Methodology for the 2004 National Annenberg Election Survey
This research is based on data from NAES04, a rolling cross-section telephone survey that was in the field continuously from October 7, 2003 through November 16, 2004. Respondents are selected for cross-section study samples via a two-stage process-Random Digit Dialing and random selection within household. For this study, households were first selected by randomly generating telephone numbers. Area code, exchange, and bank, representing the first eight digits of a ten-digit phone number, were randomly generated proportional to telephone company estimates of the count of residential numbers in each combination of area code, exchange, and bank. The last two digits of each phone number were generated entirely at random. Second, adult residents of a household were selected randomly based on a specific algorithm (For complete design, see Daniel Romer, et al., Capturing campaign dynamics:The National Annenberg Election Survey.)
Rolling cross-section sampling is implemented by adding new randomly generated telephone numbers on a strict schedule to the pool of numbers interviewers are calling to attempt to complete interviews. On each day of fieldwork, a set count of new numbers is added, proportional to the desired count of interviews completed daily. The intention is to maximize the representativeness of any single day's interviewing sample by including those respondents who are easy to contact and those who are more difficult to contact. Consequently, the day on which a respondent is interviewed may, for purposes of analysis, be considered a random event. Each respondent was called back a maximum number of eighteen times, and refusal conversions were employed.
|
