In Crikey today, Richard Farmer asked us to imagine how we would score this election if polls didn’t exist. It’s an interesting thought experiment, though it’s also salient to mention that a lot of what’s set up the terms and framing of the contest has been poll driven. Perhaps that’s why, in Peter Hartcher’s words:
…there is no transformative reform, and no ambitious national agenda. A modest set of offerings for a time of modest ambition.
Of that, more later perhaps.
But, in the meantime, on top of the weekend’s Galaxy Poll we have one that came out last night, the first taken during the campaign proper. The sample is the same – and it shows the parties tied at 50-50, down from 52-48 to Labor. Newspoll is also out today, showing a substantial Labor lead, with a 2PP of 55-45, produced primarily by an uptick in The Greens’ vote and a fall in the Coalition primary, with Labor steady at 42.
Numbers on both at The Poll Bludger.
Graham Young thinks this poll, like the earlier Nielsen poll purporting to show Gillard Labor with a primary of 47%, is quite possibly an outlier. He reminds us that the margin of error could mean (at the other end of the spectrum) that the result is 52-48 (same as the last Newspoll). I’d add that a drop in the Coalition’s vote and a rise in The Greens’ vote might not necessarily translate into a higher preference flow to the ALP (as Newspoll’s methodology indicates), and also suggests sampling error to me.
Newspoll wasn’t taken entirely after the campaign announcement.
And then we’ve got Essential Research, with a rolling sample, showing 55-45 in Labor’s favour, and pretty much the same results as the last one.
People will, of course, take their pick. My inclination would be to regard a slight Labor advantage as the campaign baseline. I suspect in a campaign like this one, where both parties are running against Kevin Rudd to some degree, there’s likely to be a fair bit of volatility, and that the campaign itself probably matters much more than it sometimes does.
Another fact to throw into the mix is that the average decline in the vote of a first-term government since WW2 is 1.7%. And all have gone backwards on the first outing. It’s worth remembering that 2007 was close, and that there’s not a huge amount of excess in Labor’s parliamentary majority. At one point, assumptions were being made that a second election win would result in a pattern similar to that of many state Labor governments back in the day – a landslide after an initial narrow win. The 55-45 figures would still produce that landslide, but I don’t think anyone really believes that’s a likely outcome.
Given the distribution of seats, and the likely distribution of votes, a swing of 1.7% against Labor, the historical average for governments in their first term, would produce a House composed of 75 Labor MPs, 72 Coalition MPs and 3 Independents. That’s on a uniform swing, and not taking into account seats that might fall to The Greens.
So, when interpreting the polls, it’s worth remembering that Labor could get 51% of the two party vote and lose its parliamentary majority.