Mitt Romney slumps in new U.S. poll - The Globe and Mail

This news story provides the latest polling information on the upcoming US election:

Mitt Romney slumps in new U.S. poll - The Globe and Mail:

The truth of the matter is that Obama will not lose this election.

We have to be circumspect in our assessment of polls. As the recent Alberta election demonstrated, even massive and sophisticated polling can be inaccurate and, at times, remarkably so. This is something all pollsters know. It is why they include the margin of error in their own description of their numbers. The margin of error usually reads something like this:

This poll is accurate to within plus or minus three percent, nineteen times out of twenty. What this fictional example (which is actually pretty standard wording for real polls) means is that the poll itself is about three percent wrong ninety five percent of the time. So, for instance, a poll that said "50% of Americans have a favourable view of Obama" is actually saying something like "somewhere between 47 and 53% of Americans have a favourable view of Obama. The ninety five percent (nineteen times out of twenty) means that 95% of the time the result I just examples (somewhere between ...) is right. And, five percent of the time, the "somewhere between" is even wider. In other words, this poll begins by letting us know that there is a five percent chance that it is completely wrong. The higher the number of people polls, whether or not it is a rolling poll, the more sophisticated the polling methodology and analysis, will make a poll more accurate (a poll, for instance, of 1000 people is more accurate then a poll of 100 just by virtue of random chance). Other factors include who is polls (how diverse the polling sample, for instance, is important. A poll conducted only in, say, New York is less accurate of the American national picture then a poll conducted in diverse parts of the country, etc.).

The other problem with polls is that they lend the wrong impression to politics. They get us thinking about politics as a popularity contest and voting as something akin to calling into American Idol. This might be the way people think about potential candidates between elections, but it does not reflect how people actually approach voting. As we saw in Alberta, a lot of Albertans did not like the ruling PCs but when it came time to cast their votes, they sat down, thought about it, and said "gee, given the other options, the PCs are the best choice." This is another reason why mid-term polling is often different then election time polling. If an election is far in the future, those polled may just be offering their views at the time, as opposed to a serious consideration of how they might vote because, after all, a poll is not a vote.

With all this said, then, we need to be careful about polls -- and there have been at least a couple -- that show Obama running ahead of Romney. The best we can say is that these polls might or might not be an accurate reflection of how Americans will vote on election day. What is more, there is a lot of time between now and election day and the American presidential election system is complicated. Contrary to what most Canadians believe, American  do not have a direct vote for the president. George Bush Jr., for instance, actually had fewer votes the Al Gore in his first election victory. This does not happen often -- that the winning candidate has fewer votes then a losing candidate -- but it can and does happen owing to the way the American system works (which is through something called an "Electoral College" in which votes are weighted on the basis of state population and a person getting the most votes in a state, most of the time, gets all that particular state's support in the Electoral College regardless of their margin of victory or whether or not they had a majority of the votes cast).

My blunt statement: Obama will win, is not, therefore, based on recent polling. It is based on an analysis of American voting patterns and the particular body of people who came together to elect Obama. The rise of the Tea Party disguised the fact that the Tea Party was highly unrepresentative of the US. Even thought they like to picture themselves as average Americans, they represented minority religious views and were near universally white. I have not checked this out in terms of the data we have, but my guess is that Tea Partiers were overwhelmingly small-c conservatives to begin with and likely working or lower middle class. In other words, conservative white working and lower middle class evangelicals. Judging from their protests and platforms, I'd also suggest that some of them were simply not well informed about policy issues (the idea of rolling pay dates, for example, for "public servants" if the government ran out of cash was just silly, but I blogged on that some time ago. It missed the point that the military are public servants and hence constitute a big chuck of the government budget line and it missed this question: how long would you work at your job if someone told you you'd only be paid every second or third day day?)

The Tea Party seemed to catch the Democrats by surprise. It likely shouldn't have but the Tea Party did not "play by the rules." You can call your opponent names, but you don't call him Hitler. You can oppose what your opponent is doing but you don't derail the state to meet your political objectives. Instead, the Tea Party set a whole bunch of new rules and mobilized in a way that surprised the Democrats. This combined with some pretty ineffective government from the Obama administration led to a mid-term collapse in the House of Representatives and the Senate.

Don't bet on that happening again. First, the Democrats now know what is coming. "You're going to use thinly veiled anti-Hispanic racism to mobilized an economically fragile white base. OK, we'll take that and appeal to the very Hispanic voters -- millions of Americans in key states -- and appeal to them." The Democrats have now had two years to develop a response to Tea Party rhetoric. My bet is that they have used that time to good advantage and can counter the negative Tea Partyesque campaigning that will come their way. Second, the Democrats misjudged the importance of the mid-term elections. They let the   pro-Obama coalition they mobilized lapse. Again, they won't make that mistake this time. Moreover, the base that elected Obama views the presidential election as more important. Finally, Obama and the Democrats never had the white conservative evangelical vote to begin with. Indeed, Obama actually did worse among that constituency then his predecessors had done. In a large number of ridings (what American call districts) with significant evangelical populations, Democratic vote actually went down -- despite Obama's national win -- in the last presidential election.

What is more, the Democratic base (moderate Christians, ethnic minorities, youth, women) view conservative evangelicals with something approximating fear. They have no desire to see a return to what they viewed as the horrors of the Bush administration and will not be tricked by a supposedly moderate Republican who must, of necessity, court a more extreme voting block (in the manner of Bush with his "compassionate conservatism disguising the extremist views of his supporters). In other words, the Tea Party transformed politics from politics into a war for the "soul" of the nation; the Democrats will be happy with this rhetoric and say "exactly" so ... who would you prefer to have running things? America is a fractured country; deeply divided on ethnic, religious, gendered, and class lines. The majority of Americans who supported Obama will reach the same decision they reached last time 'round. "We'd rather have someone like us running things then someone like ... say, a cross-dressed Sarah Palin."