Possums Pollytics

Politics, elections and piffle plinking

Newspolls for all!

Posted by Possum Comitatus on September 21, 2007

Over in The Oz, the latest Newspoll quarterly breakdown by both demographics and seat margins is out – both at the same time.

There is so much data here that rather than give a riding commentary, I thought that I’d just present the data first and do the analysis in subsequent posts. There are a few very interesting things that have already popped up out of the data – mostly reinforcing the view that the Coalition Firewall strategy has been playing out and they’ve been shoring up the base….. but we’ll get to that later.Suffice to say, the stuff floating around on this polling breakdown in the Oz and on SkyNews today is more twaddlish than usual – but we’ll round up those rogues later.For now, it’s on with the show.

Firstly, lets do the marginal seats, the safe government seats and the safe ALP seats:

 

margprimsq31.jpgmarginaltppq31.jpg

 

safegovprimsq31.jpgsafegovtppq31.jpg

 

safealpprimsq31.jpgsafealptppq31.jpg

These results give us the following swings from the 2004 Election:

margswingsprims1.jpg 2ppswingsmarginals1.jpg

Next, we’ll move on to the demographic and State breakdown. All the swings are the difference between the third quarter Newspoll results and the 2004 Election results. On the demographic breakdowns, the swing represents the difference between the third quarter Newspoll results and the best Newspoll estimate of the 2004 election results.

tsgp5states1.jpg

tsop5states1.jpg

stateprimswings1.jpg

 

2ppswingsstatescity1.jpg

 

demswingsprim1.jpg

 

Lots here to chew through, and chew through it we will.

As  something to ponder, the state swings point to around a 43 seat ALP gain.

You can plug these numbers into Antony Greens new spiffy calculator .

Although he only allows for swings of up to 10% ,which plays around with Victoria a bit.Ordinarily you’d think that sized swing would be enough!

Bryan over at OzPol also squeezes the data

 

 

 

AddThis Social Bookmark Button

20 Responses to “Newspolls for all!”

  1. John V K said

    Thank you for the work

  2. codger said

    Morgan 56.5 43.5; same old same old

  3. steve said

    They are not out there with baseball bats – they do not need to become thugs.

    They are out there with their sharpened pencils just waiting to receive their ballot papers and move into the cardboad booths.

    Then as electors we get to do our bit of the process.

    We get to say how our tax dollars are to be spent. We order the priorities. It is up to us.

    I find it hard to put into words the ecstasy I will feel when there is a new member for Bennelong and a new PM.

    Thanks Possum for the quality of the effort you put in. It is outstanding.

  4. Chris Mayer said

    You start playing a bit with Antony’s calculator, sepcifically the state-by-state break down, and it really starts to cement how this is a pretty big ask for the ALP. You play pessimistic (from the ALP’s perspective) and give them a couple of seats in WA, and boy it really starts to look like a rough mountain to climb. I mean a 6% swing in Qld will only give them three seats in that state.

  5. anthony said

    If you assume uniform statewide swings, Antony’s calculator shows 44 seats changing hands.

  6. anthony said

    (if you feed the numbers from the newspoll quarterly results into it, I mean)

  7. […] the best look at the polls and for good graphs the best value is at Possum Pollytics and also […]

  8. If you take the accepted view that Newspoll’s ALP 2PP is skewed 2.5 or so in the ALP’s favour and work that through Anthony Green’s calculator state by state you still get 91 ALP seats. Take another 2% off for the campaign and you get 79 ALP seats. Still a win if not all that comfortable.

    So, some room for manouver. That said, Kevin better study up on the tax system sharpish. I assume that will be his reading on the plane.

  9. Psephologists rejoice!

    Antony Green’s election site is up.
    By the way, if you were tempted to believe the Shanahan spin that “Key states, bush warm to Howard”, please be reassured that they’re still quite coolish. Possum has all the graphs you need. B…

  10. Craig McGarty said

    I was interested to look at the July August 2004 aggregation for Newspoll in comparison to this one.

    Polled primaries in the marginals were 44 L-NP and 42 ALP. The October election figures were 45 and 41 (which means that they also got the share for the major parties right).

    This time around those figures are 38 L-NP and 50 ALP (implying a 2PP of 57-58). If those figures were sustained at the election the Coalition would struggle to hold any of its 23 marginal seats (that’s not a prediction).

    In 2004 the big recovery for the Coalition primary votes was in their safe seats (47 in July-August aggregation up to 55.4 at the election).

    Possum: have you tried fitting the 2004 aggregated Newspoll data to the 2004 election results?

  11. Mercurius said

    Possum, you do things to numbers that should only be done by consenting adults in private.

    I don’t mind what you do in your own bedroom, but do you have to flaunt it like this?🙂

  12. Lord D said

    I believe that Newspoll underestimates the Grn vote, so there’s no need to adjust 2PP. All other polls have Grn vote much higher than Newspoll.

  13. Stig said

    Re: Robert Beswick #8: I think you may be mixing up Morgan with Newspoll. Bryan Palmer over at OzPolitics has done a lot on this in the past.

    Also, there is a frequently unchallenged assumption that Howard is a good campaigner, which doesn’t stack up with past evidence. With Sinodinos no longer wrangling Team Howard, I’d think their campaigning is going to be unsubtle and not particularly effective; so there is no solid rationale to assume they will make up 2% in the campaign.

  14. Stig #13.

    You’re right on both counts. I was just heavily into not counting chickens and pursued pessimistic/paranoid scenarios to see where they led us on Antonys calculator.

  15. Kramer said

    There was an article in the Canberra Times today stressing that the PM would not call the election this weekend because he was waiting for the next Newspoll. I would be very surprised if the next set of polls saw any further movement back to the Coalition.

    Labor’s primary vote has barely shifted from between 47-50% since March. The Liberal Party’s reaction to the recent 55-45 Newspoll was very telling and encapsulates the current political climate. They are well and truly behind and Howard is looking increasingly wizened by the day. He will not be able to campaign as strongly as he has done in the past. This is primarily due to the relentless campaign being waged in Bennelong. A disproportionate amount of his attention will be mired in this campaign. Rudd is not similarly encumbered because he is sitting on a healthy margin of well over 8% in Griffith.

  16. Stig: “… there is no solid rationale to assume they will make up 2% in the campaign.”

    Howard a good campaigner? Polls bound to tighten during the campaign? Mmm…

    Looking at Newspoll one month away from the five elections Howard has contested as leader, and comparing each against the respective election 2pp, gives an average change of … zero. (1987, +2; 1996, 0; 1998, +.5; 2001, -5.5; 2004, +3.) On average, Howard neither loses or gains during the campaign. Notwithstanding, if we are trying to find something positive, of his five elections he only went backwards in one, but it was a biggie of a 5.5 drop in 2001 (56.5 to 51). If we take that out, we might conclude that Howard, on his record, is a modest campaigner with the capacity to make up a couple of percentage points or so. It is hard to say anything more than that.

    I did a similar exercise some time ago looking at all elections, state and federal, in Australia since, I think, 1975 and concluded that incumbency is worth about a 4%, on average, gain during the campaign. (If I can find the reference I’ll post it.)

    Draw your own conclusions.

  17. haiku said

    I’m not sure that Howard will be “distracted by Bennelong”. He wants to be PM. No good winning his seat while losing the election. So he takes Rudd on nationally and tries to win back the voters who have gone to Rudd (via, as Possum has pointed out, the minors. So he may need to wean them off Rudd on to the minors and then back to him). If he manages this, then in all probability he wins Bennelong.

    The only scenario he regrets not campaigning harder in his own seat is if the Coalition hangs on and yet he loses his seat. Obviously that’s a fairly large downside for him, but his sycophants in the GG can work it into Teh Howard Legend, sacrificing his seat for the good of the party.

    No, for JWH it’s all or nothing: so play global, win local.

    But, as Possum has also already suggested, maybe it’s a case of shore up the base and don’t lose by too much …

  18. […] detail and context to the election analysis. For starters, Possum Comitatus has posted the latest quarterly breakdown of Newspoll data, and promises to follow up with some analyses in future posts. One quick thing that strikes me – […]

  19. Geoff Lambert said

    The One Nation effect on voting was quite interesting. While the PM was able to take over the niche it left vacated when it collapsed, the One Nation voters did not all remain in the nest. As Possum points out, nearly all the One Nation voters were drawn from Liberal voters, some of whom were presumably former ALP voters. ON did not attract ALP voters directly. When ON was at its peak, the preference flow to the ALP was a rather high (and consistent) 43%. When ON collapsed, a high proportion of these appear to have returned to the ALP fold (from whence they came originally via the Coalition), rather than back to Howard. Add that as another reason why Pearson’s analysis is off the mark.

  20. Charles said

    I took a different path, I used the ABC swing data as I don’t know what data Newspoll used for safe lib, marginal and safe labor. Being a Victorian I only did Victoria.

    I used the marginal swings and state swing as you did and the average swings of the sets as points on a line and used the line to work out seat swings.

    If I make the assumption you did, that is the swing imbalance is uniform across the country then I get a swing of 16.63 in Murray and 6.83 in Batman. The Liberals retain 3 seats. Indi Murray and Mallee.

    The trouble is we have lost some data. It may be a very hard Lib swing in NSW is swamping the other states. There is nothing in the data that I can see that lets us pull that out.

    However being a pissed off Liberal voter I am willing to accept it is liberal seats that are swinging hard and the results for Victoria are in the ball park.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
%d bloggers like this: