Wishful thinking knows no party: The Daily Kos fantasizes a non-existent African-American voter turnout in NC

From the kos:

Research 2000 for Daily Kos. July 28-30. Likely voters. MoE 4% (4/28-30 results)

Dole (R) 50 (48)
Hagan (D) 42 (41)

Not much change, with Dole right at that 50 percent mark considered "safe", though I doubt anyone really thinks Dole can just slack off.

There are several races in which Obama's strength will have a say in the outcome of the Senate race, and this is certainly one of them, so let's see how he does:

McCain (R) 47 (50)
Obama (D) 43 (41)

Yeah, this one is a tossup. And it's likely even better than this. According to the 2004 exit polls African Americans made up 26 percent of the North Carolina electorate. The sample size of this poll is 22 percent black, and does anyone think black turnout will be down this year? Didn't think so...

So all things considered, these are pretty darn good numbers. You Tarheelers are going to be really busy this year.

Never let a little stand in the way of creating gigantic White Houses in the air, as Tom Jensen of Public Policy Polling points out:

As has been the case every time a North Carolina poll is released over the last few months, people are again criticizing the 22% representation of African Americans in today's new Research 2000/Daily Kos poll.

They're doing this because the 2004 exit poll said that 26% of the North Carolina electorate was black. The reality, as I've written before, is that it was actually less than 19% black.

Here are the exact figures, courtesy of the State Board of Elections:

3,552,449 people voted in North Carolina in November 2004.

659,656 of them were black.

Do the math and that comes to an 18.6% African American electorate. That is not my opinion, that is a fact. It is somewhat conceivable the black electorate could go up to 22-23% this year but that is most likely the ceiling unless a whole heck of a lot of white people just stay at home this year. I have no doubt blacks will vote in record numbers but I wouldn't be surprised if everybody else does too, so it would take a lot for the proportion of voters who are black to make a huge leap. A 2-3% increase would still be a big deal but hoping for a 6-7% increase in the proportion is dreaming.

The 2004 North Carolina exit poll was also well off in terms of its gender (too female) and age (too young) distributions. The South Carolina exit poll from 2004 is full of mistakes as well.

All of which goes to prove that at this point the primary use of polls is to convince voters rather than to report what's happening.

But it's when you start believing your own press releases that you really get into trouble.