How the Study Was Conducted
To prepare this study, COPA conducted a
nationwide poll, focus groups, and a comprehensive review of previous
polls done by other organizations.
The poll was conducted January 26-31, 1999
with a sample of 1,204 American adults. Communications Center, Inc.
interviewed respondents by telephone on a CATI system, using a questionnaire
designed by COPA. Respondents were chosen from all households in the
continental United States by a random digit dialing sample generated
by Scientific Telephone Samples. Interviewers observed gender quotas.
The margin of error was +/-3-4%.
The order and placement of some questions
were varied to reduce any biases that might derive from question order.
Data for this survey were collected using
telephone interviews with Americans 18 years or older living in the
continental United States. The sample frame was generated, using random
digit dial methods, by Scientific Telephone Samples. The telephone exchanges
for this sample were drawn from residential working block exchanges
excluding blocks assigned exclusively for business use, mobile phones,
military or governmental purposes, and known business numbers. Selection
from these working blocks was weighted according to the estimated number
of working residential telephones within each. The exact number of RDD
numbers generated per working block was calculated proportional to the
estimated working residential telephones for the particular working
block against the total estimated working telephones for the entire
sampling frame. Estimates of household telephone coverage were derived
from census data on residential telephone incidence and updated with
information from local telephone companies and other sources and cross-checked
with Bellcore files. For the purpose of this study, a working bank was
defined as those with more than three known working residential telephones
out of the 100 possible numbers within that block.
The sample was released for interviewing
in replicates. Using replicates to order the sequence of calls eliminates
potential calling order bias based on region.
The Focus Groups
COPA conducted focus groups to help develop
poll questions that would reflect how people think and talk about the
issues of the publics role in government decisionmaking. Focus
groups provide citizens with an opportunity to talk about their views
and feelings in their own words, and to explain the underlying assumptions
behind their views.
COPA conducted three focus groups: the
first in Baltimore, Maryland on September 1, 1998; the second in Albuquerque,
New Mexico on December 26, 1998; and the third in Roanoke, Virginia
on January 7, 1999. Each discussion lasted about two and half hours
and included 11-13 participants. Maryland Market Source, Inc. recruited
participants for the Baltimore focus group; Business Information Group,
Inc., recruited for the Albuquerque group; and Martin Research, Inc.
recruited for the Roanoke group. All three organizations made a strong
effort to recruit groups that reflected the demographic makeup of their
Review of Other Polls
COPA performed a comprehensive review of
publicly released polls on issues related to public attitudes toward
the federal government. The primary sources were the Public Opinion
Location Library database of the Roper Center for Public Opinion Research
at the University of Connecticut, and the Inter-University Consortium
for Political and Social Research. (Special mention should be made of
a useful recent compendium: Everett Carll Ladd and Karlyn H. Bowman,
Whats Wrong: A Survey of American Satisfaction and Complaint (Washington:
AEI Press; Storrs, CT: The Roper Center for Public Opinion Research,
University of Connecticut), 1998.) The full archive of National Election
Studies data was studied via the NES website, www.umich.edu/~nes.