US Presidential Election 2000: Updated Thursday November 16 2000:

Our model of November 5 predicted a high probability of a Gore Electoral College victory. In our million simulations, the average Electoral College outcome was 282.4 electoral votes for Gore. As it turns out, the pivotal state clearly was Florida. If Gore were to win Florida, he would have 285 electoral votes.

Gore received all 10 states we initially assigned. Of the 17 battleground states that we assigned a probability >= .90, Gore received them. (There were no probabilities <.10, so none favored Bush that strongly). The smallest probability was for DE with p=.143, and it went to Gore. Second smallest was for NV at p=.146 which went to Bush. And so on for the moderate values which led to events consistent with the probabilities. (As for the smaller values, only DE went the other way, but it went to Gore instead of Bush.) NH had p=.73, so even though it went to Bush, it is not inconsistent with its probability. Florida had probability .89., the coin seems to have landed the other way----at least for now.

Pooling the poll predictions for the battleground states and weighting the states by the number of voters: 51.17% Gore (of two-party vote). Actual vote from the battleground states: 51.18% Gore (of the two-party vote) [using numbers in Wednesday Nov. 8 NY Times].

We expected the Electoral College to favor Gore, with a Bush win highly unlikely if the popular vote were a dead heat let alone if Gore had a lead. Why didn't this happen?

  1. Why Gore Was Not Strongly Favored in the Electoral College Under a Dead Heat Popular Vote, by Robert S. Erikson and Karl Sigman.

US Presidential Election 2000: Updated Sunday November 5:

GORE WIN LIKELY IN ELECTORAL COLLEGE

By considering current polls in the battleground states and their statistical error, we infer a likely win in the Electoral College for Gore
  1. Gore Favored in the Electoral College , by Robert S. Erikson and Karl Sigman.

US Presidential Election 2000: Updated Saturday November 4; results have sharpened

AS THE EVIDENCE FROM STATE POLLS ACCUMULATES, THE CASE FOR A GORE ADVANTAGE IN THE ELECTORAL COLLEGE UNDER A DEAD HEAT POPULAR VOTE SCENARIO HAS SHARPENED.THE MOST INTERESTING QUESTION IS WHAT IS THE CRITICAL VALUE OF THE POPULAR VOTE AT WHICH THE ODDS OF WINNING THE ELECTORAL COLLEGE IS THE SAME FOR BOTH CANDIDATES. THE ANSWER IS 51.4 BUSH/48.6 GORE (OF THE TWO-PARTY VOTE). IN OTHER WORDS THE BREAK EVEN POINT IS A BUSH ADVANTAGE OF ABOUT 2.8 POINTS IN THE POPULAR VOTE. IF THE POPULAR VOTE IS EVEN, THEN GORE WINS WITH VIRTUAL CERTAINTY. IF BUSH LEADS BY 5 POINTS IN THE POPULAR VOTE, THEN BUSH WINS AN ELECTORAL COLLEGE LANDSLIDE.
  1. A Dead Heat and the Electoral College (G.W. Bush versus Al Gore), by Robert S. Erikson and Karl Sigman.

  2. A Simple Stochastic Model For Close US Presidential Elections, by Robert S. Erikson and Karl Sigman. (The basics in this paper are applied in the above paper to the US Presidential Election 2000.)

All files are in PDF format, for a free PDF file reader go to: http://www.adobe.com/products/acrobat/readstep2.html




Karl Sigman's fun stuff

NEWEST PREDICTION!!

"After failing to garner enough votes to qualify the Green Party for federal matching funds, Ralph Nader teams up with Dr. Jack Kevorkian to create youscarethehelloutofme.com" --Will Durst (Political Humorist)

THE OFFICIAL FLORIDA PRESIDENTIAL BALLOT