(Washington, D.C., Right Commentary): In the last 48 hours, something has emerged that could very well be the early indication of McCain defeating Obama on November 4. For the first time since McCain’s address to the convention, the overall spread between Sen. Obama and Sen. McCain has dropped to an average of 3 points - or essentially the margin of error. Thus, McCain supporters need not be so glum that all is foretold and that the race is lost. While some pundits like Hannity and Limbaugh have been railing against the “Main Stream Media,” asserting that traditional news sources are trying to convince people that the race is locked, when it is not, for the first time, there is actually some evidence of that fact in polling. A strategy to victory, although improbable, for McCain exists. Senator Obama may indeed be vulnerable. Let me explain why.

First, the polling on Presidential races has been, to say the least, difficult to interpret. Although the infamous line of “Dewey defeats Truman,” has been applied to this year’s election, I’m not sure that analogy is truly warranted. In the case of the Dewey-Truman polling, especially in the case of Gallup that was heavily relied on in order to predict the race, the polling was imperfect as a matter of methodology. Gallup primarily relied on telephone polling to determine the outcome of the race - a methodology that resulted in a significantly skewed sample in favor of Dewey. Other polling, such as the now famous “popcorn poll” and “sack polls” that had been done - showed a groundswell of support for Truman. By all standards, the “Dewey Defeats Truman” upset was no upset at all - but a prediction based on incredibly faulty statistical analysis.

While I do believe some of the polling thus far that has been done is “push polling” - where the polling naturally leads people to a conclusion - the majority of the polling done, and the majority of the polling being presently conducted, is at the “state of the art,” compared to what is considered proper methodology. Although people can debate likely voters versus register voters, etc., the reality is, the polls are essentially being conducted with representative samples of the people who are going to be voting next Tuesday. That said, there is something going on the polls that hasn’t been accounted for - namely - the variance in readings.

The polls have been completely erratic. Two polls that were conducted during roughly the same time frame reached two different conclusions that were statistically significant - one found that Obama/McCain were essentially tied, another find Obama ahead by 8 points. How is that possible? For example, last week, ABC finds Obama up 8, but IBD, and GWU polling finds him up by 3 - essentially the margin of error. Moreover, the most ridiculous outcome is that of Pew Research, which finds for the same period, Obama is up by 15! That’s almost double the ABC finding!

In statistics - we say - “something is clearly up” when we see that level of median variance. I won’t go into the long detailed analysis - most of you don’t care. The bottom line is, however, that while one would expect some variance because of questions, sample size, likely v. registered voters, etc., one would not expect to see such a wide variance between numbers. The low and the high means are separated by a factor of five - that’s just plain unacceptable in conducting statistical analysis. Someone is clearly wrong.

I don’t think we’ve got one organizations doing bad polling - if this is messed up this round - the entire polling industry is going to be put under a microscope. The statistical analytical failures will be astounding. Someone is wrong - and wrong big. Either that - or the volatility in the public is so high that polling is essentially useless. I find the pollsters being off more likely to explain the variance.

But putting polls aside for the moment - the reality is that I think Obama has been up by about 2-3 points ahead of McCain from about September on. That doesn’t mean a heck of a lot in terms of votes - its about 2 million votes. However, where those two million votes have been located has given Obama a significant electoral lead. Until today, I had Obama as high as 370 delegates. I believed it was more than possible for Obama to win Florida, North Carolina, Ohio, and Virginia, in addition to his staunch base in the mid and north Atlantic and the west coasts. If he in fact did pick off Iowa, Colorado, and New Mexico, Obama would essentially have a map similar to that of Clinton v. Dole, trading Louisiana and Arkansas for Colorado.

Now as I look a the map, and the polling, Obama may still be 2-3 points ahead in the overall polling - but it is no long assured he can win any of those states, and the polling in those states is so all over the map, one has to really wonder about their predictive value. For example, Florida, from October to now, the polling has widely fluctuated, showing Obama ahead by 8 and then even, and McCain ahead by 3 and then even. Now, Obama has a 3 point lead. Given the 30 million dollars he’s pushed into Florida (or more), he should have a solid lead over McCain. He doesn’t - why? Similarly, Ohio, if you compare the 2004 polling data to the the 2008 polling data - it looks remarkably similar. Bush won Ohio in 2004 - it’s possible for McCain to win Ohio in 2008.

In order for McCain to win, he has a difficult road ahead of him, but Obama may be in fact vulnerable. McCain must win the following states, or it’s over on election night: Indiana, Ohio, Virginia, North Carolina, and Florida. If he were to win all of those states, I think it is entirely possible he would also then win Montana, North Dakota, and Missouri. The final hurdle for McCain is Colorado. The volatility of Colorado polling in the last two weeks seems to make it vulnerable for Obama. 

Now, my friends (to quote McCain), this isn’t necessarily as impossible as it all sounds. Let me deal with the obvious argument - namely - every African American of voting age in the US will vote for Obama, and that advantage will pummel McCain. First, while I expect that segment will vote for Obama in an overwhelming way, the question is not so much that one segment of the population. I also expect a fair number of new voters to vote - and they will overwhelming vote for Obama (by a margin of about 2 to 1 if one believes the polling on that issue). This is primarily because Obama has spent literally hundreds of millions of dollars on voter registration drives. 

Those populations, however, are going to have to offset surges in other areas - as all demographics appear to be surging in terms of potential turnout this election. Democratic gains among African American voters and younger voters will be offset by higher turnout among more Republican-leaning voters. In addition, there is a palpable sense that this race is tightening in battleground states, with independent voters increasingly receptive to McCain.

It is possible for McCain to win. There are significant hurdles and the odds are against him. I personally would not have followed a strategy that lulled my opponent into a “false sense of electability.” However, Obama’s arrogance in measuring the drapes beforehand, issues like Joe the Plumber, and the statements about redistributing the wealth, seem to have shaken loose some of the people in the middle in favor of McCain. Obama should be over 50% of the vote and solidly in the lead - why isn’t he? Why is the polling so erratic? Can it really be relied on? And what about the 20M votes that have likely already been cast by absentee and early voting? No one knows the impact of those factors.

If I were making Vegas odds - I’d give McCain 30 to 1. It’s not a true long shot, but a good deal of things have to fall into place to make McCain’s victory possible. But I think for the first time in a long time, a McCain victory truly is possible. He’s got five days… and in politics… 24 hours is an eternity.

Possibly Related Posts:

• Being Rod Blagojevich
• Eve of Destruction: Are we headed for a Global Recession?
• Treasury Bills “crack the floor” and deflation becomes a greater risk.
• The next “bubble” to burst - US Government debt?
• Franken: I’m stupid enough… I’m vain enough… I’m litigious enough… and doggoneit… I’m still behind.