(2) Eileen/Kell: an internal poll does not mean a biased poll. There is no question that the Moore poll is an internal one. The RCP link clearly indicates that the poll was taken by a GOP firm for the Rossi campaign — that’s the definition of internal poll.

That does not mean, of course, that the poll is biased, and I don’t think that’s what Kell meant. What she meant is that it is important to never fully trust an internal poll until you get confirmation from an independent poll. You never know what other questions were asked, how much leaners were pushed, why this poll is being released rather than other ones, etc. That’s why I wrote that we should take the results from the polls in PA-03 and OH-16 with a grain of salt, for instance.

In this particular case, the result of the internal poll and of the independent poll are not meaningfully different from each other and both confirm what we already knew — that the race will be tight.

Not quite as internal as you insinuate. See David Postman’s review of the poll…

“By the way, that poll mentioned in the Gregoire release was done by Moore Research in Portland. Blogger Daniel Kirkdorffer asked pollster Bob Moore the question that others didn’t before publicizing the results. He wanted to know specifically what question voters were asked about the governor’s race. Moore told Kirkdorffer in an e-mail the question was:

Looking ahead, if the election for Governor was held today and the candidates were (rotate names) Dino Rossi, prefers GOP party, and Christine Gregoire, prefers Democratic party, for whom would you vote?”

http://blog.seattletimes.nwsource.com/davidpostman/2008/07/insert_texas_pun_here.html

Doesn’t sound like a biased poll to me.

You are showing a Rossi sponsored “internal” poll there, versus one that is more independent. From your link:

“The poll, conducted by Moore Information, a Republican firm based in Portland, Oregon, surveyed 400 registered voters between 7/9-10 for a margin of error of +/- 4.9%. Moore Information has done work for the Republican Governors’ Association before, but this poll was not sponsored by the RGA. (Major update: The poll was conducted for Rossi’s campaign. We regret the oversight.) Gregoire and Rossi were tested.”

A majority of polls has the Democrat up a few points, which seems to be the consistent thing. This poll shows it tied and it’s from a GOP related firm.

“General Election Matchup
Gregoire……..45 (-2 from last, 7/07)
Rossi………….45 (+2)”

http://www.realclearpolitics.com/politics_nation/2008/07/all_tied_up_in_wa.html