Dole Leads Bowles in Poll; Race Stays Close

According to the poll, 47 percent of those questioned support Dole while 40 percent plan to vote for Bowles. Thirteen percent were undecided.

Libertarian Sean Haugh, the only other candidate to appear on the ballot, was not included in the poll because he did not participate in either of two televised debates and did not campaign actively across the state.

In a press release, UNC journalism Professor Robert Stevenson, the poll's director, said if undecided voters select between the two candidates in the same proportions as decided voters, the likely Dole margin is 8 percentage points.

The poll results come from interviews with a representative sample of 560 registered voters statewide in the week ending Thursday.

According to the poll, support for Bowles is especially strong among young voters and blacks, while Dole is popular with men. It also showed candidate preference is spread evenly across the state. A random sample of this size has a maximum error of about 5 percent.

The poll, in its 20th year, is carried out by journalism students under faculty supervision.

Though poll results are generally accurate, it must be taken into consideration that some people change their minds before the election, said UNC political science Professor Thad Beyle.

He said issues that were bothering a person can manifest themselves in the last days of a campaign. "When there's movement going on, you can be somewhat surprised," he said. "People are making up their minds at the last minute."

Dole will continue her grass-roots campaign in the last days of her campaign, said Dole spokeswoman Janet Bradbury. She said Dole will not increase her actions to maintain her lead.

"I think the poll that matters is Tuesday when the voters actually vote," Bradbury said. "North Carolina Senate races tend to be tight. We did expect this one to become closer, and it has."

Bradbury said she has not seen the poll results.

Bowles also will not change his campaign techniques because of one poll's results, said his spokesman, Brad Woodhouse. He said Bowles' final weekend of campaigning found him in the eastern part of the state reaching out to voters.

Woodhouse also said he does not think the Carolina Poll was conducted in a reliable manner. He criticized the fact that it was conducted over five nights and only questioned 500 registered voters. He said that at this point in the campaign, a larger sample of likely voters, taken over two to three days, is more accurate.

"The race changes over a series of days," Woodhouse said. "What happens on night one can be far different from what's happening on night five."

But Beyle said if nothing else, polls tend to influence candidates and their staffs by alerting them of the need to get out the vote. "In many cases, what these polls do is energize people on the inside."

The State & National Editor can be reached at stntdesk@unc.edu.