With a little over a year to the 2008 election, there have already been over 100 statewide polls that pit Hillary Clinton against Republican challengers in a general election match-up. These polls can be combined to give us an early glimpse of the national mood (albeit one likely soured by years of Bush administration’s military misadventurism).
I’ve been collecting these polls for awhile now, and finally got around to an analysis of the Clinton—Giuliani head-to-head state poll outcomes. My analyses differ from the typical national head-to-head polls because (1) I am concerned with estimating the electoral college results from the statewide polls, rather than estimating a national popular vote, (2) I make extensive use of Monte Carlo simulation methods to examine the probability that each candidate wins, and (3) by combining multiple polls the sample sizes are much larger.
For example, here is the result of 100,000 simulated elections using all the 2007 state poll data I’ve been able to find:
The red line is drawn at 269 electoral college votes—a tie. The area to the right of the red line are Clinton wins, those to the left are Giuliani wins. (There are a few ties. A tied election would almost certainly result in Clinton being elected since the House Democrats currently control 26 state delegations, and the Republicans control 20 state delegations. The 2008 election will probably increase the Democrat’s control of the House.)
These results suggest that Hillary Clinton would have a 73% probability of being elected President, but only if the 2007 polls in toto reflect the national mood for election day, 2008.
In reality, attitudes change with time, so it is helpful to restrict the analysis to the most recent polls whenever possible.
Here is what happens if we restrict the analysis to polls taken in October, 2007 (unless there are no polls in a state for this month, in which case we take the most recent poll, or, if there are no polls at all, we give the electoral college votes to the party that took the state in 2004):
After 10,000 simulated elections, Hillary wins 9,991 times and Giuliani wins 9 times. Clearly, over the course of the year, Hillary has become more acceptable, Rudy has become less acceptable, or some mixture of both has occurred. (In fact, further analyses reveal that Rudy has made a bit of a comeback from his worst showing in mid-summer.)
The polls to date put Clinton in an extremely strong position to win the 2008 election. Giuliani has a tough row to hoe just to become competitive. He’ll need a lot more than just fear-mongering over 11 Sep 2001 to pull off a win.
(I provide a more complete description of the methods and results here.)