At the ABC News Polling Unit we’re occasionally asked about the response rate to our polls. Such inquiries rarely, if ever, specify what is meant by the term. No wonder: it’s an issue fraught with complexity. In fact, there are serious, unresolved questions in analyzing and comparing response rates, and in evaluating the tradeoffs between response rates and other methodological considerations in survey research. As practitioners it’s incumbent on us to understand these issues as best we can, and to evaluate their practical significance in our work. The American Association for Public Opinion Research has issued standard definitions for the sample disposition codes by which response rates are calculated. These are useful, but they’re not the full story. Standards have not been promulgated for all the various sample design, sample management and respondent selection procedures that can affect response rates; indeed, AAPOR notes, “Comparisons between polls with different sample designs is not possible.” And the AAPOR standards themselves are complex, offering four separate methods of calculating cooperation rates, three for calculating refusal rates, three for contact rates, and six for overall response rates.