The do not call list does not apply to politics if I remember right. (I think there's an exemption for political phone calls.)
All the rest of what you say makes sense to me as well. Most polls I see need to have a plus or minus 3% over/under to be valid as well as a sampling of 1000 called. Someone can correct me if I'm wrong but that's what I remember.
I usually add 5-8% for Republicans to any poll the press publishes to compensate for the biases and sloppy polling. That's been pretty accurate for the past number of years. The polling companies are in business to sell their polling service so the poll they are most interested in having accurate is the final one before an election. All the prior polls can be explained away as sampling based on information at the moment, and things change daily. Both political party "internal" polls are more accurate, IMHO, because the polling company is held to a higher standard of performance.
Didn't know there was an exemption for political calls - thanks for that.
FWIW Bush has been polling fairly steadily between 40 and 42 on Rasmussen for several weeks now - I tend to trust his polls more than any other)
Here's another example of how I think the DBM tries to change Public opinion on polls that I posted to a thread Oct 4. Using headlines to make it look like Bush's # has have dropped to the 30s. It isn't until you get down in the article to find out the new poll shows his #s at 39 and there is a margin of error at 3.5%, which means his #s may have not changed!:
Here's an example of something that was on the web at msnbc.com this a.m. As w the NY Slimes and other MSM, I think they count on a lot of people to read the headline and not necessarily read the whole article
Here's the headline:
Sex scandal, Iraq book take toll on Bush, GOP
NBC/WSJ poll: Bush back in 30s; Dems favored for control of Congress
If you go down the page, you'll see this - -note that the margin of error rate is 3.5%, which could mean the President's numbers haven't changed at all or not much (at least they admit that the drop is within the margin of error) - but that's not what the headline implies to me:
Presidential approval
In the poll, Bush's job approval rating is at 39 percent among registered voters, a drop of three points since September, when his rating had increased to its highest level in months after he gave a series of speeches on national security leading into the fifth anniversary of the Sept. 11 terrorist attacks. The three-point drop, however, is within the poll's margin of error of plus-minus 3.5 percentage points.
Note again that they are polling registered voters, not likely voters. Also, this poll was conducted over the weekend - I've read many times that polls on weekends get more Dims, and the pollsters know that, as Pubbies are usually at church and/or out doing things w their families.
Here's another example: The President was in CA campaigning yesterday - even visited an elementary school named after him and will be going to AZ and CO. Wonder why that isn't on this calendar:
MSNBC Political Calendar
OCTOBER
1-3 John Edwards visits Uganda
3 Bob Riley’s birthday
John Kerry campaigns with Ohio Democrats
4 Chuck Hagel’s birthday
Ken Blackwell & Ted Strickland debate in Cincinnati, Ohio