R
ETHINKING THE
D
IGITAL
D
IVIDE
9
focus is on the pattern of Internet use and gratification gained within categories defined by the
most persistent digital divide categories — age and socioeconomic status. These individuals
(N=2,752) constitute 45.2 percent of the total sample in March survey. Also, in the following
analyses, further reductions in sample size occurred because of missing data for some variables.
Measures
Control variables. There were two sets of control variables: demographics and the basic
digital divide issues. Five demographic variables were included as exogenous controls in our
model
−
respondent’s gender (dummy variable with female coded 1), age, race (dummy variable
with white coded 1), level of education, and income (see Appendix for exact wording). Basic
digital divide issues were also controlled to rule out their potential impact on the relationship
among the variables of our interest. Internet access location (whether or not respondents use the
Internet at home), frequency of Internet use (how often respondents use the Internet), length of
Internet use (how long respondents have used the Internet), and e-mail use (whether or not
respondents ever used e-mail) were included to be controlled out.
Antecedent Variables. In order to identify patterns of Internet use, we performed both
exploratory factor analysis and reliability tests. Based on factor analysis, three types of Internet
activities were identified: surveillance use, consumption, and computer-mediated interaction.
Surveillance use of the Internet was measured by using five items (
α
= .55). Respondents were
asked if they ever did any of the following when they went online: 1) get news online, 2) do
research for school or training, 3) check weather reports and forecasts, 4) work or research online
for your job, and 5) look for news or information about politics and the campaign. An additive
scale was created based on these five items Consumptive use of the Internet was measured by three
Convention