I always wonder about the methodology of these polls: is it random, what is a sample size, who the hell watches lifetime TV, etc? In this case the phone poll was not quite random.

Republican pollster Kellyanne Conway of the polling company(TM), inc./WomanTrend and Democratic pollster Celinda Lake of Lake Research Partners conducted the scientific poll via telephone interviews July 25-29, 2008. The sample was 500 women nationwide, with 100 additional African American women and 100 additional Hispanic women. The margin of error for the main sample is +/- 4.4% at a 95% confidence level. Margins of error for the subgroups are higher.

http://www.mylifetime.com/community/my-l...

I wonder what percentage of the initial 500 women polled were white, Hispanic, African-American, Asian (they do exist and vote too), Middle-Eastern?

(I'm assuming tartu is saying that the polling was not random because of the additional groups of AA and Hispanic women who were polled. I assume that was done to keep the margin of error for those subgroups from being absurdly large.)

Polling is rarely if ever done purely at random, for very good reasons. (It's probably impossible to get a random sample of your target population, most of the time.) Generally AFAIK the big polls work from a (more-or-less) randomly chosen list of phone numbers, but they stratify their samples to try and make them more representative of the people (registered voters or whatever group) they want to poll, to make sure all relevant subgroups are being represented in the poll appropriately.

For example, I got a call yesterday from one who asked to speak to someone in the household who was a registered voter (that would be me), then asked the year I was born in, then asked if I were a man or a woman (hmmmmmm... good he didn't assume, I guess) and then finally, on finding I was a woman, asked to speak with the oldest male registered voter in the household. As it happens, there isn't one, so that was that. But clearly that poll had hit its quota of women in my age group and he was angling for another group they needed to fill.

There is a pretty good explanation of stratified random sampling and why it's used here. [Long and uses scary math notation in spots, but scroll down to "stratified random sampling" and its drawings.]

I'll now climb down off my lectern... ;)

San Juan could have put Hillary Clinton on his ticket and still would be bad on most women's issues. But, hey, it worked, at least nation-wide.

Here in Michigan, Palin seems to be losing steam. Though, I'd most likely chalk it up to the economy more than anything on Palin:

Palin's appeal slips for some women, recent poll shows

Just to note, EPIC-MRA is probably the best polling firm in Michigan. I only mention that because if anyone comes across their polls for Michigan in coming months, I want folks to know how accurate they usually are.

are the Democratic Party's most reliable pollsters. If these polls were about race, there wouldn't be questions about sample size.

Great post, lambert.

These polls appear quite conclusive.

active Republican spokesperson and dedicated pro-Republican cheerleader. Other than that, I'm sure she would never try to stir up trouble around tensions between the Democratic Party and white women.; Republicans are always unbiased, and every report from their polling professionals should be accepted without challenge.

Celinda Lake, on the other hand, actually is a Democrat, with some thoughts on the probable duration and size of this Palin ah Bump. She and Conway are a bit too close for my taste, but then I'm so Olde Skool.

Looks like they wrote a book together, What Women Really Want: How American Women Are Quietly Erasing Political, Racial, Class, and Religious Lines to Change the Way We Live and Lake Research Partners is one of the most reliable Democratic polling companies.

Celinda Lake is president of Lake Research Partners and is one of the Democratic Party's leading political strategists, serving as tactician and senior adviser to the national party committees, dozens of Democratic incumbents and challengers at all levels of the electoral process.

Celinda Lake, a leading political strategist for the Democratic party and one of the nation's foremost experts on electing women candidates, and Kellyanne Conway, a leading conservative pollster and president and CEO of The Polling Company, INC -- themselves cross the aisle to reveal the ways in which a newly defined, united power base among women is reshaping the state of our nation much more than the two-sided politics of Left and Right.

what, they're like a native american tribe now?

...anyway, polls like this both do and don't worry me. i think the more women see/hear of Palin, the more that support is going to evaporate. she's mean, and used her own kids as a prop. that stinks, to most decent women and the ones who like her were never going to vote for obama anyway. further, this election is going to be determined along the lines of the electoral college. random or not, general polls don't tell the important story.

it bothers me a little to recognize how, well, stupid a lot of people can be re: Palin's "feminism." but there are a lot of low information types out there, and propaganda and the SCLM exist because they work. i'm at the point, after years of blogging media 'pushback,' that i really don't know what is to be done about that. obama has said, iirc, that he's not interested in a new Fairness Doctrine or anything like that. it will bite him on the ass later, but you already know that.

The same number of Clinton supporters since she dropped out in June. More than half say it has nothing to do with the VP choices.

They find Obama less likable, honest, experienced and inspiring than Democrats overall do, and have a better view of McCain.