Reply #25: You may not want to call them exit polls,

[Igel]

because they didn't actually collar people as they left their precincts; you're right, the Oregon survey wasn't an exit poll. But the Oregon survey served the same purpose: to help call the state and to provide demographic breakdowns of who voted how before the actual returns were in.

In that case let's call them "polls taken for the purpose of assisting NEP clients". Exit polls are one kind, another is the absentee telephone survey Mitofsky did. For most states, the phone survey was an adjunct to the "main" exit polls; in Oregon, it was the only survey possible.

See the much bally-hooed report, p. 72, where Mitofsky compares the absentee telephone survey estimate with actuals.

Mitofsky overstated Kerry support by 1.2, and understated Bush support by 0.2.

In any event, TIA isn't exactly right: Mitofsky's figures for Oregon didn't match the certified results, but I can't say if they were within the calculated error.