November 2004
Mon Tue Wed Thu Fri Sat Sun
« Oct   Dec »
1234567
891011121314
15161718192021
22232425262728
2930  

Day November 1, 2004

Election prediction: John Kerry

I wanted to post an election prediction with about a week to go, but internet connectivity on Saltspring Island turned out to be problematic. Ah well, I guess that’s why they call it a "vacation."

By now it must be obvious to everyone that the popular vote polls really are deadlocked, and whether an individual poll shows Bush or Kerry slightly ahead, it’s within the margin of error. Statistics buffs will know that this really does mean that you can’t falsify the null hypothesis that the two data series have identical means, so small variations in the percentages don’t mean a thing.

I’m a partisan in this election, so take this with a grain of salt, but I do believe that since the final debate, the evidence has shown that Kerry has been in a strong position to win this election.

Popular vote polls notwithstanding, my prediction is that Kerry is going to win the electoral college, and likely the popular vote by a small percentage. The electoral college will be 295 +/- 5 for Kerry, as GOTV efforts and the incumbent effect solidify his position in Pennsylvania, Ohio, and Florida. New Hampshire and New Mexico really could go either way, and turnout there will likely decide things. Today’s final Gallup poll and Zogby tracking polls confirm this for me; following Chris Bower’s reasoning, if the popular polls are in a dead heat, the challenger will likely win.

The above would suggest that the race should not going to drag on past Nov. 2nd — there may be some challenges in individual states, but depending upon which state is involved in a vote challenge, the electoral college count may already have decided the race. Only a significant challenge in Florida, Ohio, or Pennsylvania would likely stall our knowledge of the final outcome.

At any rate, I wanted to get a prediction posted. On Wednesday we’ll see whether it held up.