Galaxy marginals polling

Polls from 11 seats across three states paint a broad picture of Labor losing office on the back of swings averaging 4.6%. Also, a head-to-head analysis of various pollsters and poll methods throughout the campaign.

GhostWhoVotes reports Galaxy has electorate-level polls from 10 marginal seats in New South Wales and Victoria, collectively painting a grim picture for Labor. There is also a Newspoll survey of 502 respondents showing Peter Beattie trailing by 54-46 in Forde (a swing to the Liberal National Party of about 2%), from primary votes of 38% for Labor (37.4% at the 2010 election), 48% for the Coalition (44.1%) and 5% for the Greens (12.2%). The Galaxy poll has apparently targeted 550 to 600 respondents per electorate for a margin of error of 4%. (UPDATE: These turn out to be automated polls, and not live interviewer polls like Galaxy normally does.)

In the Sydney seats:

Lindsay. A 54-46 lead to the Liberals, a swing of about 5%.

Banks. A 52-48 lead to the Liberals, a swing of 3.5%.

Werriwa. A 52-48 lead to the Liberals, a swing of 9%.

Reid. A 53-47 lead to the Liberals, a swing of 6%.

Greenway. A 51-49 lead to the Liberals, a swing of 2%.

Parramatta. 50-50, a swing of about 5.5%.

Barton. Labor ahead 52-48, a swing of 5%.

In Victoria:

La Trobe. The Liberals lead 51-49, a swing of 3%, from primary votes of 36% for Labor and 45% for the Liberals.

Corangamite. The Liberals lead 56-44, a swing of slightly over 6%.

Chisholm. Labor leads 52-48, a Liberal swing of 4%.

Today also brought a Lonergan automated poll of Kevin Rudd’s seat of Griffith, which remarkably showed him trailing Liberal National Party candidate Bill Glasson 52-48 on two-party preferred, from primary votes of 38% for Rudd (down six on 2010), 47% for Glasson (up 11% on the LNP vote in 2013) and 11% for the Greens (down four). Either in anticipation of or in reaction to the inevitably skeptical response, the company’s principal Chris Lonergan penned a rather informative piece on methodology for The Guardian.

UPDATE: The bit below has been amended to account for the fact that the Galaxy polls were automated, which means there is actually very little basis for comparing automated and live interview electorate polls.

Which makes this a timely juncture to consider how polling of various kinds has performed during the campaign. The table below shows the number of polls conducted for each pollster and poll method together with the average Labor swing, at both electorate and national level. There follows, for your convenience, basic results for every electorate-level poll of the campaign barring a small number which did not involve Labor-versus-Coalition contest, together with the swings – not a single one of which is in Labor’s favour, emphasising the point that electorate-level has been much worse for Labor than national polling. However, since nearly all of this polling has mostly been of the automated phone variety, the question arises of whether this it to do with polling method, the particular challenges of electorate-level polling, or the peculiarities of the electorates being polled.

The only live interview electorate polls featured have been two from Newspoll, which makes their performance relative to automated phone polls hard to compare. However, there is a very large gap of 4.3% between national and electorate polls for automated pollsters. Non-phone methods, which have only been employed at national level, appear to have been more favourable for Labor, although there haven’t been very many of them (note that the two-party result being used from Morgan is the previous-election measure).

ELECTORATE POLLS					#	Swing
Galaxy							10	5.1
ReachTEL						8	7.25
JWS Research						8	6
Lonergan						3	11.3
AUTOMATED						29	6.6
Newspoll (live interviewer)				2	4.5
TOTAL							31	6.5
					
NATIONAL POLLS					        #	Swing
Newspoll						2	3
Nielsen							1	2
Galaxy							2	1.5
LIVE INTERVIEW						6	2.2
ReachTEL						2	2.5
Lonergan						1	2
AUTOMATED						3	2.3
Essential						2	0
AMR Research						1	0
ONLINE							3	0
Morgan Multi-Mode					2	1.75
TOTAL							13	1.7

ELECTORATE POLL RESULTS				Sample	ALP	Swing
Griffith	Lonergan	21/08/2013	958	48	10
Werriwa		Galaxy		20/08/2013	575	48	9
Reid		Galaxy		20/08/2013	575	47	6
Parramatta	Galaxy		20/08/2013	575	50	4
Lindsay		Galaxy		20/08/2013	575	46	5
La Trobe	Galaxy		20/08/2013	575	49	3
Greenway	Galaxy		20/08/2013	575	49	2
Forde		Newspoll	20/08/2013	502	46	2
Corangamite	Galaxy		20/08/2013	575	44	6
Chisholm	Galaxy		20/08/2013	575	48	8
Barton		Galaxy		20/08/2013	575	52	5
Banks		Galaxy		20/08/2013	575	48	3
McMahon		ReachTEL	15/08/2013	631	47	11
Macquarie	JWS Research	15/08/2013	710	45	4
Lindsay		JWS Research	15/08/2013	578	39	12
Kingsford Smith	ReachTEL	15/08/2013	610	48	7
Greenway	JWS Research	15/08/2013	570	51	0
Forde		Lonergan	15/08/2013	568	40	9
Forde		JWS Research	15/08/2013	1160	40	9
Deakin		ReachTEL	15/08/2013	619	47	4
Corangamite	ReachTEL	15/08/2013	633	44	7
Corangamite	JWS Research	15/08/2013	587	47	4
Brisbane	JWS Research	15/08/2013	607	46	3
Blaxland	ReachTEL	15/08/2013	636	52	10
Bennelong	ReachTEL	15/08/2013	631	35	12
Banks		JWS Research	15/08/2013	542	47	4
Aston		JWS Research	15/08/2013	577	37	12
Lindsay		Lonergan	14/08/2013	1038	36	15
Dobell/RobertsonNewspoll	11/08/2013	505	46	7
Forde		ReachTEL	08/08/2013	725	46	3
Griffith	ReachTEL	05/08/2013	702	46	4

Author: William Bowe

William Bowe is a Perth-based election analyst and occasional teacher of political science. His blog, The Poll Bludger, has existed in one form or another since 2004, and is one of the most heavily trafficked websites on Australian politics.

952 comments on “Galaxy marginals polling”

Comments Page 1 of 20
1 2 … 20

  12. Well, from a Labor point of view – Barton can be retained, Parramatta is winnable, and I would not give up on La Trobe either.
    Losing Werriwa would be a body blow BUT Laurie Ferguson is not a Rudd supporter so I wonder how hard he is campaigning this time?
    Too much of the Daily Telegraph and 2GB blaring anti-Rudd propoganda across the West of Sydney.

  13. ReachTEL scrubbing up much better in comparisons with these Galaxy/Newspoll findings than two certain other robopollsters.

  14. [Losing Werriwa would be a body blow BUT Laurie Ferguson is not a Rudd supporter so I wonder how hard he is campaigning this time?]

    Yeah, MPs like to lose their own seats to spite leaders…

  15. GhostWhoVotes ‏@GhostWhoVotes 51s

    #Galaxy Poll Seat of Reid 2 Party Preferred: ALP 47 LIB 53 #ausvotes

    GhostWhoVotes ‏@GhostWhoVotes 22s

    #Galaxy Poll Seat of Greenway 2 Party Preferred: ALP 49 LIB 51 #ausvotes

  16. Labor holding up better in Victoria than elsewhere obviously.
    As for Forde, if they prefer a do nothing MP like Bert whatshisname, they are welcome to the atypical boring LNP hack.

  20. Jamez Giaz ,new LNP Guru

    GhostWhoVotes ‏@GhostWhoVotes 6m
    #Galaxy Poll Seat of Greenway 2 Party Preferred: ALP 49 LIB 51 #ausvotes

  24. I think there is a danger for political tragics and bloggers to overestimate the importance of the NBN in the electorate. They use media and the internet more than average, so overestimate its importance to others. Plus the NBN rollout has been slow, so the benefit is not appreciated, because it has not been eperienced.

    This is not to say the NBN is bad. Economically it will be a good investment. But politically, the implementation has been Conroyed.

  28. GhostWhoVotes ‏@GhostWhoVotes 1m

    @genericleftist The 2PP is 49-51. Primaries are ALP 36 LIB 45, 556 voters polled in La Trobe. #Galaxy #ausvotes

  31. I dont buy the Forde figures.

    LNP and ALP both get a swing? GRNs vote down from 12.2 to 5? Post PNG solution?

    Complete rubbish.

  32. Gee all out tonight even the lovely? John of Melbourne make a guest appearance again plus of course direct from Menzies House Newie, what a misnomer of a name 😀

  33. So Newspoll’s analysis that NSW is Labor’s problem and Qld is not the saviour is corroborated. So the state Labor stench has gone into the carpet!

  37. Zoid- may be gains relative to previously worse positions but from a election winning point of view it seems that Labor needs to improve significantly in either QLD or NSW or both. Thus explaining the campaigning domination in that part of the country.

  38. @Paddy O/37

    Not saying they are worse positions (depending if you think a another minority gov would be bad).

    Of course it’s why It’s over in eastern states.

    But I’m not sure if that should be the case considering the SSM issue.

    Probably work more in Victoria and perhaps South Australia.

  41. liyana

    [Posted Thursday, August 22, 2013 at 9:14 pm | Permalink

    I think I’ll stick with Saul Eastlake’s estimate of a 30 billion black hole in coalition costings]

    Me too.

  42. These polls aren’t as bad for the ALP as the last set, based on a national average of 53-47 to the Coalition. Indeed 7/10 seats have a margin of less than 5%. If the ALP get it to 49/51 or 50/50 nationally those seats could easily swing back and even Forde is tighter than the last few polls. We also need to see more figures from Queensland, WA and NT where state poll samples are better for the ALP than NSW and Victoria where most of these figures are from

  43. Simon – Yes. These polls are focused on possible problem seats for Labor.

    Bugger all chance of a similar set of polls being commissioned for troublesome Coalition Seats.

  47. Simon – It’s been inculcated in Australian journos for decades that the way to the front page is to write a ‘Labor in Trouble’ story.

    ‘Management’ are commissioning the polls and deciding the questions to feed the fire.

    So a ’boutique’ targeted poll like this one is all part of the action – ‘Please the Boss’.

  48. CTar1 – Indeed, but it could have the advantage for the ALP they could catch the Coalition on the hop if there really is a swing to the ALP in Queensland as most polls suggest

Comments Page 1 of 20
1 2 … 20

Leave a Reply

Your email address will not be published. Required fields are marked *