With pollster and electorate volatility apparently running rampant over the last few weeks, it’s worth taking a closer look at exactly what went on.
If we chart the first differences of each pollster since August – the amount that the two party preferred estimate has changed since the previous poll of that pollster – we get a good handle on the size and nature of the volatility. So if Newspoll came in at 57 up from 55 the fortnight before – that poll entry would have a value of +2. Similarly, if a Morgan face to face poll came in at 57 down from 60 in the previous Morgan face to face poll, that poll entry would have a value of -3. We’ll use the ALP for the two party preferred.
The polls were meandering along through August, through September and throughout October undergoing small changes within 3% each way. Suddenly at the very end of October, the polls all picked up significant movement.
Newspoll seemed to overshoot on the magnitude, but was the first poll off the rank to show a large, negative change for the ALP – coming in with the now infamous 7 point drop for the government. The other pollsters eventually followed, although it took some a little while to capture the zeitgeist, with Nielsen yawning at the whole affair. We can see the actual numbers involved if we table the behaviour since the beginning of November:
Interestingly enough, while the volatility has been pretty high, the actual polling trends tell a slightly more boring and probably realistic story story – the period where asylum seekers have dominated the media cycle has seen 2 points knocked off the government’s two party preferred on both the all pollster and phone pollster trend measures.
Boring is good because it’s more often than not, right.
9 Comments
That’s how I see it also.
When newscycle noise and statistical noise get washed out of the results, there’s not been a lot happening.
ALP at 44-45 L-NP at 37-38 and the TPP’s at 56/44.
The kicker is the best poll for the Libs in a long while – and they would still have gone backwards from the election. Now that normality has been restored – they are back as they were before the pigs started warming up their wings.
David Richards @ 2 has it. Over 100 seats on both phone poll and face to face. While I don’t believe those numbers will stand up at the actual polls, you’d have to think that Rudd would be thinking ‘Consolidate position’ not “Retrieve disaster’.
Given the polling around the refugee stuff, I’d have to say he probably called that one with his usual uncanny feel. Most of the criticism was that he was too soft, and further movement to the right probably wouldn’t have helped him as much as the greens.
All the same Poss, do I detect a touch of inertia amongst the people being polled? Christmas is coming, lists have to be made out, plans to be fixed. The latest refugee business is being thrashed up again, Auntie Norma and Gran have decided to go to Tuvalu for Christmas, meaning I will have to pay for all of this and we’ll be without anyone to look after the kids over Christmas, and if all this wasn’t enough I’ve got to endure Kevin Rudd and that bloke who is head of the Opposition, every time we switch on the TV.
Just a thought!
Question:
If the other pollsters are now shoing 4 or 5 point drops, doesnt that mean Newspoll’s 7-point drop was within the margin of error?
Ergo, it wasn’t an outlier?
If you factor in a 2% real decrease in the Pollytrend (and assume that Newspoll picked this up earlier than anyone else), plus a MOE of 2.9%, plus a built-in bias of 1% against Labor and an assumption that the earlier 58, 58 59 polls were partly over-estimating, then the answer to your question is “Yes”.
Similarly, the 58 and 59 polls MAY have been LOW by their MsOE. MOE is +/-, therefore either extreme is equally valid. Yes, if you max out the MOE on the “shock! horror!” Newspoll you get down to the 56/44, or you could max out the MOE of the others to get 52/48. There is some overlap if you assume extremes.
It is just as valid to assume that the later polls are too high by the MOE, and that would still leave the earlier Newspoll as the dark-skinned person in the woodpile.
If the earlier Newspoll WAS massaged for political reasons by Newscorp, they have been very careful to do so in such a way that they can now hide behind the MOE.
My view is there probably was a drop in ALP support (2% ish), but it was overcooked (either intentionally or unintentionally), and this produced a slightly exaggerated drop in the other polls in a bandwagon effect.
In any case, if the best the Libs can do is 52/48… they better get used to opposition. Maladroit isn’t going to be PM.. at least not this turn.
He could always do a Howard and get another go at Leader in 2016, and who knows.. he might have a chance at winning then, once the ultra right deadwood like Tuckey and Minchin are gone.
Hmm, five of the last six polls are either equal or below the Pollytrend line, so I’d be hesistant to call the recent drop 2% yet. Given the way Pollytrend works, if the next round of polls are roughly stable at 56 ALP TPP the value of the trend ‘now’ will actually drop (correct me if I’m wrong here Possum!).
David Richards at number 7:
The trouble for Malcontent, and any other aspiring Liberal leader, is that there will always be a steady stream of right wing nutjobs bubbling to the surface to cruel their dreams of becoming PM. Mainstream Australian endorsement of right-wing zealotry died with Howard, at least for the time being.
http://blogs.crikey.com.au/pollytics/2009/11/17/the-polling-isnt-as-exciting-as-it-looks/comment-page-1/#comment-16028