On the unknown quantities in polling

Peter Brent writes: Re. “Poll position: a dearth of ‘don’t knows’ does not mean ‘let’s make it up’” (yesterday). Australia Institute executive director Ben Oquist’s contribution to Crikey yesterday. I invite readers to look at the piece of mine he objects to instead of his characterisations of it. It already refutes much of what he claims in response.

My column went to several things, starting with the gimmick of surveying Coalition-held seats safe in the knowledge that the responses would be reported as those of “Coalition voters”. But the major point, the one Oquist devoted most energy to demolishing, was that respondents weren’t given a “don’t know” option in the survey question and that therefore many, perhaps most, of those who replied that penalty rates should “stay the same” would have opted for “don’t know/no opinion” if given the opportunity. Blogger Kevin Bonham, who Oquist also took exception to, went into it in more detail than me. 

In the article I performed a calculation (despite Oquist’s quotes I didn’t use the term “best guess” — nor is “thought experiment” mine) assuming that proportion to be 50%.

According to Oquist, in “June 2015 [the AI] first asked voters a similar, but not identical, question about penalty rates. What we found was that only 8% people chose ‘don’t know’ when asked about cutting penalty rates.” Terrific, great, let’s see it! It wouldn’t explain why December’s survey omitted that option, but yes 8% would render its absence not very important. I’m unable to find this survey on the Institute’s press releases and Oquist didn’t link to it. No luck on a Factiva search either.

We have to take his word for it that it was “similar” to December’s. But for the rest of his column that 8% becomes the percentage of undecideds in December. Oquist’s leap confused him somewhat, in that “8%” … is a fair bit less than the 50% imagined by Brent” and “Brent’s ‘best guess’ about the salience of the penalty rate issue was six times higher than the result of our June poll.”

But my assumption in that calculation wasn’t that 50% of all respondents would have chosen “don’t know” if offered, it was that 50% of those who nominated “stay the same” would have. Most Crikey readers will understand the difference.

Oquist for some reason objects to my statement, of something approximating the bleeding obvious, that “the fact of majority voter opposition to a proposal doesn’t necessarily mean much politically” but then a few pars down agrees that “Not every issue changes people’s votes.”

Indeed. For example, most surveys over time show voters prefer Labor’s health and education policies to the Coalition’s. If all elections were primarily about health and education, the ALP would rarely be in opposition. But that’s not the case.

And we don’t know how important penalty rates would be at the 2016 election. Blithely referring to the evolved mythologies about the reasons for the 2007 result doesn’t cut it.

Oquist complains that I didn’t call him before putting pen to paper. So what? Does he phone everyone he critiques?

I choose to minimise my personal exposure to politicians, lobbyists and other “mate, just between us” snake oil practitioners. It’s healthy for the soul and maintains professional distance. I was writing about the poll and the swifty the Australia Institute pulled over much of the media.

If he had been engaging in good faith, Oquist would have linked to the June survey. Potentially, if it is all he claims it to be, it would knock out my 50% assumption (though more by a factor of three than six).

But he wasn’t, and after that effort I’m even surer the “don’t know” option was left out deliberately to boost apparent support for the penalty rates status quo.

Still, it’s not too late for Oquist to share this mysterious June survey.