[starbase]

Well, regarding the Field Poll, this link goes to their home page where they list their history of major races and ballot measures, etc.

They point out that the final Field Poll taken before an election has been an excellent indicator of race outcomes, ranging from 64% in some ballot measures, to almost 100% in major races.

They also bring up the academic point that comparing the election results to the poll results may not be the most accurate measure of the poll's accuracy for the time at which the poll was taken, which is a scientific, but very fair point.

All in all, with its high record of final poll accuracy, plus the fact that it's been around 56 years, perhaps the Field Poll is "venerable".

[calcowgirl]

There are some interesting comments about Field in this interview with Arnold Steinberg, a Republican pollster/strategist that is a straight-talker.

http://www.flashreport.org/special-reports0b.php?faID=2005110202424596

[ElkGroveDan]

The Field Poll had Richard Riordan beating Bill Simon 56% to 21% in January of 2000, a mere 60 days before the election.

Big deal, they are good at the final poll in major races. Since there are rarely major upsets in the final weeks, so is anyone who reads the news.

The Field poll has some major, major flaws, starting with their sampling. They sample counties based on population not on the number of people in those counties that actually ever vote. So San Francisco, for example, is always oversampled. It's why things always look bad for Republicans and conservatives early on in the Field poll, and then get better later.

The really good private pollsters can call successfully races much farther out than Field does.