A new Washington Post-ABC News poll that is making lots of headlines, both nationally and locally, reported that opposition to “gay marriage” has fallen to a “new low,” with only 39 percent of the public opposed. It would seem noteworthy that less than four in 10 Americans oppose “gay marriage” – if not for the poll’s biased sample.
As an aside, a poll’s bias can be examined fairly simply, by taking a few moments to examine the background questions asked of poll respondents. This rather quick and painless quality-assurance process would seem to be a no-brainer for any journalist aspiring to achieve balanced, quality reporting. Yet even such basic questions about sample validity are often lost on many political reporters.
Looking at the poll’s results, we find that the partisan breakdown was 32 percent Democrat, 22 percent Republican, and 38 percent independent. This 10-percentage-point Democratic advantage in the polling sample stands in stark contrast with the most recent party ID polls, which show Democrats with only a 4-point advantage in one case, and even a 2-point disadvantage in another. Clearly, a poll that artificially depresses the number of Republicans in its sample is going to find “all-time lows” for opposition to “gay marriage,” not based on any reality but rather on a skewed polling sample.
Here’s hoping that those who claim to be political journalists will act like it when it comes to numbers fed to them by pollsters.