Two polls in the last two days have demonstrated dramatically different levels of support for a proposed $241 million tax increase to replace the downtown waterfront seawall. One, produced by McGinn's longtime campaign consultant and pollster, Bill Broadhead, and paid for by McGinn himself, says that voters support the mesaure by 70 percent. The other, produced by independent pollster SurveyUSA in collaboration with KING-5 TV, shows support of just 53 percent, seven points short of the 60 percent needed to pass a bond measure in Seattle.

That's a 17 percent discrepancy—far outside any conceivable margin of error, and large enough to raise serious questions about which poll accurately represents Seattle voter opinion.

In an email, Broadhead attributed the difference to two factors: The fact that SurveyUSA polled Seattle residents at random, whereas his poll only included frequent voters (close, Broadhead told me, to "four of four" voters—those who vote in pretty much every election) who had previously voted in special elections like McGinn's proposed May election; and the fact that the SurveyUSA poll told voters how much the measure would cost the typical homeowner per year (around $48), while Broadhead's poll only revealed how much it would cost per $1,000 of property valuation (about 12 cents).

"We believe our question more accurately reflects the actual ballot title voters will see," Broadhead wrote. "We have great respect for SurveyUSA and would be most interested in their polling results if they interviewed likely voters and presented a question that more closely approximated the likely ballot title."

For several reasons, SurveyUSA's poll—though, like all polls, far from perfect—more accurately reflects voter opinion. Here's why:

1) Comparing the demographics of likely special-election voters versus a random sample of the voting-age population, you'd actually expect Broadhead's poll to reveal the exact opposite outcome.

In English: Special-election voters tend to be the oldest and most tax-averse group of voters on the planet. The more (and the less-informed) people you add to a group of voters, the easier it becomes to pass a tax measure—one reason politicians frequently put tax increases on November ballots. A 70 percent approval rating for a tax among informed, tax-averse voters, compared to mere 53 percent approval among a larger, less tax-averse population, has raised eyebrows among Seattle's plugged-in political pollsters.

"Each step you take towards a wider electorate will increase support" for tax measures, says Ian Stewart, a vice president at the polling firm EMC Research. His colleague, EMC VP Andrew Thibault, adds: "You want the highest possible turnout in a tax election because it dilutes the older vote."

At 70 percent, Stewart adds, the seawall would appear to be polling higher than the most popular school levies, and "I can't imagine that that's the case."

2) Elections don't happen in a vacuum. Broadhead and others have argued that "12 cents per $1,000 of property valuation" more accurately represents the language that voters will see on the ballot. However, that ignores the fact that campaign messaging exists—and there's going to be a hell of a lot of it before any $241 million tax increase. Even more: McGinn's measure is widely viewed as a ploy to kill the tunnel—something local unions are pissed off about. Unions have money to spend, and they'll play hardball with messaging come campaign time.

"If this thing gets out of the city council, there's going to be a hell of a discussion in the community about the pros and cons" of raising taxes, says Mike Cate, producer of "Up Front" for KING-5, which jointly produced the SurveyUSA poll. "The public is going to get up to speed, and it’s going to be a different conversation in late April than it is right now."

In other words: Campaigns (ahem, "voter education efforts") exist to sway voter opinion, and it's naive (or faux-naive) to pretend that voters will simply ignore the inevitable anti-tax fusillade that accompanies any major property-tax expansion measure and make their decision based purely on the language on their ballot.

That's especially true, by the way, of the highly engaged, politically astute voters that show up for off-year and special elections.

3) Broadhead, unlike SurveyUSA, has not released his methodology; nor has he revealed what other questions he asked during the seawall survey, or in what order.

Despite the fact that the firm failed to predict the outcome of the mayor's race, SurveyUSA has a generally strong track record; for example, the web site fivethirtyeight.com routinely ranks the polling firm among its top two or three pollsters in terms of the accuracy of its predictions.

4) The seawall measure isn't likely to go to the ballot in May; therefore, comparisons to past May elections—which are, in Seattle anyway, exceedingly rare—are irrelevant. The city council has made it abundantly clear that they aren't ready to move forward on McGinn's proposal any time soon, so the next potential election is the August 2010 primary, when voter demographics will be different (i.e., broader) than an off-year or a special election.
Filed under
Share
Show Comments