Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Opinion polls: the political agenda for government and opposition is greatly influenced by the continuous process of voter opinion polling. The polls also tend to influence the outcomes of elections they predict

JIM WALEY: Tomorrow, Opposition Leader Andrew Peacock begins a five-State full dress rehearsal for the next Federal Election campaign - spurred in part, perhaps, by opinion polls showing his approval rating well behind that of the Prime Minister.

But there are many who would say that those same opinion polls have rendered Peacock's dry run redundant, because the constant sampling of voter moods and misgivings creates a continuous election campaign in which pollsters, not MPs, represent the people, and Party leaders react regularly to the changing demands of the electorate.

Justin Murphy reports on the pros and cons of opinion polls and whether they mimic democracy or mock it.

EXTRACT (November 1986):

JOE BJELKE-PETERSON: One, I will be Premier without any doubt at all. There's no question in my mind or your mind or anybody's mind. REPORTER: Tomorrow's poll shows he may have to make peace with the Liberals to fulfil that boast. Statewide, the Nationals have lost 8 per cent, well below the 36 per cent they need to achieve government in their own right. The drop was 7 per cent in Brisbane and a massive 10 per cent in the country. JOE BJELKE-PETERSEN: That is wrong. It is completely wrong.

REPORTER: The Liberals are holding their position, down only one per cent Statewide. Labor on the other hand, has rocketed 9 per cent. The increase in rural support could mean Labor will take seats off the National in the country. JOE BJELKE-PETERSEN: Who took this last one? REPORTER: It's a Morgan Gallup poll. JOH BJELKE-PETERSEN: Oh. They come and go like a morning mist, don't they? REPORTER: Less than one hour after voting finished, the writing was on the wall for the Liberals. The figures were promising for the Nationals who had always maintained they'd win a majority.

UNIDENTIFIED: Fairly obviously, the National Party look as though they'll be able to govern in their own right again, which is the sort of outstanding result, and sort of goes against all of those polls that we were watching. RUSS HINZE: You know, when you take into consideration what the news media have been telling us and what these phoney polls have been telling us for the last three weeks ... REPORTER: You mean the phone-in polls? RUSS HINZE: No, the phoney polls, and you know what I'm talking about.

JOE BJELKE-PETERSEN: Well, why didn't the polls show their support? Simply because, as I've tried educate you people for so long, polls you can't bank on them. And I hope I can get that message over to you people again tonight.

EXTRACT (Federal Election 1984):

MIKE WILLESE: An election which the polls told us would be all over pretty early, very clear-cut, big swing to the Labor Party, maybe 100 seats, and that's not what's happening at the moment.

ROD CAMERON: The first time that we saw the polls played a fairly dramatic effect on the eventual result, was when the sensationalising of the polls hurt Labor in 1984. Now, one of the reasons for the fact that Peacock did reasonably well in that election, was the fact that the polls were quite wrongly overstating Labor's lead. So, the ordinary voter does what many Australians often do, they say, `Well, let's pull that bloke down a peg or two and we'll reduce Labor's lead'.

JUSTIN MURPHY: So, Peacock got the underdog effect?

ROD CAMERON: They thought that he was so far behind that they felt a bit sorry for him.

JUSTIN MURPHY: A not so random sample of opinion polls that got it wrong. But in spite of those embarrassments, you can hardly pick up a newspaper or magazine these days without seeing which Party is flavour of the week, which leader is the more alluring, which issue the most worrying. In place of the brief triennial blitzkrieg, election campaigns are now continuous, with pollsters charting each change of mood, each crisis of confidence, live as it happens. A constant stream of pros, cons and couldn't care lesses that, almost without our knowing it, is changing the rules of politics.

ROD CAMERON: On one level they cheapen the political process. They reduce it down to a horse race, but it's far worse than that. It actually does influence the result in certain directions, and it hurts both parties, and I think there should be a degree of bipartisan support for my call, which is to limit the amount of publicity that polls receive in this country.

GARY MORGAN: Why shouldn't a politician be measured on a day by day basis? We pay them. The electorate pays them. It's wrong for the politician not to be measured. What we're doing is measuring them. We're making them accountable.

JUSTIN MURPHY: For the pollsters, the collecting, measuring and interpreting of opinion may be no more than a statistical exercise, but behind the figures are real people, the voters for whom the issues may compete with Party loyalties and the people they vote for, the politicians whose careers and ambitions are on the line.

ERNIE CHAPLES: If the opinion polls seem to be reflecting something like, Howard being considered a little less trustworthy or a little less of a real leader in the eyes of the public, then many people who might not have seen that initially for themselves, will start looking for that ingredient. I think that can be a problem for tomorrow - yes, it can be a danger for tomorrow's ...

GARY MORGAN: Honest public opinion polls stop the journalist trying to manipulate public opinion, trying to set the agenda. And it's also been used to accurately measure what the public thinks and believes about a lot of issues, whether it's test tube babies or whether we should be a republic or whether we should have a new flag.

JUSTIN MURPHY: So, you bolster democracy?

GARY MORGAN: It's absolutely an essential part of democracy.

JUSTIN MURPHY: Democracy as we now know it, began in Australia nearly 50 years ago, when Gary Morgan's father Roy, a finance reporter, was sent by the Melbourne Herald to America to study George Gallup's newfangled science. Morgan's first survey on equal pay for women appeared in the Herald on October 4, 1941; and he quickly established a reputation for accuracy in predicting election outcomes.

His first mistake, though, was a real lulu. In 1961 Morgan, like everyone else, expected another comfortable win for Bob Menzies. The voters had other ideas and took Menzies to within an inch of the greatest political upset in our history. He scraped back by two seats on a handful of Communist preferences.

GARY MORGAN: We try to do a good job. We work very hard on it and we assess what we're doing all the time, looking at the questions we ask and how we ask them and the results we get. And we make mistakes. Everyone makes a mistake.

ROD CAMERON: But media polls are often wrong because they are done usually with very small samples, often over the telephone, and often as adjuncts to ongoing commercial surveys, so that when the interviewer has got through with talking about Rice Bubbles and Coco Cola, then they'll throw in a few questions on who are you going to vote for. Well, the interviewee treats this about as seriously as they would how many jars of coffee do you have in your pantry?. It's a surface, top of the head judgment, and it's not the way to conduct proper political research.


ANGELA: Good evening, my name is Angela from the Roy Morgan Research Centre, the people who conduct the Morgan Gallup Poll. If a Federal election were being held today, which Party would receive your first preference - Labor, Liberal, National, Australian Democrats or someone else?

JUSTIN MURPHY: The key to getting opinion polls right is the random sample. According to the theory, eliminating human bias from the selection of voters to be surveyed, increases the prospect of getting a representative cross-section of society.

GARY MORGAN: We survey door-knock every weekend. We ask an Australia-wide cross-section of 1,200 people, their voting intention every weekend. And sometimes when we've got a lot of commercial clients, market research clients or social surveys to do where people buy survey data from us, which is our main business, we might ask 2,400 people a weekend. Now, we also do a large number of telephone polls. The telephone polls are getting better; they're getting it closer, but they've still got a way to go.

ERNIE CHAPLES: The problem is that that very simple theory of giving every voter an equal chance to phone in his sample, is easier said than done. The number of people who have telephones is not a random sample of the population. We know that the people who are in more remote locations, who are at the bottom of the socio-economic scale, who've come from more recent migrant groups, who don't speak English for instance, are (a) less likely to have telephones, and less likely to be able to talk to interviewers by telephones. The pollsters will tell you that they have a device for dealing with this, that they load the interviews, that they weight the interviews but, really, that's a pretty unsatisfactory way of substituting real people.

JUSTIN MURPHY: But it's real people who vote, and while media pollsters count mainly how people vote, what the political parties want to know is why they vote that way. The parties find that out through another form of opinion survey - group discussion, which reveals attitudes and how strongly they're held, a practice that raises a further question, answered by Labor's private opinion pollster, Rod Cameron.

JUSTIN MURPHY: So, that sort of research is presumably useful to the parties in determining policy?

ROD CAMERON: Well, evaluating policy, Justin.

JUSTIN MURPHY: Evaluating it? Evaluating...

ROD CAMERON: Evaluating policy.

JUSTIN MURPHY: The current policy...

ROD CAMERON: Not setting. Evaluating - polls and research are used to evaluate the effects of policy rather than - it's very seldom in my experience used to actually set policy. I'll give you one example. I presented findings some years ago on capital gains tax, and to cut a long story short, I said, `Look, you're going to have a great deal of trouble in selling this policy to Australian voters', and the Prime Minister said, `Well, thanks very much for that advice. Now, your next job is to go out and tell us how we can sell this difficult policy - that's what your job is'.

TONY EGGLETON: We're not blowing in the wind of public opinion, but we do like to know the way the wind is blowing. We use research to help us communicate, to know what our strengths and weaknesses are in the community, what attitudes are. But we don't use polls to tell us what we believe in.

BOB HOGG: I mean, you may find negatives, but you don't respond by devising some opportunistic policy to it, because it just isn't in line with what the Government ought to be doing or the national interest.

JUSTIN MURPHY: At the time of the 1985 tax summit, however, some commentators felt the Hawke Government abandoned the reformist option (c), after private polling showed that a consumption tax would cost it office.

ROD CAMERON: Well, I read stories to that effect. You often don't believe stories about polls because often they're planted by people with a specific axe to grind. No, I won't comment on the private research we do. I never have and I never will, and I would suggest that listeners to your program should also treat some of the stories about polls with a bit of caution.

GARY MORGAN: The results of our polls have been used by politicians to change what they're saying or to reset the agenda or try to reset the agenda. Don Dunstan in South Australia changed his whole campaign on one State election, when a poll came out showing that he was going to get beaten.

JUSTIN MURPHY: What - one of yours?

GARY MORGAN: Yes. There's no doubt that the politicians have changed the way in which they handle issues when they read the polls.


Oh beautiful, beautiful Queensland Up where the wildflowers grow...

JUSTIN MURPHY: The pollsters' track record is pretty good. They get it right a lot of the time, except here. This is the pollsters' nightmare. Queensland is different. The zonal system, commonly referred to as the gerrymander, plus the presence of three major political parties, makes the result very, very hard to call.

PETER COALDRAKE: I think that in Queensland, one has to look at the three Party preferred vote and then to look according to where that seat is, as to what might happen and assess the contests very individually.

JUSTIN MURPHY: Here in Brisbane, the political grapevine says that the National Government will lose the referendum on electoral reform that it's called for October, and that it will go to the people on the old gerrymandered boundaries before the end of the year. In the uncertainty created by the Fitzgerald Commission, the pollsters' who got it so wrong in 1986, are finding Queensland just as hard to read this time, with surveys of the same electorate producing different results.

This is Morgan Gallup's graph of the fortunes of Queensland Labor over the past year. Newspoll tracked another path. The Brisbane Sunday Mail saw it this way, while ANOP chartered a steady improvement. Pollsters would point to the variables of time, place, question and sample size, but the gap between them is a huge 8 points, the difference between victory and obliteration on election day.

WAYNE SWAN: There is a very distinct possibility of a change of government in Queensland at the next State election. That is not reflected in the Morgan Poll. That creates a disadvantage for the Labor Party and, as a result of that, we must spend a great deal of time and effort in talking to people in the community and letting the media know that we're going a lot better than the Morgan Poll indicates.

JUSTIN MURPHY: Labor's campaign is already up and running.

EXTRACT (Advertisement):

UNIDENTIFIED: The one thing you don't need is a handicap, like a coalition. Queensland deserves a leader with both feet on the ground, not a coalition with three feet in the air.

JUSTIN MURPHY: The Liberals too, are confident.

DAVID FRASER: We actually began polling for the '89 election way back in 1987, so it's an ongoing thing. You can't do a poll four months before an election and hope to get all the answers. You've got to look at trends over a long period of time.

We're not saying at this stage that we'll have enough members to form a Liberal government, but we'll certainly be in government.

JUSTIN MURPHY: But for both Liberal and Labor, the battle ground is not only here in the mortgage belt marginals; they have to fight the electoral system as well.

PETER COALDRAKE: I mean, one should be very careful about counting the Nationals out in Queensland, from previous experience. It's quite possible that the next State election for the Liberal Party to out poll the National Party by 5 or 7 per cent, yet I would imagine the Nationals will probably have twice as many seats as the Liberal Party and that's simply because of the electoral structure we have here. On the basis of present evidence, I think that there'll be a non-Labor majority after the next election. But I would say that in the case of this year's election, one has to be closer to the mark because there is going to be so much happening in Queensland over the next three months, that we just can't say exactly where the Nationals will be.

JUSTIN MURPHY: While many criticise the opinion pollsters, especially when they get it wrong, few question their vehicle - the media, for the pollsters can argue that they're merely taking a snapshot, and the parties argue that they're just staying in touch with the real world. But with few exceptions, the media reduces the complexities of scientific sampling to sensational headlines, superficial comment and the satisfying simplicity of `swing'.

ERNIE CHAPLES: Much of what we see in the newspapers nowadays is pretty quick and dirty, relatively sloppy polling. I think we have to be very careful of them because I think there is a tendency to have them used by the media in a precise sort of manor without reporting accurately the methodologies or without very often the reporters who are reporting on the polls, even understanding the methodologies.

TONY EGGLETON: We had an example back in 1980, in fact, in the Fraser campaign in 1980, where I almost felt there were two elections being fought. There was the real election which I though I was involved in, and there was the other layer of campaigning which was the day to day almost of research results which the media became very preoccupied with, so much so, in fact, that there was a mood throughout the campaign that the Liberals were certain losers and that Mr Hayden and the Labor Party were certain winners. And it was a very strange atmosphere throughout that campaign because what I felt were the important issues and the focus we were trying to put on the campaign, was almost daily being distorted by the Press focusing rather on the opinion polls rather than politics.

WAYNE SWAN: Well, I think it's pretty unhealthy the way we've gone with opinion polling in Australia and the way the results are misinterpreted in the media, the way they are sensationalised and the way they are therefore read incorrectly. There's no doubt that the inaccurate reading of polls and inaccurate polling, and sometimes both those things combined, produces advantage for one side of politics over the other.

JUSTIN MURPHY: Reporters may be literate but they're seldom numerate. Foolishly perhaps, they trust the pollsters to get the numbers right, and for all their criticism, the new numbers men in the political parties, the handful of backroom boys who know which way the wind is blowing, reveal the results of their secret research only when it suits them.

TONY EGGLETON: And I believe myself, that the electorate as a whole has become quite blase about polls during elections. I think we in politics like to say the only poll that really counts is the one on election day and, I think, in a sense, a lot of people are coming to that conclusion too.

BOB HOGG: I mean, I don't believe and never have believed that the electorate is sort of like some yo-yo that goes up and down and fluctuates or at the rate at which you will see on a Morgan poll or a Saulwick poll for that matter; it just doesn't function like that. The electorate doesn't operate like that.

CHRIS WALEY: The rights and wrongs of opinion polls, our cover story this Sunday, the reporter Justin Murphy.