[Moosilauke]

Exit polling is not something one does as a regular job. It's likely that many, or maybe most, of these exit pollers were academic types--grad students, professional students, instructors. Working tax-payers don't have a day to sacrifice working a one-off menial job like exit polling.

Who are these pollers most likely to try to interview? People like themselves. And who is most likely to respond to one of these pollers? Someone who recognizes the poller as being "like me".

It's no surprise the pollers found almost everyone had voted for Kerry. It's like the famous Pauline Kael complaint: "How did Nixon get elected. I don't know anyone who voted for him!"

[Strategerist]

Having worked part-time for the 2000 Census, I think it's more who is willing to stand around and answer a long list of questions from a stranger.

By a huge margin, I had the most difficult time, and the most doors slammed in my face, in wealthy, white areas (in a part of the country where these areas are overwhelmingly Republican.)

I had the easiest time and the most cooperation in lower-income and minority areas (which sort of made me wonder about all the constant claims that the census undercounts minorities.)