Pollster Stuart Elway of Seattle's Elway Research, Inc.
OLYMPIA, May 22.—The latest polls in Washington’s closely watched governor’s race might make you wonder how fickle voters can be. For most of the last year, they have shown Republican Rob McKenna with a whopping lead over Democrat Jay Inslee, sometimes as much as 10 points. But if you can believe what you’ve been hearing lately, the race has become a dead heat.
Four polls in a row say so. A February poll from Public Policy Polling, a Democratic-leaning firm, showed it tied up at 42-42. Another that month from Grove Insights, commissioned by the Service Employees International Union, showed the two candidates even at 38-38, and a March poll from the same source actually showed Inslee ahead 38-34. A poll this month from Survey USA showed McKenna ahead by a mere 40-38.
Is it a trend? Are the two candidates pulling even?
No, says veteran Washington-state pollster Stuart Elway. It doesn’t make any sense. “Certainly the campaigns are ramping up, but there haven’t been any events or debates or anything like that that might move big numbers.”
More likely the results reflect differences in the way polls are conducted, he said. Washington State Wire spoke with Elway Saturday following a presentation to the annual conference of the Mainstream Republicans of Washington. Elway, who has spent 37 years in the rather competitive business of polling in this state, says you can’t just take a spate of polls and plot them on a chart and draw a trend line. Not all polls are created equal. He maintains his own February poll, which showed McKenna ahead 45-36, represented the sharpest snapshot that could have been taken at the time, and it’s hard to believe that anything has changed much since then.
Inslee Campaign Claims Momentum
Since the beginning of the year, polls in the gubernatorial race have varied widely. About the only trend in evidence, Elway argues, is that different pollsters come up with different numbers.
As Elway observes, there really isn’t anything happening in the governor’s race right now, at least as far as the public can tell. That’s part of the problem. The polls are among the few things anyone can talk about. Ever since the two candidates announced their campaigns 11 months ago they have been doing what they are expected to do at this stage, building organizations and raising money. So far they have raised a total $10 million, most of which they’ll spend in the fall on TV ads. But of course, none of those ads have aired so far. They’re making speeches to constituent groups as they beg for endorsements, and along the way they’ve been releasing the occasional white paper on jobs, education and reform of state government. Such are the things that attract note in Olympia. But the first real head-to-head battle that might draw attention from anyone outside the political trade is still three weeks away, a debate sponsored by the Association of Washington Business in Spokane June 12.
In the absence of other news, the Inslee campaign is arguing that the poll numbers demonstrate a trend, plotting them on a chart in precisely the way Elway says shouldn’t be done, and declaring that the race is slowly but surely turning blue. A recent fund-raising letter calls the three most recent polls “a trend that shows this race is a dead heat,” and urges a $3 contribution to “help us keep up our momentum.”
Elway says anyone who reads polling data that way is bound to have a surprise on election day.
Some Snapshots are Blurry
Inslee campaign cites the same polls to demonstrate "momentum."
Yes, polls are a snapshot of what people are thinking at any given point in time, he says, and that can change, and sometimes you can see movement. But you have to realize that some snapshots are better focused than others. “I think the broad answer is that people have not settled their minds,” he says. “But a more technical answer is that there are different methodologies, different turnout models, different sampling, and so the polls are going to show different things.”
For one thing, he says, all the other polls these days use robo-calls rather than human interrogation. An automated voice asks whomever answers the phone to respond by punching a few buttons. “They ask you if you are a registered voter, and you punch yes. Are you going to vote this time? And you punch yes. And then, what is your age? Well, let’s see, I’m 65, says the 14-year-old girl who is answering the questions.”
Elway says random-dial surveys have another problem – telephone exchanges place the preponderance of numbers in urban areas. He says his surveys use a more reliable methodology, which starts with a list of registered voters to determine who will be called, balanced for geographic areas. To make allowance for the growing number of people who have only cell phones, he uses a cell phone list as well as a landline list. About 10 percent of his surveys are completed via cell phones; the cell-phone-only population is perhaps double that, he says, but he does try to account for the trend in telephone usage. Elway says he doesn’t believe the widespread use of caller I.D. has skewed phone-poll results, and he cites recent research to support that. But even if it did, that might be considered a non-unique disadvantage – all pollsters rely on the phone.
Results also can be altered by the way questions are asked. Elway favors a soft approach: “As you know, there is an election coming up; the candidates are Republican Rob McKenna and Democrat Jay Inslee; as things stand today, how are you inclined to vote?” He rotates the names and offers voters a point scale to determine how strongly they feel. Elway notes that his poll results typically reflect a greater number of undecided voters than others, which he believes demonstrates greater accuracy.
What it means, he says, is that polls offer such wildly different results that you can’t really take two that come from different sources and are done in different ways and declare that there is a trend.
Reason for Suspicion
You might get a little suspicious, Elway says, if you consider that Grove Insights took two polls for SEIU, four weeks apart, at a time when nothing was really happening in the campaign, and one found the race dead-even and the other showed Inslee ahead. “They didn’t get enough interviews around Green Lake, so they went back and got a different number,” he jokes.
There’s one way to measure accuracy: Even though things can change dramatically between the start of a race and the finish, there’s always an official poll at the end, the one that takes place on election day. Elway likes to point out that in 2008, other early polls showed the gubernatorial race to be neck and neck; his showed Gov. Christine Gregoire cruising to reelection, and that’s the way it turned out. His favorite story from the polling wars comes from the 2006 race between Republican Mike McGavick and incumbent U.S. Sen. Maria Cantwell. Survey USA released a poll in late September that showed McGavick gaining against Cantwell – he was down just 12 points. Another poll by Rasmussen Reports that week showed Cantwell ahead by only six points. Earlier polls showed Cantwell with nearly a 20-point lead. It was surely a trend. “The Seattle P-I ran a story above the fold – it was like World War II,” Elway says.
Meanwhile, Elway’s polling at the time was showing Cantwell ahead by 18.
On election day in November, Cantwell won by 17.
Reading the Tea Leaves
Elway addresses the annual Cascade Conference of the Mainstream Republicans of Washington.
Six months out, Elway says it makes as much sense to look at the fundamentals as it does head-to-head survey numbers. Historical elections statistics and polling data give Washington a decidedly bluish tint. The state hasn’t voted for a Republican president since 1984 or a Republican governor since 1980. For statewide constitutional offices, Democrats have won 46 of 57 races since 1992, and of course the Legislature has been in Democratic hands since 2004. Voters identifying themselves as Democrats number about a third; as Republicans, about a quarter.
“It looks like we’re a blue state, no getting around that,” he says. “There used to be a time when Washington would appear in the list of battleground states early in the [presidential] campaign. That no longer happens. If you see Barack Obama out here anywhere but Medina [for a fund-raiser] you know he’s in big trouble.”
On the other hand, he points out that Washington voters have a tradition of voting for the candidate rather than the party. And his polling finds that voters have a favorable impression of McKenna by a three-to-one margin; Inslee meanwhile, runs even.
“Can McKenna thread the needle? By that, I mean, can he run the perfect campaign to overcome these odds that we have been looking at in this state, to win in a blue state? There is a lot of optimism for McKenna, and ultimately the reason is that he is running a strong campaign. Moderate Republicans are the only kind that ever win in Washington state. You look back almost forever, and it is only moderate Republicans who win. When Republicans drift right a little bit, they get slaughtered at the polls.”
Democrats will have to portray McKenna as an extremist, Elway says, and he noted that Democrats have scored gains nationally by promoting a “gender gap” strategy that plays well with college-educated female voters. Which may explain recent efforts to inject abortion into the governor’s race in this state, for example. “We are talking about all sorts of variables, and campaigns and personalities and the candidates matter and can move those perceptions. But with that sort of background that makes it pretty clear why the Democrats in this state are driving that message at Rob McKenna.”
Your support matters.
Public service journalism is important today as ever. If you get something from our coverage, please consider making a donation to support our work. Thanks for reading our stuff.
