The Morning After

In the supercharged atmosphere of a primary election, the urge to be first with the news can lead to hasty -- and flawed -- predictions. Such was the case with early television and radio reports about Pat Buchanan's showing in New Hampshire. Minutes after the polls closed, commentators announced that Buchanan was running close with the President, with each candidate getting more than 40% of the vote. As the night wore on, they lowered estimates of Buchanan's support to 40%. On Wednesday, when all the ballots had been counted, Buchanan's tally had dropped to 37%, 16 points behind Bush's.

Most of the early reports were based on an exit poll of 1,848 people by Voter Research and Surveys, the polling arm of the four major television networks. The survey overestimated Buchanan's support partly because the questioners did not expect so much of the Republican vote -- 10% -- to go to candidates other than Bush and Buchanan. In addition, Buchanan voters were far more eager than Bush supporters to talk to the survey takers, further skewing the initial estimates.

In most elections Bush's margin would be considered a comfortable victory. But in New Hampshire the early overestimates of Buchanan's strength created the mistaken impression that he had come close to scoring an upset, an impression that lingered after the results became known.