and something that we in the netroots could follow up on on election day.

In 1992 when I was one of the target demographics for the "Rock the Vote" election, youth turnout was really visible. The surge was obvious even before the exit polls came in.

20somethings were lined up and poll workers were pretty dumbfounded.  It seemed to have happened overnight.

We saw that in Mpls.  We heard about from other centers with high rates of single, 20something residents.

We should cover this story in the lead up to the election.  My bet is that Mike's predictions will bear out.

Great question to which I don't currently have the answer.  I've  only had time to look through the Press Release and skim the Exec Summary.  

I've sat in on some discussion about proper poll construction, and correcting for factors like that, but I'm by no means an expert and probably wouldn't even consider myself an informed amateur.  I'd guess that there might be a way to correct for that bias by looking at historical trends from previous midterm elections, but I don't know if Harvard did that.

I'll look through the full poll/results later tonight and report back if they did in fact correct for any biases.  FWIW though, I've heard that the Harvard IOP Polls are usually pretty accurate.  

And, as someone who worked very closely with young voters in 2004, I share your view that there are cultural shifts happening and I think the trend is real.

Does that sort of follow through happen with other groups of voters?

There is a correlation that is important in these polling numbers, and one that backs up your gut and Mike's post.

The #s in 2004 that were key indicators of the higher turnout in that election were the numbers on how much young people were paying attention, how much young people felt like the election mattered, and how much they were talking about it.

In 2004, all of those numbers were in the 70-80% range.  This poll did not ask the same exact questions, but if you look at the "Definitely going to vote" and "Following the election in the news" #s (54% and 60%), they are lower than 2004, but still higher than anyone would expect for a midterm and almost surely going to lead to a higher turnout than usual and possibly than ever in a midterm.  Over the course of time mid-term turnout is usually 22%.  I think it was 17% in 02.  

It really could go as high as 30 in this election based on the numbers here and in the Pew poll:

From Young Voter PAC:

According to the polling of the Pew Research Center, the number of Americans 18-25 giving a lot of thought to the mid-term elections has almost doubled from 2002 to 2006 (27% to 48%).  Andy Kohut, the chief pollster for the Pew Research Center, believes that "we may again see higher turnout among this generation in this congressional election as we did in the last presidential election."

So, yes.  Turnout is going to go up.  Even better is that Democrats are pounding Republicans 2-1 among those mostly likely to turnout.

Do you have any reason to think young people are less truthful than the average poll respondent? I don't. We'll know in a week though. ;)

For some reason liberals always think the youth vote is going to be liberal.  It's a fallacious assumption that has never been true.  18 - 21 year olds tend to vote exactly as their parents do but don't show up at the polls very well.

The only thing special about the youth vote is their apathy.