ANALYSIS: Our analysis revealed that if there had been a dead-even national result in 1996, Democrat Bill Clinton still would have likely won. Under the Electoral College system, the presidential race is decided by state-by-state results, not the national vote, and after 1996, the Democratic vote was dispersed in such a way to give them a slight advantage in the Electoral College map. In 2000, however, it appears possible the Democrat Al Gore will lose the presidency despite narrowly winning the popular vote -- a shift from 1996 due primarily to the growth of the Democratic vote in California and New York. At this point, neither party gains an edge because of the Electoral College. The fact that Republicans gain a boost because of winning more of the small states is balanced by the Democrats' edge in most of the large states In 2004, then, a dead-even vote in the national vote would once again likely be reflected by a very close finish in the Electoral vote, dependent on who wins a handful of close races. METHODOLOGY: The numbers under "partisan indication" for 2000 were based on what would have happened in an even national vote in 1996. The number was generated by a simple calculation. Bill Clinton won 49% of the popular vote in 1996 (compared to Bob Dole's 40%). If Clinton won 51% in a given state, then, his performance in that state was 2% above his national average. If the national race had been dead-even, we assume that this state would have given Clinton 52% of the two-party vote (two percent above a 50%-50% national two-party divide). Note that this method of predicting elections has proven quite reliable in U.S. House elections, particularly in open seat races. [See Monopoly Politics] For 2004, we have chosen to use the relative two-party vote as our basis of defining partisan indications rather than the actual vote. Doing so for 2000 would have slightly improved our accuracy and is a way to equalize the impact of third party candidates, who often have different rates of success in different states. Given that the 2000 presidential race in fact was dead-even, the partisan indication for 2004 is a simple calculation of the two-party vote from the 2000 election |