Friday, March 10, 2006

Salvi’s poll: McSweeney and Salvi in Statistical tie for First

Revised at 2:00 pm on Friday

The Salvi spin machine kicked into high gear last week with a March 2, 2006 memo from its pollster that it circulated to the media. Some of the media took the bait: hook, line and sinker.

But to put the recent poll in perspective, we should note that Team Salvi [that would be Kathy and Al] had previously commissioned at least two other polls from Public Opinion Strategies [“POS”], meaning that firm has pulled in at least $50,000 from Team Salvi. Not a bad days work.

More than seven months ago, the “Al Salvi” campaign team commissioned a poll from POS. Neil Newhouse of POS, in his memo for Team Salvi, which was leaked around to the press, told us, “Al Salvi is well-positioned to both win [sic] the GOP primary election in this district and to defeat …Rep. Melissa Bean in the election.” Now, that was a convenient set of results and analysis by Newhouse because at that time, Al was thinking of running—so he couldn’t have been happier.

Now, that poll, assuming it was done right, did show Al Salvi with a statistically significant lead over Rep. Bob Churchill and a much stronger lead over Bartels and McSweeney, neither of whom had started to engage in a vigorous campaign. However, at that point, it was a name recognition poll, and the two who had run for office previously were clearly ahead of the others—but that was similar to “being full of Sound and Fury, signifying nothing.” Although it did signify at least about seventeen thousand dollars for Newhouse and POS.

According to Newhouse’s memo of last week, POS did another poll last December, showing McSweeney ahead of Kathy Salvi and Bob Churchill. Newhouse does not report the margin of error, so we can’t tell if the McSweeney lead was statistically significant. If I had to guess based on the margin of error for the other polls by POS, I would guess it probably was a statistical tie. This is what we would expect: McSweeney had campaigned quite a bit from June to December, done direct mail and therefore caught up in name recognition with both Rep. Churchill and Kathy Salvi.

Kathy has acknowledged that she benefited quite a bit from the Salvi name recognition, which recognition had been built by husband Al. As Kathy Salvi said last fall, the Salvi name is like Kleenex. I think she realized later that likening Salvi to a disposable tissue used to wipe your nose was perhaps not the image she was seeking. Oh yes, ring up another seventeen thousand dollars, or so, for Newhouse’s polling firm. Those two page memos are quite profitable.

That takes us to last week’s release of yet another Salvi commissioned poll, seventeen thousand dollars for yet another two-page memo to spin the media.

Newhouse writes breathlessly in the March 2, 2006 poll: ”Salvi has vaulted to the top of the ballot…and has built a solid lead over opponents David McSweeney and Churchill.”

But, hold your breath, boys and girls. That statement is not true. Let’s remember some basic statistics. When you take a poll, you are sampling to estimate what the true population is. Key to drawing any conclusions is the margin of error, which Newhouse professionally tells us to be +/- 5.66 %, which we will round off to +/- 5.7%. So, when Newhouse tells us the sample poll his firm did found Salvi to be leading with 29% to McSweeney’s 19% and Churchill’s 17%, his poll could reflect a true population with those numbers. But, it could also reflect a true population in which Salvi had 29% minus 5.7%, or 23.3% and one in which McSweeney had 19% plus 5.7 %, or 24. 7%.

In other words, the polling done by Newhouse’s firm is also consistent with David McSweeney being ahead of Kathy Salvi 24.7% to 23. 3%. So, in short, the margin of error for the poll is too high relative to the distribution of numbers to conclude that Kathy Salvi is ahead of David McSweeney or vice versa. I could qualify that and say, statistically speaking. But, for polling analysis, statistically speaking is the only kind of speaking to do.Yes, Kathy could say she did have a statistically significant lead over Rep. Churchill.

The question to ask: does the lead in the poll exceed double the margin of error. If yes, the lead is significant. If no, it is not. And never leave home without your margin of error.

Salvi’s lead over McSweeney in the POS poll? 10%. Double the margin of error is 11. 4%. Is the lead significant? No.

Salvi’s lead over Churchill? 12%. Double the margin of error is 11.4%. Is the lead significant? Yes. Moreover, another important item Newhouse forgot to disclose were the negative Robo calls the Salvi campaign was pumping into the 8th CD households at the same time the Salvi commissioned poll was being conducted.

And, negative calls into a household being polled are not just like a negative mailer or negative ad for the person being polled. Think about it: the person gets a call asking for her opinions about six different candidates. Just before that call, or so, comes a call telling the person being polled what a scoundrel one of the candidates is. Could that bias the poll? What do you think?

Indeed, cynics might think that after a day of not such good results the Salvi team might have ordered up the negative Robo calls to give them a little help, kind of like calling for air support when things are not going well for the troops on the ground.

Why would Team Salvi want to orchestrate their polls to show a false lead, one, which was not statistically significant? Well, to get some positive spin in the media.

And, look- it worked. The Northwest Herald reported [See here], “Poll: Salvi leading by 10 percentage points.” The Northwest Herald reported the margin of error, but it neglected to use it in its analysis or in its headline, which should have read, “Salvi and McSweeney in statistical tie, notwithstanding Salvi’s negative Robo calls.” In other words, without the bias of the robo calls, the true population might be that McSweeney has a statistically significant lead over Salvi.

The true irony here, apparently missed by the Northwest Herald, is that the speaker on the negative Robo calls was Teresa Bartels, who dropped out of the 8th CD race at the end of January. But, two days before she dropped out, Teresa went negative at a Candidate Forum in Round Lake, arguing all the other 8th CD candidates had “political baggage,” that would keep them from winning.

Bartels argued that Kathy Salvi’s baggage is that she is a trial lawyer who “spent a good deal of time working behind the scenes in her husband’s campaigns,” losing two statewide campaigns. Bartels argued that David McSweeney’s baggage is his decision to run against Phil Crane in 1998 and to spend money on his campaign. Bartels argued that Churchill’s baggage is his ties to George Ryan and Lee Daniels.

Quite a peculiar exhibition by Teresa Bartels, which judging by the lackluster reception of the audience, might have caused Bartels to question the wisdom of the person recommending this to Bartels [Mike Zolnierowicz, then the Bartels campaign manager, was recently named the Communications Director for the Illinois GOP, replacing Matt Leffingwell. Zolnierowicz also has Katherine Harris on his resume. Congresswoman Harris is someone who helped "W" when she was Florida SOS with a few key election rulings on his way to winning Florida and becoming President in 2000, but Harris' Florida Senate primary campaign seems to be in trouble, and the Florida Republican Party may be looking for a replacement Senate candidate].

Two days after Bartels' harsh negative attacks on her major opponents, Bartels drops out and the next day, she announces her support for Kathy Salvi. Engaging in some intellectual somersaults and fancy footwork, Bartels tells us what a great candidate Kathy is, notwithstanding her harsh words for Kathy and husband Al just a few days before.

Then, to top it off, Bartels does negative Robo calls in which Bartels has the chutzpah to complain of McSweeney’s alleged negative campaigning six years prior to that. Further the negative Robo calls are likely to have helped create misleading poll results which are then marketed deceptively to the media by Team Salvi, and then, topping it off-- even further-- Team Salvi runs around bad mouthing McSweeney for pointing out, accurately, that Kathy Salvi opposes [as she said on Public Affairs, see last few minutes of the TV program with Kathy Salvi, taped on Oct. 9, 2005] caps on Non-Economic Damage awards for the clients of trial lawyers, which husband Al and she are.[See here].

In short, as of last week, the best argument that Kathy Salvi could make, as to her standing in the polls, is that Salvi is in a statistical tie for first place. This is the best case Salvi could make because, if you take away the Robo call distortions, the polling data might have shown McSweeney with a statistically significant lead. Further, Kathy and Teresa, the odd couple in light of Teresa’s comments about Kathy and other candidates, when Teresa was in the race, don’t have standing to criticize anyone for negative ads or mailers. Not unless they want to be hypocrites.
Jeff Berkowitz, Show Host/Producer of "Public Affairs," and Executive Legal Recruiter doing legal search can be reached at