Bob McMullan (Fraser, Australian Labor Party) Share this | Link to this | Hansard source

The issue I wish to raise in today’s grievance debate is the manner in which the media in Australia cover opinion polls on voting intentions and related issues. I raise it not for its short-term significance to the Labor Party at a federal level—it may or may not have some; that is not clear and it is not something on which there would be unanimous views—but for the long-term perspective, as one of a number of issues raising concerns for me and others about the nature and health of our democratic processes. We find ourselves in September 2006 in a position where objective evidence and public perception are miles apart. On a sustained basis, this cannot be a healthy thing.

Many Australians, probably most, would be surprised to know that the unequivocal evidence from the published opinion polls is that the Labor Party has been clearly and consistently ahead in federal voting intention surveys for the last six months, and not just a little bit—between two and four per cent ahead on either trend or average analysis of both Newspoll and the Nielsen Sydney Morning Herald or Age polls. Over the last six months—which is in reality the short term, although for much political analysis it can pass for the long term—there have been 15 polls by those two major reputable polling organisations in this country, published by the Australian, the Sydney Morning Herald and the Age. Of those 15 polls, the ALP has been ahead in 12: six out of six by ACNielsen, nine out of 12 by Newspoll. The trend line analysis suggests the normal statistical variation around a pretty stable trend, which has the ALP on between 51 and 52 per cent two-party preferred. Yet this is a million miles away from public perception. Why is that?

It clearly has something to do with the way the polls are reported. It is not the polls themselves; I believe Australia is well served by competent and ethical major polling organisations in this country, although I do think—and I will try to find time to make a comment—that some could make the statistical validity and circumstances of their polling a little clearer on their websites. But I do think they are first-class organisations. I think there are three possible explanations for the distance between the reporting of the polls and the polls themselves: biased reporting, incompetent reporting or a mindset which influences perceptions and flows through to the coverage of the polls. It is always very easy to complain about media bias—that is not what I am doing, and I do not think the individuals who write the stories are incompetent. In my view, the principal problem is in the third category. Commentators believe that the incumbent government is likely to win and therefore they interpret the material which comes before them in that light.

Let me give you an example. On 15 August, when the coalition had one of its few reported leads of the previous six months, the commentary in the Australian of that day was:

THE Howard Government has opened an election-winning lead …

In fact, the poll showed a 51 per cent to 49 per cent lead to the coalition. Yet on 29 August when the result was reversed, with a 51 per cent to 49 per cent lead to the opposition, there was no similar report. There are other aspects of the 29 August coverage of the poll result which I will come back to later. The point I am trying to make here is that we had mirror-image polls but significantly different reporting of them by the same newspaper. The polling organisation accurately reported the change, but the representation of it was dramatically different.

My point is not that I believe that the ALP should have been reported as having ‘opened an election-winning lead’. I do not think we do have such an election-winning lead; I do not think anybody does. Properly reported, we would have a consensus that no-one is significantly ahead. The opposition is, and has been for at least six months, seriously competitive, and at this stage the election is too close to call. I am not here trying to make a partisan Labor Party point—it may well not be in the Labor Party’s interests to have such a perception abroad—but that is the accurate situation. My real concern is a long-term one: the credibility and quality of public discourse around our major democratic institutions is an important aspect of any attempt to lift the standing of and respect for our political processes and governance in Australia. There are many other aspects of this issue which need to be addressed as we endeavour to review, reform and rehabilitate our political processes. This is one which should be addressed.

Too often, the Australian media reports on opinion polls as if the mathematical laws of probability had been abolished and the inevitable margin of error in sample based polling did not exist or did not matter. I must acknowledge, particularly today, that there was this morning a noteworthy and laudable exception to this unsatisfactory approach, from Phillip Coorey in the Sydney Morning Herald. He accurately reported that a mere one-point movement over the past month in the Nielsen poll on which he was commenting falls into the margin of error of the best conducted of polls. I congratulate Mr Coorey on that; it is a welcome exception to the pattern of coverage.

This is in stark contrast to the report of 29 August in the Australian to which I referred before, which breathlessly said, ‘T3 row has ALP in front again,’ and purported to analyse the fact that the poll had gone from 49-51 to 51-49. Everybody knows that is a statistically meaningless situation and had they been reporting properly they would have said that the stability around the trend of a small lead for the Labor Party was sustained. I might also say that that poll breathlessly reported, as if it were a fact, that the polls suggested that there had been a five per cent increase in the primary vote for the Labor Party. That would mean that 600,000 people in Australia had changed their minds in a fortnight outside the election period. Nobody seriously thinks that that has ever happened. I certainly do not think that happened in that fortnight. It would have been nice because it would have been 600,000 changing in my direction had it been true. But I regret to say I am sure it is not true.

Let me make a contrast to the way in which the polls are reported elsewhere. The example that I am going to use just happens to be the one that was most readily available to me on the web today. It is a CNN USA Today gallup poll which was taken during the American election. At the foot of that web poll there was a note which said:

One can say with 95 percent confidence that the margin of error is plus or minus 4 percent.

That is an absolutely proper point to put and, to be fair, if you go the Sydney Morning Herald website you will find in the fine print—not exactly the same because the poll is slightly different—a proper reference to the margin of error. You do not tend to find it in the news coverage but it is properly reported there, so I do not wish to mislead with regard to that.

This is not something that the parliament should attempt to fix by changing the law. Laws about how things should be reported are dangerous. If you do it once, even for the best of purposes, it can lead to somebody doing it for not such good purposes next time, even if the High Court were to allow such a law following its heroic interpretation of implied freedom of speech in the Constitution. I think that we need some sort of code between the major polling organisations and their publishers to treat the public with more respect and publish and report opinion polls within a framework which reflects the reality of sample based polling on a very frequent basis.

That is my essential thesis today. We have a dangerous disconnect between public perception and the reality. It is being fed by distorted reporting—I do not claim bias—in our major media by otherwise very reputable journalists in reputable newspapers. I hope that they will take this in the manner in which it is intended as the start of a process of a better debate about politics, governance and our democracy, because in this area and in many others over the months ahead I intend in parliament and outside to raise a number of issues about the state of our political system, our governance and our democracy. This has been just the first instance.