The research team at Zogby International had the pleasure of attending the American Association for Public Opinion Research (AAPOR) recently, and we viewed some very interesting studies devoted to various issues in polling the 2008 US presidential election. An interesting paper presented by the Associated Press and GfK Roper Public Affairs and Media looks into the dilemma between the need to collect and report data quickly in a fast-changing news cycle and the need to devote enough survey time to avoid potential pitfalls that may result from low response rates (Note: the response rate refers to the number of respondents who completed the survey divided by the number of eligible respondents in the sample.) Their pre-election polls were conducted over a period of 4 to 5 days and allowed for up to 10 callbacks, more frequent than regular 4 to 6 callback attempts. Results suggest that the additional calling attempts had little impact on the horserace estimates and the representativeness of the achieved sample. The impact of staggering calls over a longer survey period varies depending on the volatility of the campaign dynamics. The impact is strong during the period when a campaign is particularly volatile.

Do you think it is necessary to make repeated attempts over an extended period of time to achieve a representative sample of the voters? Do you think the best pre-election polls should be “snapshots” at a certain point in time or “slideshows” reflecting potential movements over a longer period of time?

Grace methodology international, media, polls, race, zogby