KEY POINTS:
There is only one poll that matters, the one on election day. So say many politicians when opinion poll results are less than favourable in the lead-up to an election. They are right, to a point, although they ignore the effect that the publication and broadcast of snapshot polls may have on actual outcomes on election day.
This year a new refrain came from those behind in the polls: public opinion polling was no longer relevant because in the age of the mobile phone, fewer young or poor voters had telephone landlines to receive the pollsters' calls.
This myth gained remarkable currency among the left elite, keen to undo perceptions of a mood for a change of government.
The outgoing Prime Minister and Labour's president, one a political scientist and the other a professional market researcher, joined this chorus. Last Saturday the only poll that matters disproved their theory emphatically.
The election result closely resembled the final standings of three of five major national polls that regularly charted the public mood.
Give or take a percentage point, the two major parties were correctly assessed and, with the honourable exception of our own Herald-DigiPoll survey, only the Greens' underperformance on the night was missed. All five polling companies read the public's willingness to vote for National, Labour, the Greens and New Zealand First in the correct order. The Herald's poll of polls, an average calculated for us by columnist Colin James, was almost exact in reflecting the outcome of the left and right blocs.
So the published polls were not deficient nor were they "rogue", as the cliche would have it. They reflected a striking and lasting trend that manifested itself in the election-night counts from the ballot box.
The truth is that polling conducted by Labour and others during the campaign would, if honest, have been telling that party the same thing.
The tactic of politicians and pundits publicly undermining political polling was as hollow as it was desperate. As it turns out, New Zealanders were well served by the pollsters and the media organisations that commission and publicise them.
Public opinion surveys are more science than art. Their methods factor in potential lower response rates due to landline use and the like, with many more calls made to compensate. Low representations of voters for small third parties are weighted to ensure integrity of the overall result. Moreover, the polls do not pretend to be anything but the snapshots in time that they are.
In the United States election, polling companies also had a good year. The Real Clear Politics average margin in major polls of about a 7 per cent lead for Barack Obama over John McCain revealed days before the vote was almost spot-on to the actual outcome of a 6.5 per cent margin. Polls in many key individual states also precisely caught the public mood for President-elect Obama. Attempts among conservative pundits to throw doubt upon the poll results were futile. Theories of people conning the pollsters by intimating support for the black senator while being unable, for reason of colour, to actually vote for him proved wrong.
The enduring value of snapshot opinion polls is that they give voters a voice, perhaps their only one, between elections.
They are the equivalent of television ratings and can have an equally devastating effect on a policy - remember Closing the Gaps, for example, which was withdrawn after damaging public reaction - as viewers tuning away from an unfavoured series.
Rightly, our politics takes its lead from public thinking on some matters but challenges public opinion to lead in other areas. Where parties choose to ignore that balance, deny trends or attack the messenger, they do so at their own risk.
