Mumble

In the Oz, Mr Shanahan here (tables at bottom). They show big swings against Labor – except in the one seat with an around average Green vote, Page (NSW). (The rest have low Green support.)

We don’t know sample sizes. [Update: tables now at Newspoll, 600 in each of the NSW seats, and 600 across the three Queensland ones.]

They surveyed Lindsay on the same weekend as the Penrith by-election, which very likely boosted Liberal support at the expense of Labor.

That leaves the three Queensland seats of Dawson, Flynn and Longman which, assuming a good overall sample [update: 600, just large enough to escape a Mickey and about a 4 percent MoE], are bad news for the Rudd government with a six percent 2pp swing from 2007. Labor is doing very poorly overall in that state and WA overall, and as Richo said on Q&A last night (possibly because he has heard an RSPT compromise is coming soon) the government needs to get the RSPT out of the way.

Labor doing well in Page assumes Green preference flows similar to the last election. This does again make one wonder about those. (Despite Q&A, Greens how to vote cards don’t have a great deal of influence on preference flows. Well, they have some, but not as much as most seem to think.)

We really don’t know. Possum’s analysis of Nielsen suggests the preference flow will be more even, but that’s just from one small set of numbers (and ignored, as you must, the rather fluctuating number of Green supporters who didn’t nominate a major party).

It’s the only data we have, so we should probably expect a smaller flow to Labor than 80 percent.

But recall that back in 2004, Newspoll and Morgan got very good (ie very close to actual result) final week primary votes but not good two party preferred ones because they asked non major party supporters who would get their preferences. It was after this that they began estimating themselves based on the previous election, which gave them both very good final week 2007 results.

And who are the “others” in 2010? If Family First, One Nation etc, then Coalition-favouring. But we don’t know this either.

And taking a step back, if Labor does get 52 percent after preferences across the country, as the weekend’s Newspoll had, they would almost certainly win. Coupled with today’s results (if we take literally) it means they would do badly in traditional marginals but better elsewhere.

This is an interesting election.

A reasonably happy result for Rudd. From primary support of 35, 40 and 15.

Main tables here. Also perceived standard of living here.

In defence of two party preferred

Some people say: forget the estimated opinion poll two party preferred support, Labor can’t win with 35 percent primary vote. They have a point: it’s hard to imagine a Labor victory with such a vote.

But it is even harder to imagine the Coalition winning with 40 percent if Labor gets 35 and the Greens 15.

People are also a bit right when they reckon estimating 2pp from such low major party primaries is a tad unreliable.

But in the end, respondents who tell pollsters they support a minor party/independent are either going to move back to a major party or give their preferences to them. We need some idea of two party preferred. In each seat, it’s what wins. 

If the pollsters didn’t give us 2pps we’d have to do it ourselves from the primaries.

The worst argument against 2pp is that apparatchiks in the PM’s office or somewhere don’t believe in them, that when they phone thay say “mate, just give me the primary votes and who has best haircut/nicest personality etc rating.”

We are supposed to defer to these institutionalised, spotty youths who probably had political strategy drummed into them by someone like Wayne “keep it simple mate and stay on message – never mind the personality” Swan?

I think not! I’ll remain a 2pp man, thanks.

Update: Peter (not me) in comments makes the point that if you assume the Green flow to Labor is as low as 2/3 (instead of 4/5 as at last election), and that “others” slightly favour the Coalition as they did in 2007, you end up with the opposition a little ahead on 2pp – say 51 to 49.

But these are of course just assumptions as well.

It would be great if someone did a huge survey of current Green voters and asked them which party came from and where their preferences are likely to go.

Assuming still awake ….

Penrith by-election today. Main question is whether two party preferred swing to Liberals begins with  ‘1′, a ‘2′ or even - you never know - a ‘3′. [Update: it's about 25. William Bowe reckons this "certainly doesn’t bode well for [Labor] in [federal seat] Lindsay”, but the two things don’t have much to do with each other. Under John Howard the federal Libs won Lindsay four times while state Libs were flogged in Penrith.]

When Kevin Rudd was riding high, bashings of NSW Labor at by-elections weren’t seen as some kind of test for him. They are today, it’s silly, but that’s the way it works.

The Oz reckons Julia Gillard may be leader this time next week. Are there really so many fools in Caucus?

It all seems to hinge on a bunch of about 1200 Australians of voting age who happen to be at home when Newspoll phones them this weekend.

One single survey will tell us bugger all about how the election’s gonna go, but will mightily influence reporting and hence the behaviour of the ALP.

If you get through this one Kevin, perhaps that earlier election is worth considering.