Skip to main content

View Diary: Weekly Tracking Poll: Ch-Ch-Ch-Changes (92 comments)

Comment Preferences

  •  and here's that right-wing troll (2+ / 0-)
    Recommended by:
    dotster, mdmslle

    you mentioned.

    Likely voters in March 2010 are not likely voters in November 2010. Every approach has its drawbacks and will generally fail to be an accurate predictor of future events. Likely voter models are obviously better for approaching elections but just using registered voter models gives you a useful snapshot of the entire potential electorate. As long as the pollster is being transparent about the model, there's no problem and neither is more "correct" than the other. The only problem, evidently, is that it burns your ass that a majority of Americans have a favorable opinion of Obama.

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site