Who's to blame for that bogus arts tax poll?

Kari Chisholm

As noted in our Punditology Challenge, there was an overwhelming consensus among the smart set that the city arts funding tax was going to lose. Just 22% of the consultants, staffers, elected officials, academics, and other self-appointed pundits thought it would pass. And of course, it passed with flying colors - with 62% of Portlanders voting in favor.

In large degree, that was because of the poll released October 29 by the Oregonian and their pollster Elway Research. That poll showed that 48% of voters were opposed, and just 28% in favor.

As noted at the time by JS May (the top fundraising executive at the Portland Art Museum), that poll was deeply flawed. And in the most basic way. Here's the question that Elway asked:

"There is a measure on the ballot that would assess each adult above the poverty level $35, would you/did you vote for the arts tax?"

Seriously. That question, of course, bears no resemblance to what's actually on the ballot - and actually obscures what the proposal would actually do. Here's what the ballot said:

26-146 Restore School Arts, Music Education; Fund Arts through Limited Tax.

Question: Shall Portland restore arts, music for schools and fund arts through income tax of 35 dollars per year?

It's fair to argue about whether ballot titles and questions are written accurately and fairly. But the fact is, that's what's on the ballot - and if you're trying to project what voters will do when faced with that question, it's probably worth asking that question. At a minimum, the poll probably should have included the word "schools" or "education".

Over at Oregon Arts Watch, Barry Johnson tried to figure out what happened. And it's a fascinating tale:

I emailed and called Stuart Elway of Elway Research, a respected polling company based in Washington, for an explanation. I also emailed my old boss at The Oregonian, Peter Bhatia, the paper’s editor, to ask him for his post-election thoughts on the poll and some other questions about the way the paper conducted its coverage of the measure.

I reached Elway, and we talked by phone. He agreed that the problem with his poll was the question itself. “I don’t think that adequately represented the measure to the respondents,” he said. “The wording we used was just not accurate.”

The paper sent him the question it wanted polled, he said, and he edited it a little before the poll was conducted, “but obviously not enough.” Generally, Elway said, his firm’s practice is to read the ballot title as it is on the ballot. This time, it did not, and the results were way off. The question appeared in the Portland part of a larger statewide survey of various races (Elway Research over-sampled the Portland part of it to get representative results of the city races), and the rest of the results were reasonably accurate. But the arts tax question “kind of dropped off the table,” Elway said, as he was making sure his samples would work on both statewide and city races.

What if he had noticed that the question The Oregonian generated was not in line with Elway Research’s usual practice of reading the measure’s ballot title? Elway says he would have suggested that they change it and discussed it with them. He maintained that The Oregonian wanted accurate results from its poll.

In other words, while taking ownership for lax oversight, Elway laid the blame for the bad poll squarely at the feet of his client, the Oregonian. (Which begs the question - why is the newspaper writing poll questions, instead of relying on their pollster to do it? Why hire experts if you're not going to use 'em?)

Bhatia declined to discuss it, other than to say that they hadn't discussed the polling result with Elway.

I strongly encourage you to pop over and read Johnson's full post. He delves beyond the polling into an analysis of the overall media coverage of the campaign for 26-146 and suggests that the Oregonian and Willamette Week may have made a newsroom decision to put their thumb on the scale. Here's a taste:

Neither paper ever did the basic explanatory work on the measure itself. ... Neither asked about what sort of education and outreach programs arts organizations are expected to do now or try to figure out how important that work is. If part of the money is going to arts groups engaged in programs like these, what can we reasonably expect from our investment?

Check it out.

A double hat tip to JS May for the initial pre-election note about the errant polling, and the pointer to Johnson's excellent coverage at Oregon Arts Watch.

By Kari Chisholm Nov. 26, 2012