Props to Survey USA, Roanoke College.

I don’t give Survey USA a lot of credit. Their methodology strikes me as sloppy, their results wild, and their whole approach sort of like McPolling. Just the day before the election, I said that they don’t know their ass from their elbow.

But I was wrong. Or, at least, they were right. They released polling numbers the day before the election that seemed nuts — they had Lt. Governor Tim Kaine up by 9% and then, later in the day, they released more updated numbers that had him up by 4%. Nine percent was a number so wildly above all expectations that it was scarcely believable. But with Kaine winning by 6%, it turned out that their numbers were closer than anybody else’s.

Ditto for Roanoke College. They released the results of a poll on November 2 that had Kaine up 8%. Nobody paid it much mind because a) what does some college know about polling and b) it was just so out of the blue that it wasn’t convincing. In fact, they were the first to pick up on the trend of Kaine’s quickly-rising support. And they, too, deserve credit for releasing numbers that were, at least in the blogosphere, widely dismissed. We were wrong about that.

I’m not sure that I feel much better about Survey USA’s polling practices, but the fact is that they got this one right, and that’s impressive.

Published by Waldo Jaquith

Waldo Jaquith (JAKE-with) is an open government technologist who lives near Char­lottes­­ville, VA, USA. more »

One reply on “Props to Survey USA, Roanoke College.”

  1. what does some college know about polling

    After looking at VT’s survey tool a short while back, since two organizations I belong to are interested in running fairly broadbased polls, I’m sort of wondering the same thing. But hey, VT at sourceforge?

    Too cool…

Comments are closed.