Does External Environment Affect E-Government? A Cross- Country Analysis - Information and Communication Technologies in Public Administration - page 68

table 3.1 Descriptive Statistics

Variable M SD Min Max

Dependent Variable

E-government 0.31 0.21 0 1

independent Variable

Internet usage 32.4 27.6 0.72 93.27

Democracy 5.57 2.14 1.52 9.8

Education 0.63 0.21 0.18 0.99

Economy 12.25 13.52 0.32 61.24

Peer pressure 0.31 0.16 0.02 0.94

Note: The final sample size is 129.

national income (GNI) per capita (current $1,000) in a country was used as a unit
of measure. The democracy score came from the Democracy Index of 2010 from
the Economist Intelligence Unit. Based on the report, the index provides a snapshot
of the state of democracy worldwide for 165 independent states and two territo-
ries (Economist Intelligence Unit, 2011). The Internet usage percentage from the
International Telecommunications Union's 2010 data was used to measure Internet
usage. The average e-government score of a country's land borders was used to mea-
sure peer pressure. For example, the level of peer pressure exerted upon Germany
depended on the average of the e-government scores of Austria, Belgium, Czech
Republic, Denmark, France, Luxembourg, the Netherlands, Poland, Sweden,
Switzerland, and the United Kingdom. The descriptive statistics are in Table 3.1.

3.4 Results
To estimate the accuracy of the structural model, the maximum likelihood esti-
mates for the model coefficients were first obtained with the STATA software. The
model fit the statistics, showing that the model was accurate. The root mean square
error of approximation was 0.053, which met the criteria set forth by Arbuckle
(2006), who argued that a root mean square error of approximation between 0.05
and 0.08 is typically interpreted to represent between a “close” and a “reasonable”
model fit. In addition, the comparative fit index was 0.998, and the Tucker-Lewis
index was 0.991. Lastly, the coefficient of determination was 0.863, which is also
very high. Overall, the model (as shown in Figure 3.3) proved to be strong, and
most of the hypotheses were confirmed.