- The Washington Times - Tuesday, October 23, 2001

The Virginia governor's race is awash in polls: public polls, private polls even some unpublished polls are rumored to exist. But the results of the past few governor's races have shown that Virginia voters often surprise the pollsters.
In 1989, 1993 and 1997, the Republican candidate's strength was underestimated in almost every late October poll. That could be a good sign for Republican Mark L. Earley, who has trailed Democrat Mark R. Warner in every poll published to date this year. But the polls also hold good news for Mr. Warner the person the polls showed leading in late October ended up winning each time.
Polls this year have shown Mr. Warner topping out with about 45 percent of the vote, and Mr. Earley anywhere from three to nine points behind. Libertarian candidate William Redpath polls in the low single digits.
Specifically, a poll by Mason-Dixon Polling & Research Inc. last week showed Mr. Warner with a three-point lead over Mr. Earley.
Mr. Warner's campaign says it's not that close, pointing to a week-old private campaign poll that they said showed the Democrat up nine points. There are also reports which Republicans deny of an internal Republican poll showing Mr. Earley trailing by double-digits. The most recently released Republican campaign poll, taken in early October, showed Mr. Earley trailing by four points.
Campaigns use polls as broad indicators of what issues are working, and in which regions of the state. For voters, though, a poll's usefulness is undefined.
"We all suffer with this obsession about public opinion polls," said Larry Sabato, a longtime observer of Virginia politics who catalogs polls in his "Virginia Votes" books. "It creates momentum, and stops momentum for different candidates, and that shouldn't be what elections are about. But it is."
In the last three elections, the polls have consistently underpredicted Republican gubernatorial candidates' performances.
In 1997, major polls showed Republican James S. Gilmore III winning by between seven and 12 points, but he ended up winning by 14 points. In 1993, George F. Allen won with by a 17-point margin, though major polls showed him winning by anywhere from seven points to 13 points. And it's not just the front-runner whose strength is underestimated. In 1989, the late polls showed Republican J. Marshall Coleman trailing Democrat L. Douglas Wilder by anywhere from 4 to 11 points. Mr. Coleman ended up losing by less than one point.
Each of those has explanations, though, according to Mr. Sabato. Undecided voters don't break evenly more will naturally end up going for the leading candidate, which means his or her margin of victory will expand in the final result as compared with late polls. Also, in the Wilder race, he said, the polls were probably skewed by political correctness some voters told pollsters they preferred Mr. Wilder, who is black, because it was the politically correct answer, but ended up voting for Mr. Coleman.
Republicans, though, say there's something to the theory that Republicans outperform polls. Some say it's because Republican voters are less likely to answer polls. There's also the issue of likely voters it's difficult to find a poll sample that matches the eventual electorate, and in Virginia that usually breaks for Republicans at election time, says Stephen K. Medvic, a professor at Old Dominion University.
That may hold true this year as well, he said: "How much of the results are counted for by over-representation of Democrats, it's impossible to say ahead of time. But you can say it's still probably an over-representation of Democrats in these polls."
Still, he cautioned not to take too much from broad outlines of history.
"You really do have to take it poll by poll, year by year and place by place," Mr. Medvic said.

LOAD COMMENTS ()

 

Click to Read More

Click to Hide