KEY POINTS:
National Party leaders must be kicking themselves after last Friday's alarming - for them - opinion poll results that they didn't get a parliamentary majority in 2001 for Winston Peters' legislation outlawing the publication of poll results in the run-up to a general election.
The New Zealand First leader and National wanted a pre-election moratorium on publication for fear of last-minute "rogue polls" going unanswered. But a majority of Labour, Greens and Act voted for the status quo and the poll censorship plan was history.
The two anti-National poll results shouldn't have come as any surprise.
Apart from anything else, any party that has the misfortune to put up a former international money-trader as the saviour of our country's economy, just as the world's financial system collapses all around us, must have expected at least a ripple of discontent showing up.
But following month after month of double-digit leads over the incumbent Labour Government, the Roy Morgan poll must still have come as an unwelcome bucket of cold water to those who thought they could sleepwalk their way to victory.
The September 22-October 5 Morgan poll narrowed National's lead to just 3 per cent. That was a drop from 11 per cent a fortnight before in the same poll and from 21 per cent in June and July.
Amusingly, TV3's website ran a news story that afternoon saying this poll was "very much at odds with recent surveys and could be rogue". It sniffily noted that TV3's poll a week earlier "put National 16 points ahead".
That was timed at 3.07pm. Just three hours later, TV3 was trumpeting its own "rogue" poll. After 19 months of 3 News polls in which Labour had trailed National by 12 points, "that gap has halved to six, meaning it may not be exit stage left after all".
The report said that National leader John Key "cannot believe the result".
This disbelief would have been reinforced by Sunday's TV One Colmar Brunton Poll, which had National comfortably 18 per cent ahead, with Labour continuing a drift downwards from 37 per cent in August to 33 per cent between October 4-9 when the poll was taken.
To me it does stretch credulity to believe National's lacklustre campaign so far compared with Labour's, and the gathering worldwide financial crisis, wasn't reflected in the TV One poll.
Also in the back of the mind is Colmar Brunton's bad record at the 2005 election when it and Fairfax's ACNielsen poll both inaccurately predicted National would get a higher proportion of the party vote than Labour.
The actual result was Labour 41.1 per cent, National 39.1 per cent. Colmar Brunton had Labour at 38 per cent and National 44 per cent.
Instead of 2 per cent ahead, the TV One poll had Labour 6 per cent behind. The most accurate was 3 News TNS with Labour 40.5 per cent and National 38.7 per cent.
Morgan got the placing right, Labour 38.5 per cent, National 37 per cent, and was the most accurate predicting the small margin, but overestimated the total vote of the smaller parties.
For the record, the Herald-DigiPoll survey got the winner correct, but the margin, 44.6 per cent to 37.4 per cent, proved exuberantly generous to Labour.
Of course pollsters are sensitive to criticism. Colmar's Brian Turner said in the post-2005 inquest it never got reported that all but one of its figures had been within the 3.1 per cent margin of error, and invited me to do so.
Of course their problem in that poll was that their score for National had been 4.9 per cent over-generous, and that for Labour, 3.1 per cent too small.
As to the weekend poll, he says before that one, the various competing polls "were on the same hymn sheet and we haven't done anything different this time to what we were doing two weeks ago".
Using an advertising analogy, he says "for people to become attracted to your brand you need to say something to them and get some cut-through and I honestly cannot think what was being said that was cutting through with such substance to change the polls over that period". But he thinks this has now changed.
Helen Clark is now getting "the high ground on making sure that New Zealand's savings are safe" and "their launch was a little more polished". He thinks this may have "resonated with the voters" and show up in their next poll.
As for Murray Campbell at TV3 News poll, he happily points to the accuracy of his 2005 poll. While surprised by the smallness of the gap in the Morgan poll, he "couldn't fathom why there had been absolutely no change within the Colmar Brunton".
He said when he read Monday's Herald comment pieces about "how Key has responded, or not, over the last week's events, I thought there does need to be a bit of closing of the gap".
The race is on, not just for the politicians, but the credibility of the pollsters as well.
