2007-2008 AP-Yahoo News Election Panel Study
Survey Samples
The goal of each wave of data collection was to maximize the number of follow up completes from the baseline sample. Therefore, all baseline respondents who were still available for survey sampling from the Knowledge Networks panel were fielded follow up waves, even if they had failed to respond to a previous wave. In addition to the longitudinal samples, three fresh cross-sectional samples, goal 600 completes per, were fielded at the discretion of AP. These fresh samples were conducted at wave 3, wave 6 and wave 9.
The vote preference and opinions of younger respondents, age 18-29, were particularly important to AP-Yahoo as a large young voter turnout was expected in the November election. To obtain another measure of this demographic group, Knowledge Networks conducted a survey with a small fresh cross-section of young respondents age 18-29 between waves 5 and 6. A total of 261 panel members completed the Young Sample survey and in order to increase the sample size of young respondents in the on-going longitudinal sample, these cases were followed for the remaining waves of the study.
At wave 6 an additional stand-alone survey, wave 6 AMP, was conducted with the baseline sample and young sample; the wave 6 fresh sample was not sampled for the AMP survey. The AMP survey incorporated an implicit test for racial attitudes called the "Affect Misattribution Procedure" (AMP) which was developed by B. Keith Payne, now at the University of North Carolina Chapel Hill. It consisted of a rating exercise of images of black and white people and drawings of characters from the Chinese language. Respondents were shown a quick flash of a person's face before each drawing and were asked to rate the drawing as more pleasant or less pleasant than average. Response time and the rating selected were recorded. For a description of the methodology behind the AMP, see Payne, B. Keith, Clara Michelle Cheng, Olesya Govorun, and Brandon D. Stewart. 2005. "An Inkblot for Attitudes: Affect Misattribution as Implicit Measurement." Journal of Personality and Social Psychology 89: 277–93.
The wave 10 survey, or Election Day survey, began on October 31st, 2008, however the majority of the data collection occurred from 8AM local time to 11PM Pacific time on Tuesday November 4th. Respondents of wave 9 who indicated they voted early were fielded the Election Day survey on October 31st (N=249). The remaining sample received the survey 8AM local time on the 4th. The Election Day survey was also fielded to the wave 6 fresh sample, in addition to the baseline and Young Sample. Pre-notification emails were sent to panel members on 10/30 and 11/3 to inform them of the survey. Reminder phone calls were made on the afternoon of the 4th.
A post election survey, wave 11, was conducted in November and December 2008 and included the baseline and Young Samples. In addition, 210 new Hispanic completes were collected using KnowledgePanel LatinoSM. These survey interviews were conducted online in Spanish.
