USE OF COMMUNITY ICTS 89Demographic

USE OF COMMUNITY ICTS 89
Demographic variables.—This study selected four demographic variables
for investigation: gender, age, education, and income. Each variable was measured
by a single questionnaire item.
Psychological variables.—Three psychological variables (attitude, subjective
norms, and perceived behavioral control) were measured by adopting
Ajzen’s measures [30, 33–34, 38]. Sense of community was measured by
adopting an existing psychometric scale [39]. Each variable was measured
with multiple items on seven-point Likert-type scales. Multiple-item measures
are known to enhance item variability and reliability for latent psychological
variables [40].
Alternative service accessibility.—Alternative service accessibility was measured
with a dichotomous variable. It was measured by asking, “Do you
currently use an Internet service provider other than [name of CN]?”
Length of experience.—Length of experience was measured by asking, “How
long have you been using (or did you use) [name of CN] if you have used
it before?” Respondents were asked to record the number of years of using
the service in the blank, “About __ years in total OR about __ months in
total.”
Statistical Analytic Techniques
The current study used hierarchical multiple regression (HMR) as the
major statistical data analysis technique. A regression analysis that can
explain a phenomenon with multiple predictors, HMR is a technique that
sets causal priorities and removes spurious relationships among predictors.
In this technique, a regression model was constructed by entering the
predictors in a certain sequence that is predetermined by a theoretical
ground [41]. Generally, in a hierarchical equation, the predictors that are
previously known to affect the dependent variable are entered first. Then,
the variables for which the researcher wants to show effects are added into
the equation. The test results provide estimates of the unique contribution
of the last entered variables in explaining dependent variables, above and
beyond the contribution of the variables entered into the equation earlier
[42]. Hierarchical multiple regression also assesses the unique contribution
of each predictor in explaining the dependent variable by controlling for
the influences of all other predictors in the research model. A specific
regression model testing each research question is detailed in the next
section.
Results
Four hundred twenty-two out of one thousand sampled users returned the
survey. Three hundred ninety-four usable responses were used for data