Posted on 11/02/2002 7:32:32 AM PST by BlackRazor
Chocola up 15 points in latest poll
Long Thompson's staff disputes 'flawed' survey
Sat, Nov 2, 2002
By Trevor Wendzonka
Truth Staff
Even before President Bush's visit Thursday, a new poll shows Republican Chris Chocola had opened a 15-point lead in Indiana's Second Congressional District.
The survey of voters was conducted by Brian Vargus of Indiana University's public opinion laboratory -- the same pollster who showed Chocola with a 12-point edge over Democrat Jill Long Thompson in the middle of October. The first poll was commissioned by The Truth and WNDU NewsCenter 16, and Vargus explained he did the follow-up survey for himself.
Long Thompson's staff disputes the findings, saying Vargus continues to use suspect methods in his polls.
"This poll continues to show movement by Chocola, substantial at this time," Vargus said. "I'm even doubly convinced now that the advantage is with Chocola, though the real advantage will go to whoever moves the votes on Tuesday."
The sample of 421 voters showed the Republican candidate with 49 percent support, while Long Thompson was at 34 percent. Libertarian Sharon Metheny was holding her 4 percent total from the previous poll, while undecided voters accounted for 13 percent. Sampling error remains about 5 percent.
Chocola continues to take more independent votes, according to Vargus, and is drawing more support from Democrats than Long Thompson is taking from Republicans. And, Vargus reported, Chocola is getting 25 percent support from people who said they voted two years ago for retiring U.S. Rep. Tim Roemer, D-South Bend.
"We've been pretty happy with all the polls because they're showing people are getting our message," said Laura Zuckerman, spokeswoman for Chocola's campaign. "Obviously, we're doing the right things ... but the key to victory will be turning out our supporters on Election Day."
Almost 40 percent of those polled said they were Republicans, while 29 percent sided with Democrats. That gap is just one of the items disputed by the Long Thompson group.
"I think it's incredibly flawed," said Paige Richardson, the Democrat's campaign manager. "The methodology is questionable because the response rate is irresponsibly low, and the party ID is flat-out wrong. This is a 50-50 district, and if you poll 10 percent more Republicans, of course you're going to get this result.
"But this is the same pollster that said Tim Roemer was 12 points down to John Hiler in the last days of 1990, and Roemer won by two," she said.
Vargus did say Friday Roemer's defeat of Hiler was one of the three races he's totally missed since beginning to poll in the mid-1970s.
"That's why you always have to say these polls aren't predictions, yet we're all guilty of doing it at times," he said. "But the fact is, the odds of being this wrong twice are not very likely. I am confident there is a shift to Chocola, and I imagine it's showing up in Long Thompson's numbers, too."
The response rate for both polls was about 10 to 11 percent, meaning thousands of numbers were dialed to get just a few hundred voters each time. Vargus uses a process of random dialing, which he said is a must because voter files are available only to partisan pollsters in the 60 days leading up to the election.
Anna Greenberg, a Democrat pollster based in Washington, said her firm uses random dialing often, but not in midterm congressional races when voter turnout is low. She said an accurate survey must reach only qualified voters.
"I'm not going to dismiss the methodology, but 11 percent is quite low for a response rate. You don't necessarily outright reject the poll because of it, though," Greenberg said, saying the Washington Post demands at least a 25 to 30 percent response, and many academic institutions require 50 percent. "(The process) introduces all kinds of bias into the survey. Eighty-five percent of the people will say they vote when you ask them, but you know only 50 percent do. By not using voter files, you're reaching people less engaged with the race.
"These polls (by Vargus) are also inconsistent with every other poll that's been done in the district, both public and of the ones I know privately. It's not plausible she's down 15."
Greenberg said she belongs to the same group as Vargus, the American Association of Public Opinion Research. A representative of AAPOR did not respond to an interview request Friday.
Both the Chocola and Long Thompson representatives said they hope the gap in the latest poll doesn't keep voters home on Tuesday. Vargus said no evidence exists showing polls have a direct effect on turnout, but added, "There is some sentiment in public opinion research circles that the underdog tends to benefit because supporters are further energized."
As the first to announce a significant lead for Chocola about 21/2 weeks ago, Vargus said a second visit by the president would indicate a turn in fortunes and trouble for the Republican. Following his Sept. 5 fund-raising visit, Bush stopped in South Bend for a political rally Thursday.
"I certainly thought at the time it may mean Chocola was in trouble," Vargus said. "Now, I don't think it spelled trouble at all, but just a way to ensure voters turn out on Election Day."
Vargus said he released his new, unsolicited poll to The Truth and WNDU-TV first because they are past clients.
I continue to remain highly suspicious of this outfit's methodology, and don't believe the margin is anywhere near 15 points. However, like other polling in this race, they are showing movement in Chocola's direction.
If you want on or off my poll ping list, let me know!
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.