While I'm writing about polling today--

Far be it from me to discourage Democratic overconfidence, but Immanuel Wallerstein is way over-the-top in predicting a superawesome Obama landslide this fall. He's made some pretty basic methodological errors. (This wouldn't normally attract my attention except that James Wolcott, one of my blogging heroes, seems to have swallowed it.)

Wallerstein writes:

Let no one underestimate it. Barack Obama has won big. He has not only won the Democratic nomination for president. He is going to sweep the elections with a large majority of the Electoral College...

I have just done an analysis comparing McCain's state by state strength in the latest polls and Bush's proportion of the actual votes in 2004. In 45 of the 50 states, McCain is weaker, often much weaker, than Bush was. And in the other five, he is about the same. Of course, if Bush had won a state by a large margin, McCain will still win it albeit by a smaller one. But in the states that were close in 2004, the tide is in Obama's favor.

Ok... uh... see... "analysis..." the, uh... DUDE THERE ARE UNDECIDEDS IN THESE POLLS. Of course McCain is underperforming 2004 vote counts; it's freaking June. Like 15-20% of the electorate is still undecided. I'm sure that Obama is drastically underperforming from Kerry 04 right now; does that mean his campaign is doomed too?

In fact the current polls show that Obama could easily lose if he slips in the Upper Midwest (if Michigan goes to McCain, which for some inexplicable reason is a possibility, it's game over). Visit DemConWatch and play with their excellent General Election Tracker to give yourself some idea. Wallerstein's whole thing here is a complete fiction.

Oh, and that's the one and only piece of "evidence" Wallerstein has. The rest of his argument reads like this:

Furthermore, we have to realize that McCain is currently at the top of his strength. The Democratic Party is now reunifying and hungry for winning. Obama will lose almost none of the traditional Democratic percentages among women and Jews. He will increase the national percentage among Latinos and will bring in a very large number of young people and African-Americans who otherwise would not have voted. He will also get the votes of the considerable number of independents and Republicans disillusioned with Bush. The people who will vote against Obama because he is African-American were almost all already going to vote Republican. This issue is behind him, not in front of him. (...etc)

Yeah, well, I can confidently state in my commentary that Barack Obama craps gold bricks, but that doesn't make it true. Until I see evidence for these contentions -- some of which are quite controversial (there aren't going to be racist defections? really?) -- this whole thing is meaningless.

I'm annoyed here because Immanuel Wallerstein is a towering intellect, and this is beneath him. There are way too many examples of smart people writing confidently and at length based on completely false assumptions from bad polling analysis. And consequently, often conventional wisdom is formed without a shred of real evidence behind it. This stuff is not hard -- make an effort, guys...