WA election: modelling the 2008 result

An attempt to make sense of where the two parties over- and under-performed at the last WA state election.

In a few idle moments, I’ve knocked together a linear regression model to fit the results of the 2008 election in the vague hope it might shed some insight on where Labor and Liberal over- and under-performed. I’ve limited this to metropolitan electorates, because the relationship between voting and the most potent explanatory variable – income – breaks down beyond the city limits. I’ve also excluded electorates where independents reached the final count in the preference distribution (i.e. Churchlands, Alfred Cove and Kwinana).

On my first run I simply looked at income, and two not unexpected factors stood out. Parties over-performed where they had a sitting member defending their seat, and Labor over-performed in the inner-city and Fremantle. So I added two extra variables to the model: the Greens primary vote, to serve as a proxy for inner city-ness, and another variable which assigned a value of -1 to seats with a Liberal incumbent, 1 to seats with a Labor incumbent, and 0 to seats which were vacant. From these variables, Labor’s two-party vote can be modelled as 0.681 -0.419a +0.781b +0.056c, where “a” is a measure of median family income, “b” is the Greens primary vote and “c” is the incumbency variable. This explains 86% of the variation in the two-party results.

The following table ranks seats according to how Labor performed relative to the expectation of the model, from highest to lowest. The columns show median family income from the 2011 census, actual Labor two-party preferred, Labor two-party preferred according to the model, and the candidate situation at this election. “Sophomore” refers to situations where a party has gone from not having a sitting member at the 2008 election to having one in 2013, hence excludes situations where the sitting member has changed as a result of by-elections. “Vacancy” refers to the opposite situation, meaning a retiring sitting member in four cases and a mid-term by-election loss in the fifth (Fremantle).

		MFY	2PP	MODEL	DIFF	2013 CANDIDATE
West Swan	1879	54.4%	47.7%	6.8%	ALP sophomore	
Perth		2536	57.8%	51.7%	6.1%    ALP incumbent
Warnbro		2008	59.7%	53.8%	5.9%    ALP incumbent
Nollamara	1450	62.7%	57.4%	5.4%    ALP sophomore
Maylands	1981	59.0%	54.1%	4.9%    ALP sophomore
Cockburn	1951	59.6%	55.2%	4.4%    ALP incumbent
Ocean Reef	2374	45.6%	41.6%	4.0%	LIB sophomore
Jandakot	2254	48.2%	44.4%	3.8%	LIB sophomore
Armadale	1406	64.8%	61.6%	3.2%    ALP incumbent
Mindarie	1796	58.5%	55.4%	3.1%    ALP incumbent
Willagee	1712	64.6%	62.3%	2.3%    ALP incumbent
Victoria Park	1875	59.0%	57.2%	1.8%    ALP incumbent
Scarborough	2319	44.8%	44.0%	0.8%	LIB sophomore
South Perth	2437	35.7%	35.1%	0.6%    LIB incumbent
Cannington	1521	59.0%	58.7%	0.4%	ALP sophomore
Darling Range	2028	44.4%	44.1%	0.3%    LIB incumbent
Hillarys	2331	38.6%	38.8%	-0.1%   LIB incumbent
Gosnells	1488	55.5%	56.0%	-0.5%	ALP sophomore
Mount Lawley	2079	47.8%	48.9%	-1.1%	LIB sophomore
Bateman		2258	38.6%	39.7%	-1.1%   LIB vacancy
Belmont		1713	56.7%	58.2%	-1.5%	ALP vacancy
Morley		1642	49.1%	50.9%	-1.7%	LIB sophomore
Rockingham	1474	60.6%	62.4%	-1.7%   ALP incumbent
Fremantle	2156	62.0%	63.8%	-1.7%	ALP vacancy
Wanneroo	2016	49.3%	51.1%	-1.8%	LIB sophomore
Forrestfield	1757	50.2%	52.2%	-2.0%	ALP sophomore
Southern River	2152	48.4%	50.3%	-2.0%	LIB sophomore
Riverton	2033	49.8%	52.2%	-2.4%	LIB sophomore
Bassendean	1604	60.3%	62.8%	-2.4%	ALP vacancy
Joondalup	1908	53.5%	56.1%	-2.6%   ALP incumbent
Midland		1655	58.3%	61.2%	-2.9%   ALP incumbent
Kalamunda	1871	43.7%	46.7%	-3.0%   LIB incumbent
Cottesloe	2871	30.6%	33.9%	-3.3%   LIB incumbent
Swan Hills	2011	46.5%	49.9%	-3.5%	LIB sophomore
Girrawheen	1384	61.5%	65.1%	-3.6%   ALP incumbent
Carine		2562	35.5%	39.5%	-4.0%	LIB sophomore
Balcatta	1789	52.3%	56.4%	-4.1%	ALP vacancy
Kingsley	2170	45.5%	50.8%	-5.3%	LIB sophomore

As a basis for analysis this is far from foolproof, as in every case the result would have been influenced by multiple factors outside the model. Nonetheless, it hopefully offers a few hints where parties do or don’t have room to improve. Labor’s six “best” results were achieved in seats where the Liberals did very little campaigning as they recognised them as safe for Labor, so it might then be thought notable that they are putting considerable effort into two of them (West Swan and Perth) this time. There’s no corresponding pattern at the other end of the table, although Labor did abandon Kingsley very early in the 2008 campaign after recognising it as a lost cause. The two seats which would have intuitively been expected to have been near the bottom of the table, Mount Lawley and especially Morley, are in fact not far from the middle – though I hesitate to read too much into that.

If anyone has any further ideas, I’m all ears.

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.

26 comments on “WA election: modelling the 2008 result”

  1. Thanks William that us great I didn’t previously understand why Labor thought themselves in with a chance in Swan Hills. I knew Alban is a hopeless member who turned his back on his own electorate without any protest and should have a negative sophomore vote – and I knew Radisich was running a real campaign but I had forgotten how poor Labors campaign was last time.

    Must still be a long shot but but factoring Frank and Ian’s personal factors plus your ‘missing’ 3.5% from last time one could hope for a third Radisich miracle in Swan Hills for Labor.

  2. I’d take out the Greens vote (it’s a rubbish variable) and look at adding census data. Perhaps the proportion of the adult population in each electorate with a university degree and the average age of the electorate.

  3. I agree the Greens vote is a rubbish variable, but it still hit the mark better than any census variable that I could think of. I’d be very interested to hear suggestions. I tried “no religion”, but there was too high a response in a lot of places I don’t think of as being all that trendy.

  4. I think Peter Brent (Mumble) did a Senate Vote – HOR Vote calculation to measure personal vote strength – maybe you could work that in.

  5. Hi William,

    Sorry, rubbish variable was probably a bit harsh. My only defense is that I was a little cranky yesterday.

    Have you tried the population density if each district? If the figure isn’t available it wouldn’t be hard to work out. The population and area of each district is usually listed on the WAEC website (been a while since I’ve checked, but I doubt it would have changed).

    As I said in my first post, the proportion of the population with a university education (bachelor degree or above) may be more appropriate. It usually correlates with Greens voting and density, and is likely more meaningful than both.

    If you have access to some survey data (although this is usually hard to get at the State level) you could also try doing some multilevel modelling, including both individual level data and district level predictors. However, this may be more complicated than its worth.

  6. Treasury costs Metronet at $4.3 billion, which is on balance good news for Labor as they said it would be at $3.8 billion whereas Buswell tried on $6.4 billion. Except, apparently, that the Treasury number is “not finalised”.

  7. I think the treasury figures are great for Labor – they should come out with a comment that estimates are not an exact science and that cost control in large projects is a challenge for govt and the private sector in big projects – but highlight over and over again that they will execute the project well and deliver. The costings should be used to talk about the project.

  8. “Meanwhile, Treasurer and Transport Minister Troy Buswell has admitted that more than 50 per cent of the capital cost of the Liberal Party’s big three election transport promises – the MAX light rail, the airport rail link and the Perth-Darwin Highway – will need to be funded by the Federal Government.”

    Like to be a fly on the wall when Albo gets that letter …

  9. It’s interesting to note that our local paper has moved away from the traditional election day endorsement of one particular party to its readers to an approach where they are running a 6 days per week continuous endorsement model. This obviously saves time at the end in trying to think of new and creative ways to tell readers why they should love all things Troy and Colin.

  10. The Sunday Times:

    – Editorialises for the Liberals (“(Barnett) has the skill, vision and determination to lead for another four years and The Sunday Times believes he should be given that opportunity”), while throwing more than a few bones to McGowan (who “has proven his leadership credentials”);

    – Reports Liberal polling shows Brendon Grylls’ primary vote in Pilbara “could be as high as 46 per cent”, and says the party is hopeful of knocking off Graham Jacobs in Eyre, which I thought looked a bridge too far after Labor preferenced the Liberals there;

    – Reports the key targets for the Liberals in the final week will be Forrestfield, Joondalup, Balcatta and Albany, with Kimberley, West Swan and Collie-Preston “the next tier of very winnable seats” (not quite sure what to make of Kimberley, which I thought likely to fall to the Nationals).

  11. Given how terrible the Coalition and Barnett are, according to the ALP, if the result is anything like what is being predicted, who will the ALP blame? The voters? The media? The Federal ALP? Themselves?

  12. [Given how terrible the Coalition and Barnett are, according to the ALP, if the result is anything like what is being predicted, who will the ALP blame? The voters? The media? The Federal ALP? Themselves?]

    They will accept the result gracefully, something the Federal Libs could consider in the interests of democracy, and change nothing.

  13. William

    Aren’t they ( the libs) simply going through the marginals in their hopes and estimations and factoring retiring members as chancesfor other seats.

    I think West Swan iis safe forALP ditto ColliePreston. Not sure about the Kimberely as you have said.

  14. @14 Whinger and Jokeshot betrayed their conservative base and should never have supported an ALP government. They will be slaughtered at the next election.

  15. [@14 Whinger and Jokeshot betrayed their conservative base and should never have supported an ALP government. They will be slaughtered at the next election.]

    Clearly they saw their self sacrifice in the interests of the nation.

  16. According to today’s West. The promise of a rapid transit bus service for ellenbrook and surrounds to make up for the broken promise of the rail line has itself been broken.

    This should go down well with the residents of Ellenbrook etc

  17. [According to today’s West. The promise of a rapid transit bus service for ellenbrook and surrounds to make up for the broken promise of the rail line has itself been broken.

    This should go down well with the residents of Ellenbrook etc]

    Would love to know if they are giving up on West Swan or have it in the bag. If Labor can’t capitalize on this they can’t capitalize on anything.

    It is very disappointing the Labor brand is so poor they don’t seem to be getting traction with a strong, needed vision against tired old incremental candy promises, along with the inevitable fees and charges increases and decrease in the public service and continued decay of public transport from Barnett.

  18. We want Paul

    I notice that they are focussing on Balcatta and Joondalup in recent announcements . My sense is that they have given up on WestSwan and perhaps even Swan Hills.

  19. [I notice that they are focussing on Balcatta and Joondalup in recent announcements . My sense is that they have given up on WestSwan and perhaps even Swan Hills.]

    I hope you are right, the Libs really don’t deserve any votes in either electorate they have been an absolute disgrace. I can’t believe Alban actually got preselection a second time, preselecting him for a seat they presumed Radisich would retain, and preselecting him for Swan Hills minus Ellenbrook, after seeing his wisdom, ability and courage on display for more than 4 years is beyond me.

    They say Labor has problems selecting hacks, and I think we do, but I’d put up our worst hack after 2 bottles of wine and s/he could eat Alban fresh and sober without raising a sweat.

  20. For what its worth, my brother-in-law was polled twice on the weekend and he lives in Mt Lawley. Sounded like internal polling- not a familiar company. Could there be a few worries about Sutherland hanging on. A lot of my well-to-do friends there don’t like him at all!

  21. Compact Crank, after reading your posts on the subject, I have put a few dollars on Kucera in Mt Lawley…will owe you a beer if the bet comes through for me!

  22. I’ve just returned from having dinner at the ancestral home.

    Before dining, an ad for Barnett came on the telly. Part of the pitch was that he was going to keep up the record spend on infrastructure, AND keep taking the fight to Canberra.

    My Mum, aged 74, suddenly burst out, “You lying sack of shit. All of that infrastructure depends on getting money from Canberra.”

    Never been so proud of her in all my life.

  23. Listening to the Liberal Bullshit ads you got to wonder did Barnett cost the new AFL gift 8 years ahead,have they costed any of there policy’s 8 yr in advance.
    Apart from that I keep wondering who is this wonderful Barnett who stands up to Canberra over the GST,he signed up for it,I don,t recall either he or Court saying very much about Canberra as Howard continued to Centralize power.
    I think that if the voters are as stupid as they appear he will have Abbott nothing will change including the GST distribution and Barnett will go very quite indeed,still the local media gutless as usual continue to beat the drum for the Libs.

Leave a Reply

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