Dialogue: No harm in taking public pulse before an election

Introducing restrictions on political public opinion polls would lead to a less knowledgeable and less responsive electorate, says RAYMOND MILLER*.

Winston Peters has proposed an embargo on the publication of public opinion polls in the 28 days before an election. Legislation to that effect is before Parliament.

A number of countries ban campaign polls, including France (seven days), Portugal (14 days) and Turkey (28 days). Their example notwithstanding, the banning of polls has troubling implications for the goal of an informed and responsive electorate, as well as for the public's right to know.

Although the evidence is inconclusive, instinct tells us that Mr Peters has a point in claiming that polls influence the way people vote. The dramatic surge in support for New Zealand First following Mr Peters' 1996 anti-immigration stand, rising from 6 per cent to 29 per cent in little more than three months, seems to illustrate the underdog effect that can be induced by polls.

During the 1999 campaign, the Greens enjoyed a similar if less spectacular rise in the wake of National's intemperate attack on two prominent candidates, Nandor Tanczos and the Coromandel's Jeanette Fitzsimons.

But Mr Peters is on more shaky ground in claiming that the media deliberately presents poll results in ways that benefit some individuals and parties and harm others.

Few would dispute the media's role in transforming the modern campaign into a horse race, with personal style and ranking gaining an importance hardly contemplated 30 years ago. In contrast, substantive issues such as doctrine and policy are frequently overlooked.

However undesirable this might be, there is nothing in the media's campaign coverage to suggest that its treatment of the polls lacks objectivity or is intentionally unfair.

What, then, are the main points of disagreement with Mr Peters' proposed ban?

An important function of public opinion polls is to remind us of the basic rules of MMP. Much in the manner of cramming for a school examination, we need before each election to undertake a quick revision course in proportional representation and the functions of the two votes. Public opinion polls fulfil the helpful role of allowing us to distinguish the party vote from the vote for our preferred electorate candidate.

With the assistance of computer-generated graphics, they also give us insight into the likely makeup of the next coalition, as well as the distribution of the 120 parliamentary seats.

But campaign polls also help to define the public mood and to track any last-minute fluctuations in voter loyalty and support. Given the lateness with which voters now decide, with close to half of all voters making up their minds during the campaign, it is more important than ever that the information upon which decisions are made is both objective and up to date.

No group is more dependent on opinion polls than tactical voters. In a tight contest, voters may be directed as much by the head as by the heart. Close to an election, for instance, voters need to know if their preferred party is within striking distance of the 5 per cent threshold. After the experience of the former first-past-the-post electoral system, who among us would knowingly waste our vote?

When, in 1996, campaign polls revealed that Act had fallen below the threshold, National's Wellington Central voters were left in no doubt that it was in their party's interest that Richard Prebble win the seat.

Similarly, in the event that Act falls below the threshold on the eve of next year's election, National's Tamaki voters will come under pressure to give their electorate vote to the Act candidate and deputy leader, Ken Shirley.

As well as informing voters, campaign polls also help to satisfy the public's right to know. Several countries, including Belgium and the United States, have either resisted or overturned moves to ban polls on the grounds that they unduly restrict free speech. In the US, the right to conduct polls even extends to the day of the election, when voters can be surveyed as they exit the polling booth.

In Canada, the Supreme Court used the right-to-speak and right-to-know arguments to strike down that country's three-day ban. Just before last year's election, the Canadian Parliament amended the Canada Elections Act with a view to shortening the ban to a mere 24 hours (the same as in New Zealand).

Because the New Zealand ban will exclude private polls, such as those commissioned by political parties, under the proposed legislation the right-to-know doctrine will be unfairly applied. For example, whereas party bosses, strategists and even spin doctors will have access to data showing what the public think, the public will not.

And while the large (National and Labour) and well-funded (Act) parties have the resources to commission their own private polls, the small and cash-strapped parties, presumably including NZ First, will be deprived of the sort of data upon which to contest the campaign.

This raises the question of the viability of such a ban. How secure is the internet? And what can be done if politicians decide to spread rumours or leak information in the hope of being able to influence the outcome of the campaign? What is to prevent a television interviewee, for example, from transmitting coded messages through some well-chosen words or a supremely confident smile?

In the absence of polls, media commentators will be under pressure to draw on less reliable sources, such as talkback radio, footpath interviews or a casual survey of opinion. Freed from the constraint of evidence, the politicians will be even more inclined to claim that their party alone represents the will of the majority.

This leads, finally, to the question of how a public opinion poll is to be defined. Will the ban cover phone-in polls, such as those conducted on Television New Zealand's Holmes show, or the three-yearly survey of student opinion at Tauranga Boys College?

If not, the statistically reliable data provided by professional agencies will have been replaced by the unproven, if well-intentioned, findings of amateurs.

If used badly, the ubiquitous opinion poll can be instrumental in reducing the democratic process to little more than a media-driven sporting contest. But does the potential for abuse warrant a ban, and for the entire duration of the campaign? The weight of evidence suggests not.

* Dr Raymond Miller is a senior lecturer in political studies at the University of Auckland.