In Defence Of Polling

During the 2012 presidential election, Republicans had to torture opinion poll data to suggest Mitt Romney had a chance. Australians have had their own versions of opinion poll trutherism ever since Kevin Rudd was rolled.

But in recent weeks, the standard anti-poll sentiment has taken a more desperate air. On the weekend, Wayne Swan protested that leadership polls are “disrespectful to voters”.

First, let’s accept all the criticisms. Opinion polls are imperfect. Their regular ups and downs are often within a margin of error. Respondents can only answer the questions put to them, and only in the form those questions are asked. They do not usually show how much voters care about specific issues. That level of meaning has to be teased out with other, more qualitative methods.

And polls certainly cannot predict the future. Opinions can change. So the more hypothetical a question, the more uncertain the answer is likely to be.

But even with all these problems, polls are still the only actual empirical evidence we have about what citizens think about their government in between elections.

You’d have to be pretty arrogant to dismiss that.

For all the conspiracy theories about Newspoll, for all the embarrassed self-reflection from journalists, for all the blithe dismissals of polls by Labor partisans, surely it’s good to know what the public thinks.

A bit democratic, even.

The first regular political opinion poll in Australia was held in October 1941. It was a Gallup poll, run by what is now Roy Morgan Research. (There had been another poll a year earlier, commissioned by the Department of Information, but that was done for the government’s benefit.)

The result was conclusive, and important. Fifty-nine per cent of Australians supported equal pay for men and women doing the same work.

This was exciting. Before opinion polls, the only systematic accounting of public attitudes was an election every few years. An election is a very blunt instrument. Now the mind of the public was suddenly revealed to those whose jobs depended on it.

Over the next few years, questions poured out. How did Australians rate the government’s wartime performance? How were they enjoying price controls? Should horse racing be restricted during the war? Was it time to revisit the prohibition question? Should Sunday shopping be available for servicemen?

Yet the new science of public opinion was not universally embraced.

In July 1942, the Labor premier of Tasmania, Robert Cosgrove, demanded the Commonwealth government “take steps” to control the publication of Gallup polls. He reported that the Hobart Trades Hall Council didn’t know anybody who had actually been interviewed, and were therefore convinced the public was being misled.

You often hear that we’re living in a great age of polling. In his 2010 Quarterly Essay, Trivial Pursuit, George Megalogenis said that Australian politics changed for the worse when Newspoll switched from monthly to fortnightly in 1992.

But if there was a real breakthrough moment for polls in Australia, it wasn’t Newspoll in 1992. It was the battle against Ben Chifley’s attempted nationalisation of the banks at the end of the 1940s – the first true public opinion campaign in Australian history, and one which relied heavily on polling.

The political class has had more than 70 years to get used to regular opinion polls. Australia’s poll sceptics need to admit their problem is not with polls. It’s with politicians.

Let’s not blame Parliament’s leadership soap opera on the fact that politicians now have more information about what the public thinks of them.

After all, it is pretty well established by now that Kevin Rudd wasn’t dumped in 2010 because he was doing poorly in the polls but because his most senior colleagues didn’t like him. That is, Kevin Rudd was the victim of crude factional politics. His modest Newspoll decline was just the excuse.

In the Wall Street Journal the day before the 2012 presidential election, Peggy Noonan wrote that, “Nobody knows anything. Everyone’s guessing … Who knows what to make of the weighting of the polls and the assumptions as to who will vote? Who knows the depth and breadth of each party’s turnout efforts?”

Her scepticism was in contrast to Nate Silver’s New York Times prediction that Barack Obama had a 90.9 per cent chance at victory. In retrospect, Noonan looks like a fantastist – someone who holds onto an outlandish view and dismisses all contrary evidence.

Polls are the best information we have about what people think about their government and public policy questions.

Politicians like to say they don’t pay any attention to polls. How terrible if that were true.