Pollsters for U.S. Rep. Jim Marshall and challenger Calder Clay agree that Marshall is in the lead. The only question is how big a lead.
Marshall's campaign released a poll Wednesday showing his support at 51 percent among likely voters in the 3rd District, compared to 27 percent for Clay.
In response, Clay's campaign said its own polling shows Marshall hovering at 50 percent and Clay at 37 percent.
And they could both be right, statistically speaking.
Each poll had a margin of error of about 5 percent.
Added together, said political science professor Chris Grant at Georgia College & State University, that yields a potential 10-point spread - exactly the difference between the two polls' assessment of Clay's support.
"There's no difference. They're in the margin," Grant said.
As for what the numbers mean, the two campaigns offered opposing spin.
A memo from Marshall's pollster, The Mellman Group, said Marshall is "in a very strong position to retain his seat in the 2004 General Election."
An answering memo released by Clay's pollster, Public Opinion Strategies, said, "Jim Marshall has not broadened his base at all since he won this close election in 2002."
Grant, meanwhile, said that as long as an incumbent's numbers stay above 50 percent, he's in good shape.
The polling data, he said, "shows Marshall not to be vulnerable, but to be potentially vulnerable at this stage. It also shows Clay has a long way to go in making up the difference, and needs to be able to do something quickly in terms of getting his name out."
Marshall's pollster works for Democrats and Clay's for Republicans. Each firm said it polled 400 likely voters by phone in mid-August.
The Clay camp pointed out that the numbers were gathered before the GOP convention, and do not reflect the bounce in Republican strength that followed it. The Marshall camp said the numbers don't reflect the boost to Marshall from his television ads, which began airing recently.
Marshall beat Clay by a 1 percent margin two years ago. And during the 2002 campaign, both candidates released polls that failed to predict the race's outcome.
In mid-October 2002, Marshall's poll said he had a a 17-point lead over Clay.
Clay's pollster said that Marshall hadn't broken 45 percent, and Clay was 7 points behind.
When voters went to the polls on election day, they gave Marshall 50.5 percent and Clay 49.5.
Polling, Grant said, "is an imperfect science, but it's the only process we've got."