So right in the middle of all the yelling and screaming over polling methodologies, Gallup responds to long-standing criticisms (particularly from Mark Blumenthal) and changes its ways of choose its base sample of adults in ways that will likely make its assumptions closer to the U.S. Census breakdowns of demographic groups, and perhaps its findings closer to those of other traditional polling firms.
In announcing these changes, moreover, Gallup’s Frank Newport offers one of the more succinct critiques of partisan ID “weighting,” and makes it clear Gallup’s LV screen is based on questions about likelihood to vote, not assumptions about the makeup of the electorate:
As has always been the case, we do not attempt to weight the composition of the likely voter sample in any way — such as by political party or race or age — to approximate some guess of what we or others think it should look like demographically on Election Day. That approach is precarious given that the electorate can look quite different (especially looking at political parties) from one election to the next. Also, party identification estimates are often based on exit poll results, which themselves are surveys using totally different methodologies, and we generally do not rely on judgment calls to predict what the ultimate electorate will look like. Our basic underlying sample of national adults is weighted to known population parameters on demographic and phone use variables, as noted, and the likely voter pool derives from that, based on how our randomly selected respondents answer the likely voter questions.
You can expect conservatives to claim Gallup has caved to “liberal” pressure and is counter-weighting its samples, but that will confuse an effort to keep the base sample close to the actual demographics of the country with the conservative demand that pollsters guess which party will succeed in turning out its voters.
