|
The ANES Pilot Studies During the past non-election years, ANES has undertaken a series of research and development projects, the ANES Pilot Studies. The Pilot Studies provide the opportunity for new instrumentation to be tested, the measurement of core concepts to be improved, and innovations in survey methods to be tried out. The Pilot Studies enable thorough testing and evaluation of the reliability and validity of proposed new items before a decision is made to put them on the fall election year studies. Table 1 summarizes design and content of each of the ANES Research and Development Studies. ANES undertook its first Pilot Study in 1979, the only Pilot conducted using face-to-face
interviews. A sample of 280 respondents were drawn from 30 primary areas distributed across the
country, with a re-interview conducted three weeks later. The 1983 Pilot Study followed some of the 1979 design features, altered others, and became the
model for many subsequent Pilot Studies. Rather than selecting a fresh cross-section sample, the 1983
Pilot Study drew a subsample of respondents interviewed in the previous year's National Election Study.
Two forms of the questionnaire were administered in each of the two waves. In 1985 respondents, selected from the 1984 Pre-Post Study, were stratified by political interest,
and included an oversample of the elderly. Two forms of the questionnaire were administered in both
waves. In 1987, the Pilot Study sample was disproportionately allocated across five strata defined using
1984 ANES quintile ranges for a special voter knowledge and interest index. Two forms of the questionnaire
were administered in both waves to a sample of respondents from the 1986 ANES Post Study. In 1989, a 'political information index' was used once again, modeled on the 1987 index. Four forms were
used in both waves to reinterview a subsample of 1988 ANES Pre-Post respondents. In 1991, the traditional Pilot Study design was modified to serve not only in its traditional capacity as
pilot for new instrumentation but also to accommodate the desire of the research community to understand the
political impact of the Gulf War. Three forms were administered in a single wave of reinterviews of the
respondents from the 1990 Pre-Post Study. In 1993, with so many new Representatives, special opportunity was present to examine how new members of
the House attempt to secure their districts. For this reason, the dual-purpose 1991 Pilot Study design was
replicated, however reinterviews were attempted only with the 1992 'fresh' cross-section cases, using four
forms in both waves. The 1995 Pilot Study conducted a single wave of reinterviewing with a randomly selected subset from the
fresh cross-section portion of the 1994 Post-Election Study. Two forms of the questionnaire were used. In 1997, a reinterview of a subset of respondents from the 1996 Pre-Post Study was conducted using a single
form. All 1996 Pre-and Post fresh cross section cases with telephones were included, with the balance selected
from 1996 Pre-and Post panel cases. 1998 Pilot Study was the first NES Pilot study conducted during an election season, that is,in the context
of an electoral campaign. Three high-profile gubernatorial contests were investigated, in California,
Illinois, and Georgia. The 2000 Pilot sample comprised a selection of cases from the 1998 ANES Post study. The 2000 Pilot selection
procedure first grouped all 1998 respondents according to 5 levels of political knowledge and interest. A
single pre-election interview was conducted using a single form of the questionnaire. Additional Research and Development EffortsAlthough much less than in the Pilot Studies, there is some experimentation in conjunction with the
traditional ANES Time Series studies that are conducted in presidential and midterm election years. ANES
conducted the Methods Comparison Project (MCP) together with the 1982 Post-Election Study. The MCP was
designed to examine the impact of the telephone mode of survey administration on standard NES questions.
998 interviews were conducted by Computer Assisted Telephone Interviewing (CATI) immediately after the
November 1982 elections, using questions from the 1982 Post Study. ANES has also undertaken experiments in other Time Series studies to gauge the differences between
interviewing respondents by telephone rather than face-to-face. A random half of the respondents in the Post
wave of the 1984 Pre-Post Study were re-interviewed by telephone; the other half face-to-face. Because
respondents to the pre-election interview were randomly assigned to the post-election survey mode, this study
design permits mode effects to be estimated independent of sample biases (which are different for samples
constructed over the phone rather than face-to-face). The experiment was repeated in the post-election
reinterviews of the 1996 Pre-Post Study. The 1998 Study implemented experimentation to investigate possibilities for maximizing use of telephone
interviewing, and most interviews in 1998 were conducted by phone, although traditional probability area
sampling was used and every effort was made to obtain initial face-to-face contact. In 2000, design
innovation with mixed mode was carried even further, so that the 2000 data actually represent two presidential
studies: the core study preserved past commitment to probability area sampling and face to face interviewing,
while, supporting the core study, the efficiencies of RDD sampling and telephone interviewing were used for a
second set of respondents. |