survey USA is really stinking it up lately.

just look at the internals of the poll.. Dems and republicans are tied... ummmm... no way. Dems have a 7-10% advantage in Minnesota, if not more.

they are supposed to make us believe the young support Coleman almost 2 to 1? LMAO.

bad poll. Survey USA i have no respect for lately. their polls are BAD.

Yet another chance for all of us to learn how the margin of error works.

Now, who can tell me why both of the recent Minnesota Senate polls are consistent and don't contradict?

I'm very interested in these internals. As an 18-34 year old in Minnesota, I find it extremely questionable that we break for Coleman 58-31.

Maybe they worked their poll on a listingf rom FOX Radio?

This is the fourth poll in a row that has Coleman up double-digits in Minnesota. And they are the ones who are in line with other polls of the state. The Quinnipiac poll had Coleman up 10% just last week.

It's the middle of July. The best we can expect from these serial polls is to see a trend one way or the other over an extended period. As is often said, any one of these polls is no more than a 'snapshot in time'.

Eh, they're not polling Independents either, which is a major problem for pollsters in MN, since we have a large number of independents of varied parties and no party. No polls done in MN will be truly accurate unless they take the independents into account.

Regardless of the polls, how is Franken doing on the stump? A friend of mine in St. Paul, very liberal, saw him speak a couple months ago and said he was terrible--very wooden, uninspiring, etc. This certainly surprised him, and me, given Franken's years of being in the spotlight. How's he doing now?

SurveyUSA, MasonDixon and Quinnipiac tend to overestimate Republicans in most polls. Rasmussen is better, but also tends to have above average numbers for Republicans. CBS and Gallup tend to be at the opposite end. Others are all over the place.

It is also important that sampling is different for each organization between the national and regional/state polls. It's not entirely clear what methodology would make polls more accurate because there are so many variables. And the weighting on top of that could also skew results if it relies on old demographics figures.

My guess is that SUSA underestimates the potential Democratic vote based on past elections. However, doing so consistently is also important, as stable numbers even in badly skewed polls show that the race is not changing overall. Trouble may arise if demographics somehow change between polls, e.g., one party constituents have a surge in voter registration or suddenly become more likely to show up to vote. But other comments are right on one point--it makes no sense that SUSA internals show younger voters going for Coleman. Given a number of obvious flaws in SUSA splits and weights, the actual numbers are badly off and Franken is likely closer than SUSA shows. But he has not made a significant move over time and he clearly started behind. If his numbers don't start moving by convention time, Franken will be a hard sell the rest of the way. Coleman's best strategy right now is to stay out of the news.