Princeton Election Consortium

A first draft of electoral history. Since 2004

Meta-Margins for control: House D+1.0% Senate R+4.2% Find key elections near you!

Making your own Prediction

November 1st, 2004, 6:30pm by Sam Wang

I’ve been getting criticisms of my original estimates of the undecided and turnout adjustments. Why don’t we postpone further discussion until tomorrow, when we know the true outcome. Anyway, I have given you enough information to let you make your own predictions.

Lower and upper bounds can be set using Gallup’s last poll. This shows Kerry +2 among RVs and Bush +2 among LVs. The difference betweem these sets an upper bound of +4%. To set a lower bound: where data are available, new registrations give Democrats an approximately 0.5% edge in swing states; these new registrants are first-time voters who may fail LV screens. Also, the election will be high-turnout, suggesting that at least some of the 4% gap in Gallup will be filled. For example, a turnout increase from 50% to 62% would lead to +1.0%. Using this as a very rough estimate (and using it to subsume other factors as well), these two factors sum to +1.5%, a lower bound.

Finally: new registrants and overseas voters skew anti-Bush by a large margin, and 527s are pouring massive amounts of money into GOTV activities. I do not know how large these effects will be, except that they are likely to be towards Kerry. The suggested cell phone problem in polling is probably irrelevant (more on this later). I also do not know to what extent the 4 million evangelicals or Rove’s 72-hour operation will materialize for Bush. Known unknowns…

Tags: 2004 Election

No Comments so far ↓

Like gas stations in rural Texas after 10 pm, comments are closed.