More important than national polls

If there’s one thing we should have learned from the 2000 election, it is that neither national polls nor popular vote numbers matter in predicting who wins the election. What matters, obviously, are the state-by-state results, which determine the electoral college votes. Unfortunately, the mainstream media have been nearly obsessing in the last few days over a couple of recent polls showing that McCain has gained some ground against Obama, but they’ve been ignoring the state polls that really matter.

There are several key things you will need to get an accurate picture of where the race stands at any given point. The first thing you need is access to the state polling data. One of the best sites for that data is Real Clear Politics, which tracks all of the major polls in each state. You can click on each state and see all of the polls in chronological order so you can see how the polls are trending. They also provide an average of the polls to show you were the consensus is in each state.

Results of recent polling in Michigan (Source: RealClearPolitics.com)

Results of recent polling in Michigan (Source: RealClearPolitics.com)


Secondly, you need to know how the electoral map turned out in 2004 (and perhaps even 2000) so you can see which states will need to flip from one party to another in order to change the overall result. CNN has a handy electoral map that can go back and forth between 2004 and 2000. What that map shows is that the Democrats won all of New England, all of the Pacific coast states and the upper Great Lakes states (Illinois, Michigan, Wisconsin and Minnesota); the Republicans won all of the south and middle of the country.

The last two elections have been very close, of course. In 2004, Bush won the electoral college by a vote of 286-252; in 2000, the margin was even closer, 271-266. In both elections, the swing of a single state could have changed the result. So the key question when looking at the electoral map is this: Which states do the candidates have a reasonable chance of flipping to their side in 2008? The only way to determine that is to look at the state polls and see how they’re doing in those states.

When you look at the state polls, it quickly becomes clear that the only blue states from 2004 that McCain has a reasonable shot at winning are Michigan, with 17 electoral votes, and New Hampshire, with only 4 electoral votes. Obama currently has the lead in both states, by an average of 1.4 percent in New Hampshire and by an average of 4.3 percent in Michigan. But no poll in Michigan has shown McCain with a lead since May, so the trend is clearly in Obama’s favor.

Ohio, as always, is another key swing state. There the polls show a statistical dead heat, with Obama leading the polls by an average of .5 percent. But the polls there are quite mixed, some showing a McCain lead and some showing an Obama lead, and most of them within the typical margin of error for such surveys. If Ohio and its 20 electoral votes go for Obama, it would become nearly impossible for McCain to put together a winning electoral map. It would mean he would absolutely have to flip Michigan to have any chance at all.

Obama, on the other hand, is polling strongly in many parts of the country that are traditionally Republican and has a reasonable chance of flipping several key states from red to blue. Colorado, for instance, is up for grabs with Obama consistently showing a small lead in the polls. In New Mexico, a swing state that went for Bush in 2004, Obama is leading the polls by an average of 6 points with no poll since April showing McCain leading the race. Likewise in Iowa, which went to Bush in 2004, Obama has shown a consistent lead since the beginning of the year, currently leading by an average of more than 7 points. Nevada is also in play, with Obama leading by an average of 1.7 points.

But perhaps most astonishing is the fact that Obama is polling well not just in these traditional swing states (states that frequently go back and forth between parties in elections) but also in states that have been strongly Republican over the years. For example, both Indiana and Virginia, traditionally very Republican states, are currently a toss-up in the polls, with Obama showing a slight lead on average in both (.5 percent and 1 percent respectively). Even in strongly Republican states like Missouri and North Carolina, McCain’s average lead in the polls is within the statistical margin of error. If McCain were to lose any of those states, it would be a huge upset and would make it all but impossible for him to win the election.

A word of caution, however: Let’s remember that it is still very early in the election process and polls can and do change quickly. At this point in 1988, a Dukakis presidency looked like a foregone conclusion. At this point in 2004, Gore was a virtual shoo-in. Much can happen in the next 3 months to change the public’s perceptions and thus their vote. But if we’re going to look at polls, we at least need to look at the polls that matter, the state polls rather than national polls. And those polls, as they currently stand, are very strongly favoring Obama.

Michigan and Ohio are shaping up as the key states in this election. If they stay as they were in 2004, Ohio going Republican and Michigan going Democratic, then Obama will have to win several of the smaller states to tip the scale in his favor. If McCain wins both states, Obama would have only a slight chance of winning the election; if Obama wins both states, McCain has virtually no chance at all of winning. Stay tuned to the Michigan Messenger for weekly updates on the polls in these and all the other key states in this election.