tip off
17

The Polls of Election Day

A big wad of new polling comes in today as all the pollsters vie for bragging rights of being the most accurate – which is pretty funny in itself since the pollster that ultimately ends up being closest to the actual election result does so purely through chance and chance alone. But far from be it from us to try and add sense to that most insane of all political days. The juicy details of the final poll of each pollster comes in like this:

finalpolls1

Update: The Morgan at 51.5% is using the 2007 pref flow allocation – which is what I always use for Morgan phone poll here.

Unfortunately, no pollster pulled a 53% for the ALP, so our 2004 vs. 2010 poll comparison looks a little flacid at the end, lacking that last pollster spike:

2004v2010

On something a bit more serious, the final all-pollster chart for the election campaign sees the pollsters clustering between 50 and 52:

allpollsterseday

On our trend estimates, the final phone pollster trend comes in at 51.4 while the all pollster trend comes in at 51.5. A 51.4% ALP two party preferred would deliver around 79 seats with something approaching a uniform swing.

The final short and long term pollytrend charts now look like this:

pollytrendmediumeday

pollytrendlargeeday

The big “if” of the election is the undecided vote – which the final polls had coming in between a low of 2% for Newspoll and a high of 9% for Morgan Phone poll. Depending on which way that breaks, it could skew the two party preferred of each party by up to a percent or so. There was a lot of volatility between the pollsters at the state breakdown level – which we’ll go through in a bit – with Newspoll having a primary vote for Labor in Qld of 27%. That isnt a typo – it’s also not supported by any other pollster in any way, shape or form, so it’s not really worth taking too seriously.

Labor is still the favourite, the data suggests pretty strongly that they’ll have a small win – Qld and NSW will show Coalition gains, SA and Vic will likely produce Coalition losses and the West is being it’s erratic self like it usually is with seats open to change both ways. Ultimately, we’ll all know soon enough – although with the enormous volume of postal votes this election, “soon enough” may perhaps be a little longer than usual.

But more importantly, despite the cynicism many of us feel about both sides of the campaign, despite the leaders treating us like idiots and despite the general malaise in our political system – today’s election, like every election, is an important day that we should never fail to appreciate – not because of the quality of the choices we have on offer, but simply because we, unlike so many others, actually have a choice. As we said on the morning of the 2007 election:

So go and do your part for our democracy – man the booths, assist the booth workers or simply just exercise your franchise. This day is our day, your day, the day where that little piece of paper and its accompanying little pencil make all of us equal.

Next up – the election day election simulation. What does the combined polling tell us?

15

Please login below to comment, OR simply register here :



  • 1
    Randy
    Posted August 21, 2010 at 8:35 am | Permalink

    Thank you Possum for a great blog, very informative
    You are still sticking to an ALP victory?
    All the best

  • 2
    Ulysses
    Posted August 21, 2010 at 8:38 am | Permalink

    Thanks Poss. Great job. Good luck with your job hunt.
    A little less confident but hopeful for the ALP.

  • 3
    wal kolla
    Posted August 21, 2010 at 9:11 am | Permalink

    I think labor will get 51% after TPP, but due to the Mod. Areal Unit Problem will find itself short of of a majority.

    If I weigh the polls against their sampling size, I get ~51.3% (51.2504) which has labor between 75-76 seats nationally (74-75 with Melb going green).

    Im tipping 74 seats to a minor lab gov.

  • 4
    wal kolla
    Posted August 21, 2010 at 9:13 am | Permalink

    Btw, Im not factoring in the retiring MPs like you have done.

  • 5
    Ern Malleys cat
    Posted August 21, 2010 at 9:30 am | Permalink

    The undecideds could well be impressed by the new school halls they will be voting in.

  • 6
    David Stephens
    Posted August 21, 2010 at 9:55 am | Permalink

    Neilsen has this (clipped from Michelle Grattan’s report): “Almost two thirds of voters (64 per cent, up 5 points) expect Labor to win, while just 22 per cent (down 5) predict a Coalition victory. These figures are the same as on the eve of the 2007 election.” I recall that this question is a good antidote to 2PP; why don’t we hear more about it?

  • 7
    Wakefield
    Posted August 21, 2010 at 9:58 am | Permalink

    Thanks for your excellent analysis Possum. The latest Newspoll state results some pretty random swings – do the pollsters have ways of dealing with rogue results when they show up?

  • 8
    the spectator
    Posted August 21, 2010 at 10:01 am | Permalink

    Neilsen allocated the Green preferences on actual intentions roughly 86% and all the others use the 2007 election of 80%. Given the size of the green vote this could be a large factor in the overall result this year meaning the neilsen is potentially more accurate this time?

  • 9
    DodgyKnees
    Posted August 21, 2010 at 10:35 am | Permalink

    Dr. Possum, sincere thanks for all the regular injections of political reality.
    If the media favor integrity and accuracy over spin and blather then the job offers should flood in.
    Oops ….. didn’t mean to spoil your day !

  • 10
    David Blaazer
    Posted August 21, 2010 at 10:50 am | Permalink

    Is there any evidence one way or the other about whether preferred PM is a good predictor of how the undecideds decide once they get into the booth?

  • 11
    Katielou
    Posted August 21, 2010 at 10:51 am | Permalink

    Thanks again Possum.

  • 12
    Darn
    Posted August 21, 2010 at 11:26 am | Permalink

    Poss (or anyone else reading this if Possum is unavailable)

    I read on the PB site that according to information on this site the chances of Labor winning 73 seats is 89+%
    I’m trying to find the source of that information so that I can establish what are the chances of Labor having
    74, 75, 76 etc

    Any assistance would be appreciated.

  • 13
    Posted August 21, 2010 at 11:28 am | Permalink

    Darn – the latest post:

    http://blogs.crikey.com.au/pollytics/2010/08/21/election-day-mega-simulation/

    It’s 68.3% implied probability of the ALP winning at least 73 seats, under the assumption that the combined polling results are correct.

  • 14
    Darn
    Posted August 21, 2010 at 11:44 am | Permalink

    Thanks Poss – very informative.

  • 15
    Barking
    Posted August 23, 2010 at 3:28 pm | Permalink

    And the winner was? Essential, by the look of it?

2 Trackbacks

  1. ...] This post was mentioned on Twitter by Possum Comitatus and Carole Strong, Frank Calabrese. Frank Calabrese said: RT @Pollytics: On Pollytics: Election Day http://bit.ly/b0vvpF [...

  2. ...] such as pollytics opt to provide a well researched and alternative view.  ‘Possum’ streams his page with self [...

Please login below to comment, OR simply register here :



Womens Agenda

loading...

Smart Company

loading...

StartupSmart

loading...

Property Observer

loading...