Jurnal Asli SIM Rokman

Public Administration Presence on the Web: a Cultural
Explanation
Fatma Bouaziz
Faculty of Economics and Management, University of Sfax, Tunisia
fatma.bouaziz@fsegs.rnu.tn
Abstract: UN reports (2002-2005) reveal several differences in worldwide adoption of e-Government by public administration (PA). This research conducts an analysis of the impact of the national culture on e-
Government adoption by PA to attempt to explain these differences. The impact of culture, as defined by
Hofstede, is tested at the level of Web Measure Index and at the different stages of e-Government evolution.
The regression analysis and Pearson’s coefficients show the importance of Individualism and Uncertainty
Avoidance. Power Distance has lower impact on e-Government adoption. The level of education attainment
moderates these relationships.
Keywords: e-Government, stages of e-Government evolution, Web Measure Index, national culture.
1. Introduction
e-Government programs remain at the top of most countries policy agendas (UN, 2002). The emphasis on
the e-Government may be due to the potential of information and communication technologies (ICT) to
transform public administration (PA) to an efficient system, enhance public services quality, establish trust
between PA and citizens and realize economic objectives (OECD, 2003; UN, 2004). In fact, e-Government
aims at providing an efficient government management, better service delivery, and empowerment of
citizens through access to information and participation in public policy decision making (UN, 2005). For
these benefits, the PA in most countries develops Web sites, portals and electronic services delivery (UN,
2002; 2003; 2004; 2005). However, differences in worldwide adoption of e-Government are considerable.
The e-Government readiness index estimated for each of UN countries varies between 0, 9062 in the United
States of America and 0, 0011 in Liberia (UN, 2005). The Web Measure Index varies between 0 (no
presence for countries like Haiti, Zambia, Liberia, etc.) and 1 in the United States of America.
Certainly, the diffusion of the Web in PA leans on a political will that mobilizes the necessary human,
material and financial resources. But, this diffusion depends also on social and economic variables (UN,
2002). UN (2005) affirm that e-Government readiness is a function of a country’s situation of technological
and telecommunication infrastructure and of the level of its human resource development. Nevertheless, “the
number of variables - social, legal and institutional as well as economic and technological - which might
retard the diffusion process is virtually limitless’’ (Rosenberg, 1972, cited by Erumban and de Jong, 2006,
p.303). Recently, several researchers show the importance of culture in IT adoption and justify the
divergences between countries in the adoption of ICT, in general (Erumban and de Jong, 2006) and of
specific technologies, such as Internet, PC, fax machines, cellular phone, telephone and pagers (Bagchi and
al., 2004; Gong and al., 2007) on the base of the cultural variables. According to Zghal (2001), cultures
include both favourable factors to ICT exploitation and unfavourable ones that may hinder an efficient
implementation of these technologies. So, this research tries to answer this question: what is the impact of
the national culture on the adoption of the e-Government by the PA? The analysis starts with the definition of
e-Government and the presentation of its stages of evolution. Then, an explanation of the differences
between the countries use of the e-Government is advanced on the base of a cultural perspective and the
hypotheses are formulated. After the precision of the methodology, the results and their interpretations are
presented. We end by a conclusion where the contributions and limits of this study are discussed.
2. e-Government: Definition, dimensions and stages of evolution
2.1 Definition and dimensions of e-Government
UN (2005, p.14) define e-Government as “the use of ICT and its application by the government for the
provision of information and public services to the people”. Broadly defined, e-Government includes all ICT
platforms and applications in use by the public sector. The objectives are a continuous optimization of public
services delivery and information dissemination, citizens’ participation in political processes, while
transforming the internal and external relationships and processes by the technology, the Internet and the
new medias (Moon, 2002; UN, 2002). Narrowly defined, e-Government is the use of ICT between PA
Electronic Journal of e-Government Volume 6 Issue 1 2008 (11-22)
www.ejeg.com 12 ©Academic Conferences Ltd
agencies, enterprises, and citizens to simplify and improve information diffusion, communication and public
services delivery (Moon, 2002; UN, 2002).
According to El Jamali and al. (2004), a complete definition of e-Government must consider all partners of
PA. e-Government includes interactions between the administration and the citizens “Government to
Citizens” (G2C), the administration and the business “Government to Business” (G2B), and the PA agencies
“Government to Government” (G2G). These interactions concern the information diffusion, communication
and / or the realization of online transactions. The G2C/C2G allows citizens to consult information, to
communicate and to make transactions instantaneously and easily. Several interactions such as services
delivery, social assistance, legal authorization, etc., can be developed. The G2B/B2G refers to the use of
ICT in relationships between PA agencies and private and public business for information diffusion,
communication and realization of administrative formalities. Finally, the G2G refers to the use of ICT
between PA agencies at national, regional and local levels. PA is composed of several agencies that can
intervene, each according to its domain of expertise, in the realization of the same transaction. ICT (Internet,
Intranet, Web servers, etc.) allow them to share their data and to put in common their knowledge and
resources. Moreover, they allow the development of technological and organizational platforms that integrate
operational, decisional and management processes (El Jamali and al., 2004).
2.2 Stages of e-Government evolution
According to Layne and Lee (2001), e-Government implementation follows an evolutionary process starting
with the phase of information, then the phases of interaction, transaction, and horizontal and vertical
integration (Moon, 2002; El Jamali and al., 2004). However, it is not obligatory for a PA to follow a linear
progression in the implementation of e-Government. The possibility to evolve to stages of integration before
exploiting the full potential of information or interaction phases is not excluded. UN (2005) presents a model
of e-Government evolution composed of five stages. These five stages will be used in the statistical analysis.
For this reason, the model of e-Government evolution and its components’ definitions are adopted (UN,
2005):
Table 1: Stages of e-Government evolution, (UN, 2005, p.16)
Emerging Presence is stage 1 representing information, which is limited and basic. The e-Government online presence
comprises a web page and /or an official website; links to ministries/departments of education, health, social welfare,
labour and finance may/may not exist; links to regional/local government may/may not exist; some archived information
such as the head of states' message or a document such as the constitution may be available on line, most information
remains static with the fewest options for citizens.
Enhanced presence is stage 2 in which the government provides greater public policy and governance sources of current
and archived information, such as policies, laws and regulation, reports, newsletters, and downloadable databases. The
user can search for a document and there is a help feature and a site map provided. A larger selection of public policy
documents such as an e-Government strategy, policy briefs on specific education or health issues. Though more
sophisticated, the interaction is still primarily unidirectional with information flowing essentially from government to the
citizen.
Interactive presence is stage 3 in which the online services of the government enter the interactive mode with services to
enhance convenience of the consumer such as downloadable forms for tax payment, application for license renewal.
Audio and video capability is provided for relevant public information. The government officials can be contacted via
email, fax, telephone and post. The site is updated with greater regularity to keep the information current and up to date
for the public.
Transactional presence is stage 4 that allows two-way interaction between the citizen and his/her government. It includes
options for paying taxes; applying for ID cards, birth certificates/passports, license renewals and other similar C2G
interactions by allowing him/her to submit these online 24/7. The citizens are able to pay for relevant public services,
such as motor vehicle violation, taxes, fees for postal services through their credit, bank or debit card. Providers of goods
and services are able to bid online for public contacts via secure links.
Networked presence is stage 5 which represents the most sophisticated level in the online e-Government initiatives. It
can be characterized by an integration of G2G, G2C and C2G (and reverse) interactions. The government encourages
participatory deliberative decision-making and is willing and able to involve the society in a two way open dialogue.
Through interactive features such as the web comment form, and innovative online consultation mechanisms, the
government actively solicits citizens’ views on public policy, law making, and democratic participatory decision making.
Implicit in this stage of the model is the integration of the public sector agencies with full cooperation and understanding
of the concept of collective decision-making, participatory democracy and citizen empowerment as a democratic right.
Fatma Bouaziz
www.ejeg.com 13 ISSN 1479-439X
3. e-Government adoption by PA: A cultural explanation
The Web Measure Index (WMI) (UN, 2002-2005) assesses the websites of the governments to determine if
they are employing e-Government to the fullest. This index provides an evaluation of Internet and World
Wide Web utilization by the PA for the diffusion of information and the public services delivery.
UN results (2005, p.57) reveal that PA in 179 countries are online in 2005. However, disparities between
countries are considerable: the WMI varies between 0 (no presence for countries like Haiti, Zambia, Liberia,
etc.) and 1 in the United States of America. These disparities concern countries with different levels of
economic, social and technological development, but even for industrialized countries, having similar
features, differences in the WMI exist. In addition, 22 developing countries are among the 50 first countries.
Among these 22 countries, 11 are better classified than some industrialized and high income per capita
countries such as France, Iceland and Switzerland (appendix 1).
Several researchers (Bagchi and al., 2004; Erumban and de Jong, 2006; Gong and al., 2007) justify the
worldwide divergences in the adoption of ICT, in general, or of specific technologies such as Internet, PC,
fax machines, telephone, cellular phone and pagers by the cultural differences between countries. In fact,
different cultures react differently to new products and technological innovations (Gong and al., 2007) and
deciding to adopt such technologies may be influenced by the social and cultural features and the perceived
values of the individuals within a society (Erumban and de Jong, 2006). Therefore, culture can be considered
as a factor that differentiates between countries in the adoption of ICT.
Zghal (2001, p.5) defines culture as an internal logic, a sort of habits and a tacit understanding terrain that
members in a society share and to which each adjust his/her behaviours. So, individuals share a collective
national character that shapes their values, believes, attitudes, and determine their behaviours in a society. It
is a system of values collectively shared (Hofstede) or culture that differentiates between the countries. Gong
and al. (2007) and Erumban and de Jong (2006) recognize two theoretical approaches of the national
culture: the approach of Hall (1976) and the approach of Hofstede. According to Gong and al. (2007), In
Hall’s approach, culture is seen as a unidimensional construct in which countries are grouped into either
high/low or high/medium/low context categories. This distinction is based on the way messages are
communicated within a society. In high context cultures (e.g. Japan, China), contextual significations are
important in the interpretation of a message, while in low-context cultures (e.g. USA, Canada), most of the
information is explicitly expressed in the words (Gong and al., 2007). Gong and al. (2007) highlight that the
studies that have used the Hall’s approach to explore the influence of culture on the diffusion have shown
mixed results because of an empirical measurement problem when using this approach. In addition,
according to Erumban and de Jong (2006), this approach doesn’t describe the cultural differences between
countries in a detailed manner. So, more sensitive measures of national culture are needed.
Based on data from an empirical survey carried on between 1967 and 1973 with about 116 000 employees
of IBM in 50 countries and 3 regions, Hofstede observes that the national cultures can be distinguished on
four cultural dimensions: Power Distance, Uncertainty Avoidance, Individualism/ Collectivism and
Masculinity/ Femininity. A fifth dimension was added later (Long-Term Orientation). Hofstede’s approach of
culture and in particular these dimensions are the most studied when considering ICT adoption at a national
level (Bagchi and al., 2004; Kovacic, 2005; Erumban and de Jong, 2006; Gong and al., 2007). Despite the
criticism of its methodology and its limited context (Baskerville, 2003; 2005), this approach offers the most
robust measures of national cultures and its five dimensions are considered as the most discriminating
factors between countries (Gong and al., 2007). Bagchi and al. (2004) stress their relevance to assess
culture’s influence on adoption of ICT. Besides, all these authors conclude that differences in adoption of ICT
between countries can be attributed to the national culture as it is described in Hofstede’s approach. Thus,
these dimensions provide better criteria for assessing culture’s influence on ICT adoption, and will be used to
analyze differences observed in the case of e-Government. Nevertheless, cultures include both favourable
factors to ICT exploitation and unfavourable ones that may hinder an efficient implementation of these
technologies (Zghal, 2001). So, if culture has a certain duality on behaviours, the question is: what are
among these dimensions those that may have an impact on e-Government adoption by PA? Each dimension
of culture and its expected relationship with ICT adoption are discussed below.
Electronic Journal of e-Government Volume 6 Issue 1 2008 (11-22)
www.ejeg.com 14 ©Academic Conferences Ltd
4. Culture’s dimensions and e-Government adoption by PA: an analytical
framework
4.1 Power distance (PD)
Power distance is the extent to which a society accepts differences and inequalities in power distribution
among individuals, organizations and institutions. In large (high-scoring) PD countries, individuals are more
ready to accept power differences and a large hierarchical order than individuals in small PD countries.
These later countries are characterized by decentralized and participative decision making structures
(Bagchi and al., 2004).
Since ICT can contribute to the decentralization of decision making by allowing more suitable information
diffusion and sharing, they will be more accepted in small PD countries. In these countries, Kovacic (2005)
remarks that citizen involvement to political decision processes is solicited, and requires ICT implementation.
Hence, this author suggests that the large PD countries may have negative attitudes toward the adoption of
ICT.
But, ICT are, also, a symbol of power and authority; this can encourage their adoption in large PD countries
(Bagchi and al., 2004). This adoption risks to be a sign of social status and limited to the elite of the society
(Gong and al., 2007). In large PD countries, organizations are characterized by centralized structures,
authority and application of formal rules (Erumban and de Jong, 2006). Thus, individuals called to respect
authority and differences and to follow instructions may be less ready to challenge status quo and less
motivated to adopt ICT (Gong and al., 2007).
Results from earlier studies show that when the PD is low, the adoption of PC and cellular phones increases
(Bagchi and al., 2004). It is also the case in Internet use and access (Gong and al., 2007) and in ICT
adoption independently of their types (Erumban and de Jong, 2006). Kovacic (2005) confirms the hypothesis
that the government of a large PD country has negative attitude toward the increase of the e-Government
readiness. So, we expect that e-Government adoption increases in small PD countries (H1).
4.2 Individualism/ Collectivism (ID)
Individualism describes the relationship between the group and the individual in a society. In individualistic
cultures (low collectivism) individuals are more concerned with themselves and their direct families’
members. The personal freedom and the individual decision-making are valued (Erumban and de Jong,
2006; Gong and al., 2007). In contrast, members from low individualistic cultures (high collectivism) show
more attachment to their group of adherence (extended family, community, association, etc.). They are
integrated into strong, cohesive groups in which individuals expect their relatives and others in their group to
look after them in exchange for unquestioning loyalty. Obligations and group harmony come before individual
objectives (Gong and al., 2007).
Authors (Bagchi and al., 2004; Kovacic, 2005; Erumban and de Jong, 2006; Gong and al., 2007) agree that
ICT adoption increases in individualistic cultures. On one hand, individuals in individualistic cultures have
more freedom to express their opinions and try new ideas than individuals in collectivistic cultures (Erumban
and de Jong, 2006; Gong and al., 2007). Therefore, they are more innovative and more open to the adoption
of ICT. On another side, ICT reduces time and space constraints. They assure more independence in work
or personal life in the sense that they have the option to maintain greater physical distance and allow
individuals to schedule their activities to meet the needs of several groups to which they belong (Bagchi and
al., 2004). Because of the same features, ICT promotes continuous contact between groups’ members in
collectivistic cultures (Bagchi and al., 2004). However, in such cultures, the importance of face-to-face
interactions reduces the role of ICT in maintaining relationships (Bagchi and al., 2004). Moreover, individuals
do not adopt ICT if they prove to be in contradiction with their groups norms (Erumban and de Jong, 2006).
Significant relationships between individualism and the adoption of PC, Internet, telephone, pagers (Bagchi
and al., 2004) and ICT in general (Erumban and de Jong, 2006) are showed by these authors. In the case of
e-Government, Kovacic (2005) finds that the government of an individualistic culture shows positive attitudes
toward the improvement of the e-Government readiness. In contrast, Gong and al. (2007) invalidate the
hypothesized positive effect of individualism on Internet adoption. But, they find a marginally significant
relationship and in the expected direction for narrow band penetration (t = 1, 56; p <0, 1). So, we expect that
e-Government adoption by PA increases in individualistic cultures (H2).
Fatma Bouaziz
www.ejeg.com 15 ISSN 1479-439X
4.3 Masculinity/ Femininity (MA)
This dimension of culture is based on a distinction between men and women’s roles in a society. Values
such as assertiveness, performance, success and competition characterize masculine cultures. On the
contrary, a feminine culture focuses on human relationships, quality of life, solidarity, equality, interest for
others and protection of the weak. These values are further searched by women who are more modest,
tender and more guided by the improvement of quality of life than men.
According to Hofstede, organizations in masculine cultures encourage the competition and recognize the
performance of individuals; these are the features of innovative organizations (Erumban and de Jong, 2006).
Organizations adopt ICT for more efficiency and competitiveness which are valued in masculine cultures
(Bagchi and al., 2004). In these cultures, the adoption of ICT increases because they permit to obtain useful
information to competition (Kovacic, 2005; Erumban and de Jong, 2006), influence the contribution of actors
to efficiency (Bagchi and al., 2004) and improve the performance and success chances (Kovacic, 2005). But,
it seems that earlier researches find results that contradict the authors’ expectations: masculinity has a
negative effect on the access to Internet and its use (Gong and al., 2007) and it has no relationship with the
adoption of ICT in general (Erumban and de Jong, 2006), and with the e-Government readiness index in
particular (Kovacic, 2005).
Moreover, ICT allow more cooperation at work and a better quality of life that are values of feminine cultures
(Bagchi and al., 2004). Gong and al. (2007) notice that ICT characteristics, such as Internet, are more
appropriate in the feminine cultures because they facilitate the information sharing and the communication
between people, groups and organizations. If this potential of communication is the major motive of ICT
adoption, then they are more accepted in high femininity countries (Erumban and de Jong, 2006). This idea
is supported by Bagchi and al. (2004) who find a strong relationship between feminine culture features and
the adoption of the telephone and the cellular phone. So we consider that the e-Government adoption
increases in high femininity countries (H3).
4.4 Uncertainty avoidance (UA)
Uncertainty avoidance refers to the attitude toward the risk, the uncertainty, and the new phenomena. The
new situations can include surprises and are different of the usual. High UA cultures try to structure life,
society, and minimize possibilities to have such situations by laws, formalism and measures of security.
Individuals are, generally, more resistant to change, risk averse and less innovative (Kovacic, 2005; Gong
and al., 2007). But, cultures that accept the uncertainty (low UA) are more tolerant toward others different
opinions. In these cultures, individuals show more tolerance toward the risks, like those associated to ICT,
and are more innovative (Bagchi and al., 2004; Gong and al., 2007). According to Bagchi and al. (2004), PC
routinize jobs, fax machines and e-mail leave paper records, telephone and cellular phone improve the
communication to solve problems more quickly. Consequently, ICT can respond to the needs of reducing
uncertainty in high UA cultures. However, ICT adoption is risky. Because these cultures are more risk averse
and resist to changes, their adoption of ICT is less important than in low UA countries (Bagchi and al., 2004;
Kovacic, 2005; Erumban and de Jong, 2006).
Mixed results are given by earlier studies. Supposing that the UA is negatively associated to Internet use and
access, Gong and al. (2007) don’t find a relationship between these two variables. Moreover, Bagchi and al.
(2004) conclude that the relationship between ICT adoption and UA is not obvious. Kovacic (2005) finds a
weak support to the hypothesis that the government of country with a high UA culture has a negative attitude
toward increasing the level of e-Government readiness. Contrary to these authors, Erumban and de Jong
(2006) confirm the hypothesis that high UA cultures show a lowest rate of ICT adoption than cultures with a
low UA. They notice that this cultural dimension is the most significant to the explanation of differences
between countries in the adoption of ICT. Therefore, we suggest that e-Government adoption increases in
low UA countries (H4).
4.5 Long term orientation / short term orientation (LOT)
This dimension refers to the degree to which a culture values its tradition and how much individuals focus on
their past and future (Erumban and de Jong, 2006; Gong and al., 2007). According to Hofstede, cultures with
LOT are characterized by values such as persistence, adaptation of tradition to new circumstances,
perseverance and the idea that most important events in life will occur in the future. In contrast, cultures with
a short term orientation (low LOT) expect quick results and focus on respect for past, tradition and stability.
Electronic Journal of e-Government Volume 6 Issue 1 2008 (11-22)
www.ejeg.com 16 ©Academic Conferences Ltd
LOT is the least tested dimension in cross-culture studies (Gong and al., 2007). These authors consider that
LOT cultures can better perceive the long term benefits of Internet and, therefore, they are more in favour of
the adoption of this innovation. They find that this cultural dimension has a significant positive influence on
Internet use and access. This impact increases when introducing the control variable “level of education” in
the regression model tested for the variable “broadband penetration”. In contrast, Erumban and de Jong
(2006) consider that low LOT cultures are more opened to new ideas than high LOT cultures, so that in such
cultures the adoption of ICT increases. Though, these authors don’t find conclusive results because of the
reduced size of the studied sample. So, we suggest that e-Government adoption increases in countries
characterized by LOT cultures (H5).
4.6 Control variables
Question on cultural dimensions that have an impact on the PA presence on the Web doesn’t imply a
deterministic view of the relationship between culture and ICT adoption. On one side, Bagchi and al. (2004,
p.31) argue that “when making prediction about the likely implications culture has for technology use, it is
important to avoid an overly deterministic view of either technology or culture”. Kovacic (2005) considers that
this relationship is not simply causal. Rather, it is dynamic as the e-Government can have an impact on the
national culture. On the other side, even though the UN (2005, p.91) explains disparities between countries
in the e-Government adoption by cultural differences, this organization highlights the role of differences in
political, economic and social systems, level of development, resource availability, technological
infrastructure and human capital. Kovacic (2005) affirms that e-Government adoption depends on the level of
democracy in the country, the cost of implementation and the perceived political benefits for the government
from implementing an e-Government initiative, and that those more democratic countries are higher ranked
on the e-Government readiness list than the less democratic ones.
Therefore, other variables may influence the ICT adoption: “while cultures can promote, resist, or shape
technology use, they do not wholly determine technology use” (Bagchi and al., 2004, p.31). Hofstede,
himself, argues that the five cultural dimensions are statistically correlated to several data of countries. For
example, power distance is correlated to income inequalities in a country, and individualism is correlated to
Per Capita Gross National Product. He explains that when the effect of other variables (economic variables,
for example) is significant, the cultural variables are redundant. But, if the cultural variables are still
significant despite of included economic variables, then the effect of culture on observed phenomenon could
be confirmed.
Erumban and de Jong (2006) and Gong and al. (2007), in their analysis of the relationship between the
national culture and the ICT adoption, study the impact of the level of education attainment. This variable will
be studied in the case of e-Government adoption by PA. Gong and al. (2007) find that the education level
contributes to explain the Internet access and use. They conjecture that the utilization of Internet requires
high levels of education and the researches on innovation diffusion must continue to take it in account. Thus,
we expect that e-Government adoption increases in countries where the level of education is high (H6).
5. Methodology
Qian (2007) affirms that since 2000, more than 25 global e-Government surveys are conducted. The most
recognised are surveys of “Accenture, Brown University, Economic Intelligence Unites (EIU), Taylor Nelson
Sofres (TN), United Nations, and World Economic Forum/Harvard University” (Qian, 2007). He adds that the
UN report on global e-Government readiness is one of the most read and used by practitioners and
researchers for the rigor of its methodology and the richness of data on all UN countries it contains. For this
reason, information given in the 2005 report is exploited in this study. More precisely, it is the Web Measure
Index (WMI) that serves as a measure of e-Government adoption by PA. In fact, “it is important to take a
closer look at the online public sector offerings in and off themselves. The web measure index assesses the
websites of the governments to determine if they are employing e-Government to the fullest. The web
measure rankings are different to the e-Government readiness rankings, which are based on a composite
index comprising web measure, infrastructure and human capital assessment” (UN, 2005, p.71).
The scores provided by Hofstede serve for the measure of cultural dimensions. These scores are available
in 65 countries and 3 regions (Arab world, East and West Africa) (www.geerthofstede.
com/hofstede_dimensions.php). Countries belonging to these three regions are excluded from the
present study since it can be inconceivable to consider that these countries possess the same scores on the
cultural dimensions than the region. For Arab world, for example, El Louadi (2004) argues that the imputation
of regional scores and dimensions to all Arab countries (have one same score (high) in power distance, are
Fatma Bouaziz
www.ejeg.com 17 ISSN 1479-439X
relatively collectivist, etc.) is worrying. According to the same author, the complexity of the Arab culture
necessitates the study of the Arab values and their inclusion in Hofstede’s framework. Among others
countries studied by Hofstede, Taiwan and Hong Kong are excluded because the UN report doesn’t provide
data on their WMI. Therefore, a final sample of 63 countries is examined.
Methods of correlation and regression analysis are applied to the data. Several researchers (Bagchi and al.,
2004; Kovacic, 2005; Erumban and de Jong, 2006; Gong and al., 2007) use regression analysis to examine
the relationship between national culture and ICT adoption. Regression analysis allows the prediction of
values taken by an endogenous variable from a set of explanatory variables x1, x2,…, xp. Its application in
this research seems to be adequate. In fact, no major problem of multicollinearity is detected despite of the
existence of significant correlations between the following independent variables: PD / ID; ID / LOT; ID / EDU
and PD / EDU (table 2). As mentioned by Hofstede, these cultural dimensions have relationships between
them and with socioeconomic variables such as the level of education. But, all the values of the variance
inflation factors (VIF), which permit to verify if correlations between independent variables cause a
multicollinearity problem, are comprised between 1,202 and 4,121. These values are much less than 10, a
frequently suggested cut-off value of multicollinearity (Bagchi and al., 2004).
Table 2: Correlation coefficients between independent variables
IDG1 PDG1 UAG1 MAG1 LOTG1
PDG1 -0,613**
UAG1 -0,232 0,221
MAG1 0,095 0,136 -0,040
LOTG1 -0,430* 0,341 -0,043 0,113
EDUG1 0,514** -0,409** 0,013 -0,049 0,014
** Significant at 0.01 level; * Significant at 0.05 level.
6. Results and discussion
In addition to the WMI, UN report (2005) provides data on e-Government service delivery by stages of
evolution. These stages may be related to the depth of e-Government adoption by PA. Data show that the
PA differ on their presence on the Web, but also they have different scores at the level of these stages. This
reflects different levels of e-Government adoption by PA. To verify if the cultural and socioeconomic
variables impact can vary according to the level of engagement in these stages of e-Government evolution,
coefficients of correlation are firstly calculated (table 3). Results show that:
􀂃 At the first stage of e-Government evolution, all correlation coefficients are low and insignificant.
They are positive for the variables ID, MA and EDU, and negative for the variables PD, UA and
LOT.
􀂃 At the second and the third stages, the variables ID and EDU show significant and positive
coefficients of correlation. These coefficients are significant and negative for the variable PD. For
UA, LOT and MA, coefficients are low and insignificant. They are negative in the case of UA and
LOT and positive for MA.
􀂃 At the fourth stage, the variables ID, EDU and PD show similar results as those of the second
and the third stages. In addition, the UA has a significant and negative coefficient. The
coefficients relative to LOT and MA are insignificant. Similar results are obtained at the fifth
stage and the WMI.
􀂃 Among the independent variables, EDU has the highest coefficients of correlation (in terms of
absolute values) which underlines its importance. This coefficient has a higher value at the
second stage than the others.
Table 3: Correlation coefficients : e-Government adoption/ independent variables
ID PD UA MA LOT EDU
Stage 1 0,164 -0,131 -0,014 0,156 -0,293 0,163
Stage 2 0,525** -0,405** -0,133 0,092 -0,139 0,591**
Stage 3 0,454** -0,404** -0,229 0,059 -0,093 0,521**
Stage 4 0,434** -0,424** -0,359** -0,12 -0,116 0,441**
Stage 5 0,406** -0,296* -0,298* 0,004 -0,093 0,488**
WMI 0,502** -0,422** -0,259** 0,03 -0,126 0,555**
** Significant at 0.01 level; * Significant at 0.05 level.
Electronic Journal of e-Government Volume 6 Issue 1 2008 (11-22)
www.ejeg.com 18 ©Academic Conferences Ltd
Thus, the impact of the cultural variables and of the level of education appears progressively with the e-
Government evolution. Relationships between these variables that depend on the level of e-Government
adoption may exist. To better examine these relationships and identify the most significant variables, a
regression analysis (table 4) is done for each of e-Government evolution stages and for the WMI. The
coefficient of determination R2 measures the proportion of the variation in the dependent variable as
explained by independent variables in a regression model, and standardized Beta coefficients are used to
make statements about their relative importance. A higher Beta value, for a given independent variable,
means that this variable is more relevant than the others.
Table 4: Regression results
Model 2
(Y = a + b1PD + b2ID +
b3MA + b4UA + є)
Model 3
(Y = a + b1PD + b2ID + b3MA +
b4UA + b5EDU + є)
Model 4
(Y = a + b1PD + b2ID + b3MA +
b4UA + b5LOT + b6EDU + є)
R2 Bêta t (p) R2 Bêta t (p) R2 Bêta t (p)
Phase
1
0,054 - 0,063 - 0,171 -
Phase
2
0,243 ID (0,423) 5,930 (0,005) 0,433 EDU (0,451) 3,771 (0,000) 0,639 EDU
(0,787)
3,972
(0,001)
Phase
3
0,247 ID (0,298) 1,998 (0,05) 0,367 EDU (0,416) 3,288 (0,002) 0,532 EDU
(0,635)
2,815
(0,012)
Phase
4
0,305 ID (0,277)
UA (-0,261)
1,937 (0,058)
-2,302
(0,025)
0,377 EDU (0,320)
UA (-0,309)
2,554 (0,013)
-2,812 (0,007)
0,438 EDU
(0,501)
2,024
(0,059)
Phase
5
0,211 ID (0,366)
UA (-0,212)
2,204 (0,032)
-1,761
(0,084)
0,308 EDU (0,374)
UA (-0,269)
2,832 (0,006)
-2,323 (0,024)
0,607 EDU
(0,685)
PD (-
0,607)
3,311
(0,004)
-2,238
(0,039)
WMI 0,290 ID (0,364) 2,519 (0,015) 0,421 EDU (0,433)
UA (-0,201)
3,577 (0,001)
-1,901 (0,062)
0,606 EDU
(0,699)
3,376
(0,004)
In a first step of analysis, a first regression model (results of this model are not presented in the table 4),
where the five cultural dimensions are included, is tested. Results of this model show that these variables
have insignificant relationships with the e-Government adoption by PA at all evolution stages. It is also the
case for the WMI. Thus, the hypotheses 1 to 5 are rejected.
Since scores of the variable LOT are only available in 24 countries, this low number may influence the
results. For this reason, a second regression model is tested for four cultural variables: PD, ID, UA and MA
(Model 2). Results of model 2 (table 4) show that the cultural variables have insignificant coefficient at the
first stage of e-Government evolution and that the coefficient R2 is low (R2 = 0,054). In this phase, basic and
limited information is presented on static Web sites conceived for consultation without possibility of
communication or interaction with the PA. Characteristics of these sites don’t contradict with the cultural
dimensions of a nation. The PA can adopt this stage without cultural constraints or stimulants.
The contribution of cultural variables to the explained variance of the adoption of e-Government increases
from the second stage (R2 is between 0,211 and 0,305). Among the cultural variables, ID shows positive and
significant coefficients at stages 2, 3, 4 and 5 and at the WMI. Looking for a better public services delivery
and an improvement of its efficiency, PA in individualistic cultures adopts the ICT more then in collectivistic
cultures. Therefore, H2 is supported. The coefficient relative to UA is negative and significant at stages 4 and
5, negative insignificant at stage 3 and at the WMI and positive insignificant at stages 1 and 2. Consequently,
H4 (the e-Government adoption by the PA increases in low UA cultures) is partially supported. PD has
negative and insignificant coefficients at all stages of e-Government evolution and at the WMI. So this
variable doesn’t have a relationship with e-Government adoption by the PA (but the sign of the relationship is
as expected), contrary to what was expected in the hypothesis H1. MA shows insignificant coefficients at all
stages of e-Government evolution. Therefore, this variable doesn’t have any impact on the adoption of e-
Government by PA contrary to what was hypothesised (H3).
In a second step of analysis, the socioeconomic variable EDU is introduced in the regression model (Model
3, table 4). Results show that the value of the determination coefficients R2 increases. As found in model 2,
at the first stage of e-Government evolution, the cultural and the socioeconomic variables don’t have any
impact on the adoption. This stage concerns a passive presence of the PA on the Web where sites are
static. This hasn’t contradictions with the cultural variables and doesn’t require a high level of education. The
education level is an important variable at the second, third, fourth and fifth e-Government stages and at the
WMI. Since Beta coefficients are positive and significant, hypothesis H6 is supported. After the introduction
of the socioeconomic variable (EDU), ID doesn’t have any more influence (Beta are positive and
Fatma Bouaziz
www.ejeg.com 19 ISSN 1479-439X
insignificant). Therefore, hypothesis H2 is not supported. This can be due to the high correlation between
EDU and ID (table 2). In contrast, the coefficients Beta of UA are negative and their significance increases at
stages 4 and 5.
The coefficient Beta relative to UA has the highest value at the stage 5. In addition, this cultural variable has
a significant negative relationship with WMI. In the first stages (2, 3) of e-Government evolution, the
communication and interaction are essentially unidirectional with information flowing essentially from the
government to the citizen (UN, 2005). This may not require developed measures of security; thus the
importance of the variable UA can be low. But, from the stage 4, online transactions, exchange of personal
information and online payments become possible. This requires the development of more secure links (UN,
2005). If citizens consider that the use of e-services offered by PA is insufficiently secure, they may refuse
them. In addition to security requirements, the stage 5 necessitates cooperation between different public
agencies for their integration. Cooperation is often mentioned among the barriers of e-Government
development (UN, 2002; OECD, 2003), so that risk of failure increases. However, decision-makers in PA are
risk averse and prefer “to rely on the tried and true” (Bartoli, 2005, p.320). So, as expected in the hypothesis
H4, UA acts negatively on e-Government adoption by PA.
The influence of the variable PD on the e-Government adoption at all evolution stages and at the WMI is
insignificant although it is in the expected negative direction. Consequently, hypothesis H1 is not supported.
PD seems to don’t have an impact on the adoption of the e-Government by PA.
For the variable MA, the impact at all evolution stages and at the WMI is insignificant. The coefficients Beta
are negative at stages 4 and 5 and positive at stages 1, 2, 3 and at the WMI. Therefore, hypothesis H3 is not
supported, suggesting that masculine cultures as feminine cultures adopt in the same way the e-
Government.
Since Gong and al. (2007) find that the impact of LOT dimension on broadband penetration increases when
introducing the control variable “level of education” in the regression model, the variable LOT is reintroduced
among the independent variables (model 4, table 4). It appears that the five cultural variables and the
socioeconomic variable give better explanation to the variance of the e-Government adoption (the values of
the coefficient R2 increase). The Beta coefficient of the variable EDU still positive and significant becomes
higher. Therefore the hypothesis H6 is supported. But with the reintroduction of LOT, the negative Beta of
UA becomes insignificant. Consequently, the hypothesis H4 is not supported. This can be justified by the fact
that if LOT increases, the long term benefits of the e-Government can be better perceived. As a result, the
influence of UA may be counterbalanced. Although the correlation coefficient of LOT and UA is low and
insignificant (table 2) its negative sign means that these two variables act inversely.
Besides, the model 4 shows that the variable PD has a negative and significant impact on the evolution to
the stage 5 of e-Government adoption and that this impact is limited to this stage. Then, the hypothesis H1 is
partially supported. This stage implies the networking of various agencies of PA, cooperation between these
agencies, information sharing and major organizational changes (Ebrahim and Irani, 2005). Moreover, it
encourages citizens’ participation in political processes and decision making and involves the society in a
two way open dialogue (ONU, 2005). This can be in contradiction with cultural values of a country where the
PD is high: centralized structures, respect of authority, application of formal rules and instructions,
preference for status quo. The Beta coefficients of the variables LOT, ID and MA are insignificant. They are
negative for LOT and ID and positive for MA. Contrary to expectations, the hypothesis H5, H2 and H3 are
not supported and the signs directions are opposite to the proposed relationships. Table 5 gives a summary
of results:
Table 5: Summary of results
Model 1 Model 2 Model 3 Model 4
H 1 (PD) Not supported Not supported Not supported Partially supported
H 2 (ID) Not supported Supported Not supported Not supported
H 3 (MA) Not supported Not supported Not supported Not supported
H 4 (UA) Not supported Partially supported Supported Not supported
H 5 (LOT) Not supported - - Not supported
H 6 (EDU) - - Supported Supported
So, we can conclude that:
􀂃 The hypothesis H1 is rejected in the three first models. This contradicts the results of Kovacic
(2005) who finds that PD act negatively on WMI, and the results of Bagchi and al. (2004) and
Electronic Journal of e-Government Volume 6 Issue 1 2008 (11-22)
www.ejeg.com 20 ©Academic Conferences Ltd
Gong and al. (2007) according to whom the adoption of PC, telephones, phones cellular (Bagchi
and al., 2004), and Internet (Gong and., 2007) increases in low PD countries. But, the model 4
gives a minor support for the hypothesis H1, suggesting that PD may be negatively associated
with the e-Government adoption by PA.
􀂃 The hypothesis H2 is rejected in models 1, 3 and 4 and supported in model 2. In the last case,
results corroborate with those found by Kovacic (2005), Bagchi and al. (2004) and Erumban and
de Jong (2006). According to the second model, e-Government adoption increases in
individualistic cultures and ID is the most significant cultural variable. But this relationship doesn’t
remain significant with the introduction of the socioeconomic variable EDU (Model 3) and the
cultural variable LOT (Model 4).
􀂃 The hypothesis H3 is rejected in all models. So, cultural values, as masculinity/ femininity, don’t
have any impact on the e-Government adoption by PA. Similar results show that masculinity has
no relationship with the adoption of ICT in general (Erumban and de Jong, 2006), and with the e-
Government readiness index in particular (Kovacic, 2005). But, this result is contrary to those of
Gong and al. (2007) who find that masculinity has a negative effect on the access to Internet and
its use and Bagchi and al. (2004) who reveal a strong relationship between feminine culture
features and the adoption of the telephone and the cellular phone. This result may be explained
by the fact that e-Government seems to respond to both feminine and masculine values: it aims
at providing an efficient government management, empowerment of citizens through access to
information and participation in the public policy decision making (UN, 2005), enhancing the
public services quality, establishing trust between PA and the citizens and achieving economic
objectives (OECD, 2003; UN, 2004).
􀂃 The hypothesis H4 is rejected in models 1 and 4, partially supported in model 2 and supported in
model 3. So, the suggestion that adoption of e-Government increases in countries where UA is
low gives mixed results. In the same vein, Bagchi and al. (2004) find that the relationship
between UA and ICT adoption is not obvious. In addition, the test of the relationship UA – the
level of e-Government readiness is only weakly supported in the work of Kovacic (2005).
􀂃 The hypothesis H5, i.e. e-Government adoption by PA increases in the countries characterized
by LOT cultures, is not supported. This result contradicts the findings of Gong and al. (2007) who
note that this cultural dimension has a significant positive influence on Internet use and access
and that this impact increases when including the control variable “level of education” in the
regression model tested for the variable “broadband penetration”. It seems that both researchers
and practitioners have highlighted the advantages of e-Government and e-Government
programs remain at the top of most countries policy agendas (UN, 2002) independently of the
cultural variable long term orientation/ short term orientation.
􀂃 The hypothesis H6 is supported. The Beta coefficients of the variable EDU (table 4) are higher
than those of the cultural variables suggesting that EDU is more important in the e-Government
adoption. The highest coefficient is attained at the stage 2 (table 4). Although, the
communication and information flows are still unidirectional (PA 􀃆 Society), this stage is more
sophisticated than the first (UN, 2005). In the beginning, citizens and employees in the PA can
face technical problems; the resolution of these problems may require high levels of knowledge.
Then, the accumulation of experiences generates know-how and new skills that can be used to
engage in the more advanced stages of e-Government evolution.
7. Conclusion
This study is designed to examine whether cultural characteristics explain differences in worldwide e-
Government adoption by PA. It is found that results vary according to the regression model tested, and
national cultural dimensions influence the e-Government adoption. Among the cultural variables ID and UA
are the most significant and PD is of minor importance. However, results show that the cultural and
socioeconomic variables, taken together, explain better the e-Government adoption by PA. The level of
education attainment is more important since the e-Government requires the recombination of efforts and the
expertise of several stakeholders (PA, private partners in the case of outsourcing, etc.) for the development
of e-services that respond to the needs of a society. Reciprocally, to be able to use e-services, society must
detain the requisite skills.
This study has implications both for practice and theory. e-Government adoption by PA is measured by WMI
and by stages of e-Government evolution. So results identify the most significant cultural dimensions by
Fatma Bouaziz
www.ejeg.com 21 ISSN 1479-439X
stages. To take in to account the significant cultural variables when elaborating policies for e-Government
adoption by PA may be important. At a theoretical level, this study suggests to add cultural variables to other
theories and models, such as the diffusion of innovations (DOI) theory, the theory of planned behaviour
(TPB) and the technology acceptance model (TAM), to study e-Government adoption. Horst (2007), for
example, finds that perceived risk by citizens is inversely related to perceived usefulness of e-Government
services, and that this later is an important determinant of citizens’ intention to adopt e-Government services.
This research is not exempt of reservations. The first limit relates to the measurement of the national culture.
Dimensions are derived from Hofstede cultural approach. In addition to its criticisms (Baskerville, 2003;
2005), Hofstede provides regional scores for Arab world, East and West Africa. Since these scores are
regional and that countries can have different scores from those of the region, the countries belonging to
these regions are excluded from this study despite the availability of data on e-Government adoption in the
UN report (2005). Data in this report show that fewer countries evolve to stages 4 and 5 and that the full
potential at these stages is unexploited. Different approaches of the e-Government according to countries
must to be considered (OECD, 2003). Thus, to be unaware of variables such as political systems, strategic
priorities of the governments, and technical and organizational challenges (Layne and Lee, 2001; Ebrahim
and Irani, 2005) in the e-Government adoption and particularly the evolution to stages 4 and 5, is another
limit.
References
Bagchi, K., Hart, P. et Peterson, M.F., 2004, « National Culture and Information Technology Product Adoption”, Journal
of Global Information Technology Management, 7, 4, pp.29-46.
Bartoli, A., 2005, Le management dans les organisations publiques, Edition Dunod, Paris.
Baskerville, R.F., 2003, “Hofstede never studied culture”, Accounting, Organizations and Society, 28, pp.1–14.
Baskerville, R.F., 2005, “A research note: the unfinished business of culture”, Accounting, Organizations and Society, 30,
pp.389–391.
Ebrahim, Z. et Irani, Z., 2005, “e-Government adoption: architecture and barriers”, Business Process Management
Journal, Vol.11, N°5, pp.589-611.
El Jamali, T., Plaisent, M., Benyahia, H., Bernard, P. et Maguiraga, L., 2004, « La France à l’heure du e-gouvernement »,
9ème Colloque de l’AIM, « Systèmes d’information : Perspectives critiques », 26, 27 et 28, mai 2004, INT Evry
France.
El Louadi, M., 2004, « Cultures et communication électronique dans le monde arabe », Systèmes d’information et
Management, Vol.9, N°3, pp.117-143.
Erumban, A.A. et de Jong, S.B., 2006, “Cross-country differences in ICT adoption: A consequence of culture?”, Journal
of World Business, 41, pp.302–314.
Gong, W, Li, Z.G. et Stump, R.L., 2007, « Global internet use and access: cultural considerations”, Asia Pacific Journal
of Marketing and Logistics, Vol. 19, N°1, pp. 57-74.
Hall, E.T., 1976, Beyond Culture, Anchor Books, New York, NY.
Hofstede, G., www.geert-hofstede.com/hofstede_dimensions.php, access : 14-09-07.
Horst, M., Kuttschreuter, M.T., Gutteling, J.M, 2007, “Perceived usefulness, personal experiences, risk perception and
trust as determinants of adoption of e-Government services in The Netherlands”, Computers in Human Behavior, 23,
pp.1838–1852.
Kovacic, Z.J., 2005, “The Impact of National Culture on Worldwide e-Government Readiness”, Informing Science
Journal, Vol.8, pp.143-158.
Layne, K. et Lee, J., 2001, “Developing fully functional e-Government: a four stage model”, Government Information
Quarterly, Vol.18, N°2, pp.122-136.
Moon, M.J., 2002, “The evolution of e-Government among municipalities: Rhetoric or reality?”, Public Administration
Review, 62, 4, pp.424-443.
OCDE, 2003, “The e-Government imperative”,www1.worlbank.org/publicsector/egov/e-GovernmentImperative.pdf, 199p,
access: 5-03-04.
ONU, 2002, “Benchmarking e-Government: A Global perspective, assessing the progress of the UN member states”,
81pp. www.golconferenceca/presentations/e-Government_UN.pdf, access: 5-02-04.
ONU, 2003, “World Public Sector Report 2003: e-Government at the Crossroads”,
unpan1.un.org/intradoc/groups/public/documents/UN/UNPAN012733.pdf, access: 05-09-05, 129p.
ONU, 2004, “Global e-Government readiness report 2004: Towards access for Opportunity”, New York, 182p.
http://unpan1.un.org/intradoc/groups/public/documents/un/unpan019207.pdf, access: 29-08-05.
ONU, 2005, “From e-Government to E-inclusion”, UN Global e-Government Readiness Report,
unpan1.un.org/intradoc/groups/public/documents/un/unpan021888.pdf, pp.1-270, access: 21-09-07.
Qian, H., 2007, “UNDESA Initiative on Global e-Government Assessment”, pp.1-10, access: 21-09-07,
unpan1.un.org/intradoc/groups/public/documents/APCITY/UNPAN024987.pdf.
Rosenberg, N., 1972, “Factors affecting diffusion of technology”, Explorations in Economic History, 10, 1, pp.3-33.
World Bank list of economies (July 2007), siteresources.worldbank.org/DATASTATISTICS/Resources/CLASS.XLS,
access: 20-09-07.
Electronic Journal of e-Government Volume 6 Issue 1 2008 (11-22)
www.ejeg.com 22 ©Academic Conferences Ltd
Zghal, R., 2001, « Technologies de l’information et de la communication, administration publique et culture en Tunisie »,
5ème conférence internationale de management des réseaux d’entreprises, Octobre 2001, Tunisie - Mahdia, pp.1-8.
Appendix 1
WMI and characteristics of the 50 first countries (UN, 2005)
Countries WMI Rang Income/ capita* Development state
1 USA 1.0000 1 High
2 United Kingdom 0.9962 2 High
3 Singapore 0.9962 2 High Developing
4 Republic Korea 0.9769 3 High
5 Denmark 0.9731 4 High
6 Chile 0.9115 5 Superior to the average Developing
7 Australia 0.9038 6 High
8 Canada 0.8923 7 High
9 Sweden 0.8654 8 High
10 Germany 0.8423 9 High
11 Finland 0.8269 10 High
12 Mexico 0.8192 11 Superior to the average Developing
13 Japan 0.8154 12 High
14 New Zealand 0.8038 13 High
15 Norway 0.7962 14 High
16 Malta 0.7923 15 High
17 Brazil 0.7500 16 Superior to the average Developing
18 Austria 0.7423 17 High
19 Philippines 0.7423 17 Lower to the average Developing
20 Holland 0.7346 18 High
21 Israel 0.7308 19 High
22 Belgium 0.7115 20 High
23 Ireland 0.7115 20 High
24 Hungry 0.7038 21 Superior to the average Developing
25 Estonia 0.6962 22 High
26 Thailand 0.6654 23 Lower to the average Developing
27 Argentina 0.6577 24 Superior to the average Developing
28 Romania 0.6423 25 Superior to the average Developing
29 Mauritius 0.6288 26 Superior to the average Developing
30 Italy 0.6269 27 High
31 Colombia 0.615 28 Lower to the average Developing
32 France 0.6115 29 High
33 United Arab Emirates 0.6115 29 High
34 Iceland 0.6077 30 High
35 Suisse 0.6038 31 High
36 Slovenia 0.5923 32 High
37 Czech Republic 0.5885 33 High
38 India 0.5827 33 Low Developing
39 Ukraine 0.5808 34 Lower to the average Developing
40 Malaysia 0.5769 34 Superior to the average Developing
41 Venezuela 0.5769 34 Superior to the average Developing
42 Chine 0.5692 35 Lower to the average Developing
43 South Africa 0.5692 35 Superior to the average
44 Peru 0.5577 36 Lower to the average Developing
45 Slovakia 0.5385 37 Superior to the average
46 Lithuania 0.5231 38 Superior to the average
47 Turkey 0.5231 38 Superior to the average Developing
48 Bulgaria 0.5192 39 Superior to the average Developing
49 Greece 0.5115 40 High
50 Poland 0.5115 40 Superior to the average
*World Bank list of economies (July 2007), siteresources.worldbank.org/DATASTATISTICS/Resources/CLASS.XLS, access: 20-
09-07.