example
ADOP1). Here PT is both an exogenous variable and endogenous
variable (like a mediator variable for adoption). Therefore, for two
models, i.e., GAM-S and GAM-I we have inputted two different
1212-correlation matrices. Final Path models and t indices are
shown inAppendices F and G. For path analysis we have tested 11
hypotheses, through path analysis for our research questions of this
research.
After conducting path analysis for the adoption of e-Gov at static
level, we nd that PA and PATU have signicant causal relations with
24 M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
9/19
ADOP1 with tvalues of 3.98 and 7.01 respectively. Therefore, these
two factors are signicant at the 0.05 level (z score for the 0.05 level is
1.96). Even these two factors are signicant at the 0.01 level (z score
for the 0.01 level is 2.576). PFB is signicant at the 0.1 level which has
a tvalue of 1.76 (z score for the 0.1 level is 1.645). PATU, PU, and PS
are signicant predictors on PT (Appendix F, Fig. 1).However,sincePT
is not a signicant predictor on ADOP1, subsequently PT and its
predictors are not related to adoption of EG. Apart from any issues
related to the adoption of e-Gov, separate causal relations of PT withother exogenous variablesare beyond thescope of this research.From
path analysis for adoption of EG at the interaction level, we nd that
PA, PI, PT, PIQ, and PATU have signicant causal relations with
ADOP2 at the 0.05 level. PATU, PU, and PS are signicant predictors on
PT (Appendix F, Fig.2). The nal accepted hypotheses for the adoption
of e-Gov at the static and interaction levels are listed inAppendix H.
Numerical formulation of the Path models for the static and
interaction levels are shown inAppendix I.
6. Discussion
As was previously mentioned, e-Gov implementation passes
through different phases of evolution as the services offered mature.
Though these phases are mutually exclusive and not distinctive,
however, as the services provided by e-Gov mature, its levels of
interaction improve from the static level to the interaction,
transaction, and integration levels. Especially in terms of technology,
organizational structure, service quality, reliability, security, and
privacy, the potential characteristics of the interaction level and
static level might be signicantly different. From the end users'
perspectives, thetwo stages of services havesignicant differences in
characteristics and functionality. ADOP1 indicates the adoption of EG
at the static stage. This is the rst stage of government service and
information presence online. Consequently, ADOP1 has a higher
mean between ADOP1 (3.673) and ADOP2 (3.2164). We found that
PA, PATU,and PFB are signicant predictors for the adoptionof e-Gov
at the static phase. Therefore, our hypotheses that PA, PATU, and PFB
have positive effects on the adoption of e-Gov at the static stage are
supported.When citizens are aware that there is an alternative sourceof brick
and mortar government service and information, for example e-Gov,
they might be interested in looking for this. After that, if they nd that
they have sufcient technological and psychological ability to use it
and also perceive that it provides absolute and relative advantages
they will most likely adopt it.
Since this is only the static stage of e-Gov, at this stage citizens can
only view, read, and collect government information relating to
government policies, services, rules and regulations, and different
other issues. Citizens do not adopt this stage to interact with
government agencies; rather they do that only to be informed of
government services that, instead, they could collect by physically
going to different government ofces. At this stage, availability
of resources (AOR), perceived information quality (PIQ), perceivedtrust (PT), computer self-efcacy (CSE), multilingual option (MLO),
and perceived image (PI) are not important for citizens to adopt
e-Gov. These constructs basically contribute very little to the
variances explained on ADOP1. Therefore, these hypotheses are not
signicant.
As Canada is a developed country and also advanced in modern
ICT, AOR might not be an important predictor for the adoption of
e-Gov. To adopt e-Gov, resources are mostly available. Here the
internet adoption rate is around 84% (Internet World Stats, 2008).
PIQ could be a potential predictor. However, since PFB is a signicant
predictor of e-Gov adoption, whencitizens perceive that using e-Gov
static webpage provides them with absolute and relative benets,
they are not concerned about PIQ. Same argument can be drawn for
CSE. From our demographic analysis, since more than 80% of the
respondents have at least undergraduate education and at least 80%
have online experiences formore than 3 years, CSE is not an issue for
citizens to adopt e-Gov. And, moreover, since PATU, which shared
some essence of CSE, has already contributed enough in ADOP1, CSE
is not a signicant predictor. Since adoption of the static stage of e-
Gov is very private, not an exposed matter to any other (as there are
no communications anywhere), perceived image (PI)is expectednot
to be a predictor of e-Gov at the static stage. At the static stage,
citizens are not interacting with government by any means, so theyare not disclosing any personal or nancial information in the virtual
environment. So,PT hasno signicantcausaleffect on theadoption of
e-Gov at this stage. MLO could be an important issue for citizens
whose mother tongue is not English when they need to collect
government information from published information in web pages.
However, from demographic analysis, we found that around 83% of
the respondents have mother tongue of English/French. Since
Canadian government web pages are written in English and French,
MLO cannot be an issue for use for respondents who speak both
English and French. Respondents who have a mother tongue other
than English or French are mostly new immigrants. Since the
Canadian immigration policy is highly dependent on level of
education, we found from cross tab analysis that around 90% of the
respondents whose mother tongues are not Englishor Frenchhave at
least an undergraduate education level. As we expected, language is
not a barrier for them to collect information from government web
pages that are written in English/French.
Now if we look at TAM (Davis, 1986) and DOI (Rogers, 1995), we
can get support for our analysis that PATU, which captures the
integrated view of PEOU of TAMand the complexity and compatibility
of DOI, and PFB, which captures the overlapping essences of PU
(perceived usefulness) of TAM and the relative advantage of DOI, are
the critical factors for the adoption of e-Gov, a system driven by
modern ICT. Since, government simply cannot abandon traditional
government ofces after launchinge-Gov andsincecitizenshave been
concerned about the traditional service system for decades, citizens
need to be aware of the alternative outlet of government service
system that is offered in e-Gov. Therefore, PA is also an important
predictor of e-Gov adoption that is not reected in simple technologyadoption as discovered byDavis (1986). From TPB (Ajzen & Fishbein,
1980), we know that a person's behavior is determined by the
person's intention to perform the behavior and that this intention is,
in turn, a function of the person's attitude and belief toward the
behavior. Therefore, adopting e-Gov at the static stage, which we can
view as the outcome of behavioral intention, is inuenced by the
beliefs of that person about the outcome of that behavior. We can
clearly see that PATU and PFB are those beliefs of positive attitude
toward adopting e-Gov, which in turn affect intention to use and,
nally, the adoption of e-Gov at the static stage. However, also
applicable in this theory that, for a system which has multidimen-
sional aspects (including technological, social, cultural, behavioral,
economic, organizational, and political) and which has an alternative
outlet reecting different behavioral outcomes (e.g., adoptingtraditional government services), for developing beliefs of the attitude
and intention to adopt the system, perceived awareness (PA) is an
important aspect to impart that beliefs.
From statistical analysis (see Appendices F and I), we observed
that at both stagesof service maturity of e-Gov, in theadoptionmodel,
perceived trust on e-Gov (PT) is affected by the causal effects of PATU,
PU, and PS. While in our model development, we have conjectured
that PU and PS have causal effects on PT; we did not make the
hypothesis that PATU also has a causal effect on PT. Our path analysis
has provided the result. However, since PT is not a potential factor for
the adoption of e-Gov at the static stage, we do not pay attention to
the causal relationships of PATU, PU, and PS with PT in this paragraph.
Rather, in the next section, we will do that while explaining the
signicance of ourndingsfor theadoptionof e-Gov at the interaction
25M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
10/19
stage where PT is also a signicant predictor of adopting e-Gov at the
interaction stage. We also conducted multiple regression analyses to
verify our path analyses and got concrete support for the results
(Appendix J).
For the adoption of the interaction phase, we found that PA, PATU,
PI, PT, and PIQ are the signicant predictors. Therefore, our
hypotheses that PA, PATU, PI, and PT have positive effects on the
adoption of e-Gov at the interaction stage are supported. When
citizens are aware that instead of going to brick and mortargovernment departments for seeking service and information, they
can alternatively communicate with the respective departments
through e-mail, chat rooms, etc., which are integrated with e-Gov
websites, they develop an intention to use it. After that, if they nd
that they have the technological andpsychological ability to useit and
also perceive that this electronic communication is trustworthy, they
will most likely adopt it. Since, in the interaction stage, citizens have
some communication with others through modern ICT, they feel it to
be prestigious. Citizens feel that instead of going to a physical
government ofce, online communication with government depart-
ments can enhance their social status. As a result, PI was hypothesized
to have a positive relation to the adoption of e-Gov at the interaction
stage, and our ndings support that hypothesis.
PIQ is also a signicant predictor of e-Gov adoption at the
interaction phase. However, it has a surprisingly negative relation to
ADOP2, although our primary hypothesis predicted the causal relation
to be positive.Althoughregression analysis (Appendix J) supports this
nding, the correlation matrix shows a positive relation between
ADOP2 and PIQ. Basically PIQ has the lowest correlation coefcient
among the other signicant predictors of ADOP2 (0.239). From our
literature review of EC, we have found that PIQ has positive relation
with the use of EC. However, we should notice and consider some
subtle issues in this respect. For EC or general adoption criteria of e-
Gov (irrespective to service maturity level), better information quality
encourages consumers or users to use that online system. However,
the interaction stage of e-Gov has certain specic attributes. At this
stage, citizens generally do some queries to get specic information
that they do not nd on the e-Gov websites, or to provide some
information to the service providers. Therefore, if they nd thatinformation on the e-Gov websites is systematic, sequential, up-to-
date, effective, complete, and also provide sufcient links to
supplement information shortage, if any, citizens do not feel any
urge to contact or interact with government agencies through
electronic media. As a result, perceived information quality (PIQ) of
e-Gov automatically reduces the interaction of citizens with govern-
ment agencies through e-Gov websites. This argument justies thendings of the statistical analysis, which depicts negative causeeffect
relations between ADOP2 and PIQ.
Other exogenous variablesavailability of resources (AOR), com-
puter self-efcacy (CSE), multilingual option (MLO), and perceived
functional benets (PFB)are not important for citizens to adopt e-
Gov. These constructs, basically contribute very little to the variances
explained on ADOP2. Therefore, these hypotheses are not signicant.We have already explained the plausible reasons for the hypotheses
formed by the exogenous variables AOR, CSE, and MLO to be non-
signicant on theadoptionof e-Gov forCanadian users in theprevious
section. However, PFB wasfoundsignicant for ADOP1 while it isnon-
signicant for ADOP2. The hypothesis revealing a positive relation
between PFB and ADOP2 was not proven from the statistical analysis.
This is an interesting phenomenon. As we already explained, the
interaction stage of e-Gov has certain characteristics. This stage is
generally used to seek further information related to government
services, policies, and rules and regulations; to provide government
agencies with certain personal information; and to interact to ask
some queries. Citizens typically perform these interactions through e-
mail, chat rooms, etc. In the interaction stage, government agencies
might not communicate with citizens. Therefore, in this stage, citizens
are concerned mostly with the trustworthiness of e-Gov websites, not
PFB. Contacting through e-mail is so common in this era that doing
something (e-mail) through e-Gov websites does not create any
perceptions of functional benets among citizens as the plausible
reasons of using e-Gov websites at the interaction stage (Ong & Wang,
2009). Reasonably, after PATU, PT is the second inuential predictor,
and PFB is a non-signicant cause of the adoption of e-Gov at the
interaction stage.
Based on TAM (Davis, 1986) and DOI (Rogers, 1995) we cannotexactly explain the adoption criteria of e-Gov at the interaction stage.
The construct PATU, which captures the integrated view of PEOU of
TAM and complexity and compatibility of DOI, is the most contrib-
uting construct of EG adoption at the interaction stage. We also nd
that perceived image (PI) inuences the acceptance and use of e-Gov
at the interaction stage. This nding is supported by DOI proposed by
Moore and Benbasat (1991). Image refers to citizens' perceptions of
adopting e-Gov to present themselves as superior to others in the
society. Interaction with e-Gov systems, instead of using traditional
government ofces, reects a perception by citizens of superiorstatus.
PI exactly captures this superior status perception. In this phase, PA is
also a signicant predictor, like ADOP1 which is an addition to TAM
and DOI specic to e-Gov. Static and interaction stages of e-Gov are
the primary stages when a country rst launches e-Gov projects
(Accenture, 2005). Since these two phases are the rst introduction to
online service of government and are the alternative of traditional
government service systems, at these two levels of service of e-Gov,
citizens' awareness is presumably an important critical factor for
attitude and behavioral intention to use those systems. However, e-
Gov, particularly at the interaction stage, has more aspects than
simply adopting an innovation. Citizens' perceptions and expectations
differ remarkably at this stage from simply viewing government
information through static websites of e-Gov. Therefore, in this phase,
PT and PIQ are also additions to traditional TAM and DOI constructs,
which are introduced to capture the specic interaction character-
istics of e-Gov services. We have already explained the reason for the
negative impact of PIQ on ADOP2.
If we look at the sources of trust that citizens have while
interacting with the government, based onEaston (1965), we seethat the trust that develops the satisfaction of citizens with
governments due to their credible performance is very signicant.
This institutional-based trust is an important component while
interacting in e-Gov web pages (Parent et al., 2005). Thomas
(1998)also supported that citizens required trust while interacting
with e-Gov, a trust that derives from institutional credibility. If
actors of e-Gov do not feel institutional trust, according to the
institutional theory, they might not follow the institutional norms
of e-Gov. Therefore, PT is a reasonable addition to TAM and DOI in
predicting the adoption behavior of e-Gov at the interaction stage,
which is not reected in simple technology adoption as discovered
byDavis (1986). From TPB (Ajzen & Fishbein, 1980), we can clearly
see that PATU, PT, and PIQ are beliefs of attitude toward adopting e-
Gov, which in turn affect intention to use and, nally, the adoptionof e-Gov at the interaction stage. However, also related to this
theory is that for a system which has multidimensional aspects
(including technological, social, cultural, behavioral, economic,
organizational, and political) and which has alternative outlets
reecting different behavioral outcomes (e.g., adopting traditional
government services), PA and PI are important aspects for
developing beliefs of attitude and intention of adopting this
modern ICT-intense system.
From Appendix F we can see that at both stages of service
maturity of e-Gov, in adoption modelperceived trust on e-Gov
(PT)is affected by the causal effects of PATU, PU, and PS. We nd
that PS is the strongest predictor of PT. In all phases of e-Gov,
citizens have trust in e-Gov if they perceive that the e-Gov websites
ensure security ofnancial and personal information. Although we
26 M.A. Shareef et al. / Government Information Quarterly 28 (2011) 1735
-
8/11/2019 s2 sembarang
11/19
did not conjecture any relation between PATU and PT in our
hypothesis development process, we nd a strong causal relation of
PATU on PT. The perceived technological and psychological ability
to use e-Gov increases trust in e-Gov. This relation is suggested by
path analysis for better tness of the model and we nd several
logical aspects for this relation.Warkentin et al. (2002) and Parent
et al. (2005)both discovered that an important component of trust
is characteristic and process-based. If citizens perceive that they are
suf
ciently capable of handling online systems, their characteristicand process-based trust also increase (Warkentin et al., 2002). In
light of TPB, we also see that if a user believes that he/she is capable
of using an online system, he/she feels attitude and intention to use
that system, which implies his/her development of characteristic-
based trust. Therefore, the causal relation of PATU with PT has a
strong theoretical base. We conjectured that perceived uncertainty
(PU) has a negative relation to PT; however, surprisingly, we nd
that PU has a positive relation to PT (loading factor 0.25, signicant
at the 0.01 level). That is, the perception of uncertainty basically
increases the perception of trust in EG. Apparently, this result is
confusing. Therefore, we looked at the result from different
statistical and conceptual perspectives. The correlation matrix
shows that PU has a very weak positive correlation with PT
(0.109). PU has weak negative correlations with PATU and PS. We
have also done regression analysis of PU, PS, and PATU (indepen-
dent variables) with PT (dependent variable) (Appendix J). We see
that all these variables are signicant positive predictors of PT.
However, in stepwise regression, while only PU is inputted, we see
that PU has a negative non-signicant effect on PT. While inputting
PU, PS, and PATUeither in path analysis or in regression analysis
PU has the least positive effect on PT with very weak correlation
with PT. This simply signies that, while measuring the combined
effect of PATU, PU, and PS, PU has a positive effect on PT. In our
sample, more t