Beware of candidates bearing polls

After the last two presidential cycles, Americans who follow politics closely have come to accept practically daily reports on polls as a fact of political life.

There are so many public polls out there these days that there are few surprises on election day, although "too close to call" won both the 2000 and 2004 presidential races.

But those are polls meant for public consumption, usually sponsored by a media outlet. Want to know what people think about Bush's Social Security plan, or Tom DeLay's troubles, or abortion, or the New York mayor's race? You can probably find it at the Polling Report.

Candidate or campaign polls are different. They are, as one esteemed Wisconsin legislator used to say, a horse of a different feather.

Those are polls taken for a candidate's campaign. This early in the 2006 governor's race, a campaign might take what is called a benchmark poll, to get the lay of the land, determine which issues are uppermost on the minds of the voters, and perhaps test some general themes.

Such polls will also ask respondents to say how much they like or dislike a series of public officials and candidates, whether they think they will vote to re-elect the incumbent, and maybe how they would vote in a hypothetical head-to-head matchup between two candidates (or more, if there is a bigger primary field.)

The "horse race" questions -- the head-to-head matchups -- are the least useful information to be gathered from a poll at this point. Why? Because the general election is more than 18 months away (the primary a mere 16), people are not focused on the race and know little or nothing about most of the prospective candidates.

Polls done for candidate campaigns are useful for strategic purposes. It is rare that they are leaked or released for public consumption, as selected portions of Scott Walker's campaign poll were last week.

In 20 years of working on campaigns in Wisconsin, I've seen dozens if not hundreds of polls. I cannot think of more than a couple of instances, however, where campaigns I have been connected with have made polling numbers public. The media ask all of the time if we've done any polling or what our polls show. We just say, "We don't talk about our polling. It is for internal strategic use." Period.

There are times when it's tempting, because you have numbers better than the ones someone else is giving to the newspaper. But once you open the door and begin to share that information, it is very hard to close it again. Why won't you show the media the next poll, they'll ask. Is it because it's bad news this time? It is a very slippery slope.

That brings us to Scott Walker, who released a memo from his pollster delivering two messages: Walker is the leader for the Republican nomination for governor, and the incumbent, Jim Doyle, is vulnerable.

Why would Walker do that? Obviously, to try to take the luster off of Mark Green's weekend entry into the race. To try to persuade Republican donors to give to Walker. To try to create some momentum and a bandwagon effect. Or, at the very least, to get some Republicans to hold off for awhile before making a commitment to Green.

The numbers on Doyle's vulnerability are intended for Republicans outside of the state as well, to persuade the national GOP to target Wisconsin and pour money in here. That could happen, but if it does it won't be based on any poll from Walker. And it won't happen until October 2006, when the Republicans have a nominee and can see how close the race really is at that point.

If you take Walker's poll at face value, maybe it accomplishes some of those objectives. But the media would be well advised to be much more skeptical when it is given part of a candidate's poll. The Walker memo is typical. It shares a few selected pieces of good news for Walker from a long questionnaire that asked all sorts of other things they are not telling us about.

Do you suppose there was a head-to-head matchup with Walker and Doyle, for example? Walker tells us he beats Green in a primary matchup, and gives some generic numbers on whether people would vote for Doyle, whom they know, or some nameless Republican who right now has no negatives because he doesn't even have a name in the survey. How did Walker do against Doyle? If he beat him, we would have heard by now.

What were Walker's positives and negatives among the general population? Did more than 25 per cent of the people even know who he was? How many outside of the Milwaukee area?

Again, if the numbers were good, he would have shared them.

So next time a candidate wants to share responses to one or two questions, the media should ask to see the survey. The whole thing. What's the part you aren't telling us? What was the context for what you are releasing? What was the bad news?

(Green's response was that Walker's release of the poll "demonstrates they have already lost focus on what matters most to voters in Wisconsin. I’ll give them a hint – it’s not a poll 18 months before an election." OK, I'll give Green a hint, too: What matters most to Milwaukee voters is not photo ID cards for voters, which is what he chose to highlight in his announcement news conferemce there. Maybe he should pay for a poll and discover there are other issues, like jobs, health care, and more.)

One other important fact, but one that virtually everyone in the media ignores: You can't really compare a poll taken by one firm for one client with one done by another firm for another client. The methodology and samples will be different. In most cases, it's the old "apples and oranges" comparison. If you want to see any trends, you need to be comparing two similar polls.

The Wisconsin media, however, usually treat all polls equally -- one taken by a college class is as good as one taken by one of the country's top political pollsters. So you can be up one day in a professional poll and have a story saying you are falling the next day, based on another poll that isn't half as reliable. If voters really are affected by those stories, it is like Russian roulette for the candidates, especially in the final weeks when perceptions do matter and can affect turnout, too.

By releasing a poll now that shows himself to be 14 points ahead of Mark Green, Walker has guaranteed himself a series of future stories showing that Green is gaining on him and narrowing the gap. If Walker does win the primary, the margin is likely to be closer to 1.4 per cent than 14 per cent. So, no matter how things go, Green is guaranteed to show "momentum" as the race tightens up.

That's one reason many candidates prefer to be the underdog. If you lower expectations, the next poll showing you close or ahead will seem like great news for your campaign. If you have proclaimed yourself the frontrunner with a big lead, you risk the perception that your campaign has stalled when polls show that your opponent is gaining on you -- which he inevitably will.

It is too much to hope that the news media would refuse to report on candidate polls. They love those horse race stories. But the Journal Sentinel does get some credit for where the Walker story played -- back on page 4 of the Metro section, at least in my edition. And it did point out the lack of any head-to-head numbers. That's a good start for the '06 cycle.