Nine days ago The Weekend Australian ran the results of some remarkably comprehensive Newspoll work in the marginals, based on a sample of some 3,500 people.
Labor needs to hold all the electorates it currently has and take 16 seats from the Coalition to win.
Newspoll omitted Western Australia, Tasmania and the Northern Territory, but gave a best-case scenario for the Coalition of a loss of only seven seats and a worse-case where it lost 25.
Apply the best-case scenario across the country, the Coalition could maybe lose Bass and Solomon, but even gain a seat in the West and still remain in office. It’s not worthwhile taking the worse-case scenario further, since it would add at least another three seats to Labor’s swag.
So, two thirds of the way through this campaign, what can we say? The commentariat appears to have been stung by its calls in 2001 and 2004.
The Possum has written extensively here and on his own site about “the narrowing”. The idea that the gap between the parties may significantly change is predicted on the curious notion that the direction of the votes of those who say they are undecided are somehow likely to flow in ways different to the trends indicated by the polls.
Poll commentators such as Sol Lebovic, and even the pollsters themselves, have been suggesting for some months that there may be something that the published polls are missing, something the polls are not measuring that would justify a higher confidence in a Coalition win or clawback than a simple reading of the poll results suggest.
This is bizarre. It suggests the pollsters are disavowing the accuracy of their own polls – the very core of their businesses.
The statistical and methodological underpinnings of quality polling are solid and well known and the interpretation of the results is well-grounded. So long as the caveats arising from the statistical issues are accounted for we know that polls are excellent tools to show us trends and directions.
The votes of those who tell a pollster that they are yet to make up their minds are likely to fall in a similar pattern as the wider poll outcome. The longer and more stable the trend, the more likely it is an accurate reflection of what is going to happen in the only poll that counts.
There have been some remarkable past campaigning performances that have seen the polls clawed back from the commencement of a campaign to the actual election result by around 5% (Kim Beazley gave the most recent example of these), but there is nothing to indicate that such a clawback by the Coalition is now underway or is imminent.
Even were we about to see the narrowing, the problem for the Coalition is that the gap is somewhere between 8% and 12%.
Oppositions want to always be the underdog so Labor will stick to its line that the election will go right down to the wire.
But all races run the full length. Some are still won by a nose and others by three or more lengths. Twelve days out from the finish line, the bookies on this one are pointing to a strong win by a clear favourite.