ABSTRACT
After pre-election polls predicted the wrong winner of the 2008 Democratic primary in New Hampshire, a Washington Post headline asked “Can we ever trust the polls again?”1 Concerns about the increasing methodological challenges facing survey research in recent years have undermined confidence in the entire survey enterprise. Surveys rely on the cooperation of people to check boxes and answer questions, yet people today are harder to reach, and when contacted they are less likely to answer questions. At the same time, there has been a proliferation in the amount of polling-from horserace numbers in the newspaper headlines to opt-in “polls” predicting sports outcomes on ESPN.com or judging celebrity outfits in US Weekly magazine. With so many polls, it is hard to figure out which ones are accurate and reliable.
