Comparing 2008 to 2004

By this point, I’m sure most of you don’t need much convincing that the contours of the current Presidential election are very different from 2004. But its worth looking at just how different. First, let’s take a very wide view. Below I’ve created a graph that plots the electoral votes that each candidate is projected to win based on Electoral-Vote.com, over the 30 days leading up to the election in both 2004 and this year.

There’s some interesting things here, but first we need to get some basics out of the way. The methodology of Electoral-Vote.com in 2004 was to award each state’s electoral votes based on the leader of the most current poll. That’s why you see such big swings back and forth between Kerry and Bush. There were a bunch of states that were very close and therefore polls kept showing the lead flip from one candidate to another. This year, however, Electoral-Vote.com decided to switch up its methods and now they use an average of the most recent polls (going back a week, I believe). The difference in methodologies explains why the 2008 lines are much more stable than the 2004 lines.

Stability issues aside, the graph reveals quite dramatically how different this race is right now. Obama has a lead that is greater than anything either Bush or Kerry ever achieved. Moreover, McCain is dwelling down in vote totals lower than anything suffered by Bush or Kerry. Indeed, neither Bush nor Kerry ever broke 300 electoral votes, and neither ever fell below 225. By contrast, Obama has not yet fallen below 350 electoral votes and McCain has yet to climb above 200.
So that’s the national picture. What about at the state level. Well, let’s look at my favorite state, Pennsylvania. Below is a graph of a the polling in Pennsylvania over the last 32 days leading up to election day this year and four years ago. Each day’s level is a simple average of all polls released that day and four days prior (a five-day rolling average, in other words).

This graph has a lot of the same patterns as the national picture. Obama is far outperforming Kerry and McCain is way behind Bush. The result is a big lead for Obama.

Finally, let’s look at one more state: Florida, that swingiest of swing states. Here’s the same graph for Florida:

Now this is really interesting. Again, Obama is outperforming Kerry and McCain is underperforming compared to Bush. But what I think is fascinating is how similar the Obama line is to the Bush line and how similar the McCain line is to the Kerry line. Of course, past performance is no guarantee of future behavior, but one wonders if the pattern will hold through the next few weeks. If so, we can expect polls to tighten some in Florida, but the averages will favor Obama.

One final note about Florida. Notice that in the final tally, Bush actually did far better than final polling suggested (by about 4 points). Since it is unlikely that all the undecideds broke his way (exit polls indicated that late deciders split about evenly), this disparity suggests the effects of a very successful get-out-the-vote effort. Indeed, we know that President Bush implemented an unprecedented GOTV operation in both Ohio and Florida (he similarly over-performed in Ohio by about 3 points). What does that mean for 2008? Well, if Obama does have a GOTV advantage, it could be worth up to four points.

Speculation aside, there is really no doubt that 2008 looks very different from 2004.