Skip to main content

View Diary: Doctors Like Public Option, Especially As A Choice (126 comments)

Comment Preferences

  •  I don't know that that's really the case (0+ / 0-)

    all polls are like that (theoretical, don't hold me to it; if the election were today, even though it's not...)  and that leads to the guy behind saying "the only poll that counts is election day." When you hear that, you know the guy is behind, and he's challenging the poll. Nonetheless, he loses.

    Exit polls have a true differential non-response issue... young college kid surveyors might not get a good response from conservative seniors (the "shy" voter). But given the combo of equivalent demographics including, importantly, regional matching and the very large difference in preferences for public option/private option vs. private alone, I don't see a case being made for invalidation of the survey based on non-responders being "different". A greater case could be made if the results were closer.

    Phone polls often land in the 20% response rate, with weighting making the demographics match the census. Once again, non-responders have to have a good reason to not match the responders before dismissing the poll. It isn't non-response, it's differential non-response that matters more. Again, accuracy may suffer, but when the guy is up 73-27, a few points here or there won't change the election.

    The key is how well the sample is chosen, not simply the size of the response, though it is an issue. But the sample size and methodology seems big enough and sound enough in this case to avoid the type II errors seen in tiny differences in clinical trials, which this should not be compared to.

    "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

    by Greg Dworkin on Tue Sep 15, 2009 at 05:02:09 AM PDT

    [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site