8
Residential telephone numbers were drawn from the Fall 2001 edition of the
national Select Phone telephone software package on compact disc and database
published by InfoUSA (Select Phone, 2001). Random residential telephone numbers were
generated on a state-by-state basis from the database of over 100 million residential and
commercial telephone numbers using a table of random numbers and the random number
selection function built into the Select Phone software.
2
Each resulting state roster of
chosen residential numbers was adjusted using the one-up and one-down last digit
method to include unlisted numbers. If there was no answer, a busy signal, or an
answering machine, up to two more attempts were made later that evening or during
another evening to reach the number. The response rate, excluding invalid contacts, was
56%.
Other variables included a series of demographics and self-reported political
orientations: gender, age in years, race collapsed as White and Black, Hispanic ethnicity
(yes or no), a five-point scale of income categories, a seven-point scale of education level
categories, and a five-point self-reported political orientation scale ranging from very
conservative to very liberal with a “middle of the road” midpoint.
News Use
Newspaper use. This was measured by the item: “How many days a week, if any,
do you read a printed newspaper?” Only those respondents who read a newspaper one
day a week or more were examined so long as they also qualified as an online news user.
Online news use. Respondents were asked whether they used the Internet or the
Web. If they were Web users, they were then asked how many days a week, if any, they
read an online news media outlet and then how many days a week, if any, they accessed
Convention