ReachTEL: 52-48 to LNP in Queensland

After a dead heat in a ReachTEL poll conducted on the evening the election was announced, a second poll for the Seven Network finds the Liberal National Party moving ahead.

The Seven Network reports that the second ReachTEL poll of the Queensland election suggests the tide is going in the Liberal National Party’s direction, which is consistent with what I’ve been able to observe of the campaign. Whereas the first poll on the night of the election announcement had it dead level, this poll has the LNP leading 52-48 from primary votes of 42.0% for the LNP (up 1.7%), 36.7% for Labor (down 1.4%), 8.4% to the Greens (up 0.8%) and 5.2% to Palmer United (down 0.9%). Two-thirds expect the LNP to win, and a question on preferred LNP leader if Campbell Newman loses Ashgrove has Fiona Simpson moving ahead of Lawrence Springborg, with likely nominee Tim Nicholls well down the list.

UPDATE: Full results here. The automated phone poll was conducted last night from a sample of 1635.

UPDATE 2: For the purposes of illustrating a point I’m making in an article I’m writing for Crikey, I feature below calculations of the LNP’s two-party vote at the federal election transposed on to state election boundaries. In the other two columns are the 2012 state result (asterisk indicating an estimate because no LNP-versus-Labor count was conducted), and what I’ve called “SSD”, or “standardised state difference”. This equals the relevant electorate’s deviation from the mean at the state election, minus the equivalent result for the federal election. So for example, Nanango, Kawana and Burnett were seats where the LNP had relatively stronger results at the state than the federal election; Curtis Pitt in Mulgrave and Kerry Shine in Toowoomba North appear to have been two Labor members who performed particularly well in spite of everything; and Mansfield, ever the bellwether, was bang on the statewide average both times. There have been a few seats where I deemed the state 2PP to be unestimateable (inestimable?), and you might well think there should have been a few more.

 Federal State SSD
Albert
59.8%
61.9%
-1.2%
Algester
51.5%
59.1%
4.2%
Ashgrove
56.4%
55.7%
-2.3%
Aspley
57.2%
71.7%
10.6%
Barron River
60.0%
59.5%
-2.6%
Beaudesert
69.0%
77.1%*
8.7%
Brisbane Central
53.0%
54.9%
0.5%
Broadwater
70.2%
61.3%
-10.0%
Buderim
71.3%
76.0%
4.6%
Bulimba
51.5%
50.1%
-4.9%
Bundaberg
58.2%
68.2%
6.6%
Bundamba
38.8%
48.2%
5.0%
Burdekin
61.0%
71.5%*
7.3%
Burleigh
66.2%
61.0%
-7.0%
Burnett
63.8%
76.1%*
10.8%
Cairns
59.9%
58.9%
-4.3%
Callide
72.9%
78.2%*
5.0%
Caloundra
65.8%
71.2%
6.3%
Capalaba
54.0%
53.7%
-3.7%
Chatsworth
57.4%
64.1%
3.4%
Clayfield
60.3%
70.6%
8.0%
Cleveland
63.9%
68.1%
1.4%
Condamine
78.2%
80.9%*
2.9%
Cook
55.1%
53.4%
-6.3%
Coomera
68.4%
73.3%
3.9%
Currumbin
62.7%
70.2%
6.4%
Dalrymple
67.0%
Everton
60.5%
63.2%
-0.2%
Ferny Grove
59.0%
59.5%
-1.6%
Gaven
63.3%
69.1%
3.6%
Gladstone
45.3%
Glass House
64.3%
70.4%
7.3%
Greenslopes
54.4%
52.5%
-4.8%
Gregory
69.8%
75.5%
4.9%
Gympie
69.0%
76.1%*
7.9%
Hervey Bay
64.3%
71.7%
5.4%
Hinchinbrook
73.7%
73.1%*
-2.0%
Inala
36.3%
43.1%
1.3%
Indooroopilly
65.8%
69.5%
3.4%
Ipswich
41.7%
54.2%
8.1%
Ipswich West
49.0%
57.2%
4.3%
Kallangur
54.1%
62.4%
5.4%
Kawana
66.2%
76.3%
11.7%
Keppel
56.0%
56.4%
-2.4%
Lockyer
66.6%
71.3%*
5.5%
Logan
49.0%
54.8%
2.0%
Lytton
50.2%
51.6%
-1.9%
Mackay
55.4%
49.5%
-9.0%
Mansfield
56.3%
61.1%
1.8%
Maroochydore
66.1%
70.9%
4.1%
Maryborough
64.9%
Mermaid Beach
73.6%
76.0%
0.8%
Mirani
55.7%
61.2%
3.2%
Moggill
69.8%
73.9%
3.0%
Morayfield
57.4%
55.6%
-4.9%
Mount Coot-tha
57.5%
55.4%
-2.1%
Mount Isa
69.8%
56.0%*
-15.7%
Mount Ommaney
59.3%
66.5%
4.1%
Mudgeeraba
72.0%
75.9%
2.4%
Mulgrave
59.7%
48.9%
-14.9%
Mundingburra
58.3%
60.2%
-0.9%
Murrumba
51.2%
59.5%
3.0%
Nanango
70.9%
82.8%*
12.0%
Nicklin
64.4%
Noosa
69.1%
77.5%*
8.2%
Nudgee
47.4%
53.1%
1.3%
Pine Rivers
57.3%
63.7%
3.8%
Pumicestone
59.8%
62.1%
-1.0%
Redcliffe
50.2%
60.1%
4.4%
Redlands
63.6%
71.1%
4.5%
Rockhampton
47.8%
46.1%
-5.8%
Sandgate
45.6%
52.9%
2.8%
South Brisbane
44.2%
45.3%
-0.4%
Southern Downs
76.2%
80.1%*
4.5%
Southport
67.5%
64.7%
-4.6%
Springwood
57.9%
65.4%
4.6%
Stafford
54.3%
57.1%
-0.5%
Stretton
50.7%
59.6%
5.3%
Sunnybank
51.5%
60.2%
4.7%
Surfers Paradise
74.5%
79.5%
2.9%
Thuringowa
61.6%
61.3%*
-2.3%
Toowoomba North
67.5%
59.6%
-9.0%
Toowoomba South
68.4%
71.6%
2.4%
Townsville
58.5%
54.8%
-6.0%
Warrego
78.2%
81.0%*
2.2%
Waterford
51.2%
51.0%
-3.9%
Whitsunday
62.8%
60.7%
-4.0%
Woodridge
36.1%
44.2%
4.1%
Yeerongpilly
45.6%
51.4%
3.7%

UPDATE 3: And below are full results for the weekend’s Galaxy poll, which I don’t think have been published. Swings from 2012 election noted in italics.

LNP
ALP
GRN
PUP
LNP 2PP
Sample
Cairns
40
44
8
5
47
610
-3
+17
+1
-11.9
Barron River
43
42
10
5
50
700
-3
+14
+1
-9.5
Mulgrave
31
52
4
5
39
600
-1
+18
–
-9.9
Greenslopes
36
50
12
41
511
-12
+11
-1
-11.5
Pumicestone
46
40
7
3
52
567
-7
+10
+1
-10.1
Mundingburra
43
44
8
5
49
644
–
+18
+3
-11.2
Thuringowa
38
42
5
5
48
696
+2
+15
+1
-8.7
Townsville
33
46
10
6
42
611
-6
+17
+2
-12.8

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.

176 comments on “ReachTEL: 52-48 to LNP in Queensland”

Comments Page 3 of 4
1 2 3 4

  3. The assumption is that one or the other main parties will have a majority. Given the fracturing of voting in recent elections and even in Queensland, this election looks likely to result in a hung Parliament. Labor 36-40, LNP 40 -43 and the rest to be distributed throughout the pot porri of Independents and doggy scraps.

    This could end up being the phony election.

  7. I think the “Murdoch Poll” meme is bullshit. Seriously guys, get over it. That said, according to Possum, his internal ReachTEL polling for the unions is hovering around 51.5 to the LNP on 2012 prefs, and has been for 6 months. Every recent poll is with the MOE from that starting point. It seems likely RM favours the ALP a bit generally on these state polls across the country.
    [Leroy ‏@Leroy_Lynch
    @Pollytics any thoughts on the latest public ReachTEL?
    7:28 PM – 21 Jan 2015 · Details

    Possum Comitatus ‏@Pollytics
    @Leroy_Lynch 6 months now and nothing has changed. Averaging out bouncing around 51.5% on 2012 prefs. On we go…. again
    7:29 PM – 21 Jan 2015]
    And…
    [Possum Comitatus @Pollytics
    I have current polling-not just public-as this:
    http://www.abc.net.au/news/qld-election-2015/calculator/?mode=overall&overall=11.5&SEQ=custom&REG=custom&retiringmps=false&mary=ind
    BUT with 4 ALP wins and 3 LNP wins wrong, & 3 further coin tosses
    7:53 PM – 21 Jan 2015]

  10. William,

    My misunderstanding. I thought with your comment at 39 you were updating the Queensland polling for the duration of the election.

  13. I misunderstood too, GG – I thought you were referring to the federal BludgerTrack, which I promised I would update overnight but didn’t. Anyway, the answer to both questions is no. I’m about to investigate the record of Morgan’s SMS polling in Queensland to see if it’s worth including.

  14. John Holznagel@96

    I’ve just received an email from Morgan Gallup with its latest poll (today) and the result shows that it is 50.5% – 49.5% to lnp. This is unchanged from November (not on Morgan Gallup website yet). This poll is also untainted by the murdoch press unlike Galaxy or NewsPoll.

    It is however tainted by rampant overestimation of the Green vote and wildly volatile results. At least that is what the Victorian election showed us. Morgan SMS polls were the worst public polls in the Vic election.

    I really think this stuff of judging a poll by the fact that it is hired (not owned as some insist) by the Murdoch press should stop. If association with Murdoch caused Newspoll or Galaxy to skew to Coalition then this would have shown up by now either in election results or tracking. Yet this is not the case. Murdoch may well run a lot of ludicrous right-wing editorial bias but it is not in their commercial interests to be publishing bad polling data.

    So why do people keep assuming Galaxy and Newspoll results must be tainted by association with Murdoch, as an a priori position, instead of examining whether or not that association causes a taint by reference to actual evidence?

    Indeed in Victoria it was the final Morgan result that leant to Coalition compared with Galaxy and Newspoll.

  15. Mr Bonham went:

    [I really think this stuff of judging a poll by the fact that it is hired (not owned as some insist) by the Murdoch press should stop.]

    I agree. Though your opinion seems to flip the full 180 degrees when our union hires someone to do polling o_O

  20. [We could all hold hands GG, and bore ourselves and everyone else to death]

    I have never found your posts boring Possum. I wish you would visit here more often.

  25. Yep Poss and Kevin.

    No pollster is going to doctor their results.

    But from personal experience, clients (e.g. media, even unions!) do have influence on question wording, timing of polls and whether they are actually published.

    So what’s your latest saying Poss?

  29. Winston went:

    [But from personal experience, clients (e.g. media, even unions!) do have influence on question wording, timing of polls and whether they are actually published.]

    The Question wording of course! But that’s published.

    Timing? Well the snapshot of reality (which is what polls are trying to measure) is what it is. Do we say a poll is wrong because it came on a Wednesday as opposed to a different day?

    Or even say it’s wrong because it came at a time that is inconvenient for some political body or unit?

    Does the inconvenience make the poll wrong? Or are the results still what they are regardless?

    Some polls are unpublished – for a whole lot of reasons, especially campaign ones where you’re running an under the radar campaign in a seat, you’re tracking the results, and don’t want to give the game away. Or you’re issue testing questions that you plan to campaign on and you don’t want to give the game away.

    But what the accusation usually is with these things is that only cherry picked polls are released – which is pretty amusing from a number of angles.

    Let’s say, for example, that someone did do such a thing.

    Let’s say someone only released polling results favourable to their cause (what’s ‘favourable’ about a poll result is another story) – but let’s say the polls deemed acceptable for release had results that came in more than a standard deviation above the mean.

    That would require doing, on average, five to six polls for every one you released.

    So what would be the purpose? To change public opinion because you released a poll result that was 2 points higher than normal? Even though the next poll(s) that came out by someone else would be inevitably lower?

    Polling is expensive – whether you hire someone to do it, or you build and run your own system. For the same price of running 5 or 6 polls, repeatedly, just to release the ones you like – you could run a large TV advertising campaign that would hit the eyeballs of between 50 and 100 times more people than would read the result of your poll. You could put 5000 people on the ground and doorknock 100,000 households for the same price. You could do a whole bunch of mini campaigns running sequentially to really change the mind of the the people you are targetting.

    And that’s assuming that people reading a poll result would actually change their voting behaviour because they read it (they don’t, I’ve tested it).

    So financially, it makes no sense at all. Politically, it makes no sense at all, as a campaign strategy it makes no sense at all. Even as a throw away tactic it’s pointless because the only people that care about those sorts of games have been living in the political bubble for so long, they’ve long since ceased making a difference to anyone and anything.

    Polling results from good firms are what they read as on the label, regardless of who paid the dollars for them. Because the alternative is genuinely ridiculous in nearly every way, shape and form.

  30. Super response Poss.

    Agree with a lot of that.

    But the average punter doesn’t take much notice of the question wording – only people like you and some of the Pollbludgers. Most just see the headline. And I’ve found that the media are pretty indiscriminate about what the publish – they are more interested in whether the poll throws up some newsworthy result than how the poll was conducted.

    But who commissions the poll is very relevant. Commissioned polls may cover a range of issues and use different questions to measure opinions. What gets released publicly is the results which further the clients interests. Results are in fact cherry picked. I’m not referring to basic voting intention polls here but polls on issues. Would you release any polling that wasn’t favourable to your union’s agenda?

    Also, most of the pollsters have major commercial clients who they wouldn’t want to upset by polling issues that reflect badly on them. So the pollsters would be reluctant to even do public polling on issues like attitudes to banking. Not in their commercial interests.

    And taking all polls at face value seems bit naive to me – and I know you’re not that.

  31. This may well be a super-boring election, with the LNP losing a swag of unholdable seats but cruising to inevitable re-election.

    On the other hand, there are many volatile ingredients – Liberal v National friction, the cock-sparrow Premier from Tasmania, the rookie Opposition leader, the Abbott syndrome, Clive Palmer, Bob Katter, Alan Jones, the humidity that sends people troppo, optional preferential voting, the problems for pollsters etc etc.

    With more than a week to go, anything could happen.

  32. It doesn’t need to be a question of doing a bunch of polling and releasing a selective result as your predetermined objective. A union might do a poll for purposes of its own research, and if the result happens to be helpful for publicity purposes, whack out a press release with the results attached while they’re about it. They might even do so if the result is below the mean – if 60% are opposed to privatisation and the poll comes in at 58%, they’re hardly going to let that stop them. But if the result can only generate unhelpful publicity, they would probably prefer to keep it to themselves.

    If a union (or any other interested body) made a regular habit of publishing polls, I might just treat this tendency as a house bias effect to be corrected for like any other poll series. But if there’s just one or two results here and there, I can see why you might prefer to leave it alone.

  33. Possum Comitatus@115

    Mr (sic) Bonham went:

    I really think this stuff of judging a poll by the fact that it is hired (not owned as some insist) by the Murdoch press should stop.


    I agree. Though your opinion seems to flip the full 180 degrees when our union hires someone to do polling o_O

    A severe exaggeration, eg see “Newspoll Corroborates ‘Union Push-Poll'” section here: http://kevinbonham.blogspot.com.au/2014/04/is-campbell-newman-actually-in-trouble.html And you either know it or haven’t been paying enough attention to comment.

    That said, the distinction I do draw is completely justified, because the primary purpose of newspapers in commissioning polling is to obtain information that increases the interest of their product to the public. They have a strong reason not to suppress data because they have wasted their money if they do.

    On the other hand a political group that commissions a poll of any kind may be interested in it for its own strategic benefit (irrespective of the public benefit in seeing the data) or may be interested in using it to convince the public or politicians that something about public opinion is true. Neither of these things lead to a situation in which the sponsor of the poll has a motive to reliably and systematically publish their results. Does your union do so?

    A Mr Silver who has looked at the matter passingly in the US is of the view that what I call “internal” polling is less accurate than public media polling, possibly because it is easy for the pollster to get suckered into telling the commissioning source what they want to hear.

  34. Agree with Toorak Toff it seems that many issues and irritants sent late in this campaign to create an ordeal for Cando. Alan Jones, former LNP Minister Chris Davis, Head of Police Union Leavers all formerly staunch LNP supporters and now spitting chips at the worst possible time. Reported issues swirling around current Minister Walker, Newman’s broken promises to Alan Jones and alleged subsequent lucrative donations to LNP from said mining co-op, Humorous and bizarre dancing Gold Coast member with serious questions to answer, Cando refusing point blank to sign Tony Fitzgerald’s letter re good governance (which all other parties signed no probs) – He seems under siege right now with probable humiliating loss of seat and very nervy. Everything going wrong which must affect final countdown to polls. Makes for interesting run home.

  35. Kevin@114
    To be fair Morgan wasnt alone with crazy GRN polls, IPSOS was just as bad.

    SMS only polls are a distortion because irrespective of phsyical age and whatever sort of balancing the pollsters try and do they are still biased towards those with a ‘young mind’.
    Combine this with the desperation of the fringe to be heard more than the center, and you get those distortions. (IMO)

    Newspoll has historically been biased in the opposite way, they never used to used SMS (dont know if they do now) so they where biased against the ‘modern demographic’ and hence favour the LNP.

  36. I seem to recall lots of people here claiming Morgan was biased to labor before they started doing SMS polls, people claiming face-to-face favoured ALP as well.

    We are all a bit biased towards one pollster or another i guess.

    Morgan does have good breakdowns and is fortnightly, so its the best for following trends. (there is my bias showing, but i dont trust anyones SMS polling)

  37. Possum Comitatus@129


    But what the accusation usually is with these things is that only cherry picked polls are released – which is pretty amusing from a number of angles.

    Let’s say, for example, that someone did do such a thing.

    Let’s say someone only released polling results favourable to their cause (what’s ‘favourable’ about a poll result is another story) – but let’s say the polls deemed acceptable for release had results that came in more than a standard deviation above the mean.

    That would require doing, on average, five to six polls for every one you released.

    So what would be the purpose? To change public opinion because you released a poll result that was 2 points higher than normal? Even though the next poll(s) that came out by someone else would be inevitably lower?

    Polling is expensive – whether you hire someone to do it, or you build and run your own system. For the same price of running 5 or 6 polls, repeatedly, just to release the ones you like – you could run a large TV advertising campaign that would hit the eyeballs of between 50 and 100 times more people than would read the result of your poll.

    The first faulty assumption there is that the 5 or 6 polls would be run just to release the ones you like. The accusation you refer to is not that – the accusation is that unions and other actors commission numerous polls for internal information purposes (the money being well spent irrespective of public release) but then are selective in which ones they release. The selectivity doesn’t just extend to innocent tactical reasons but also extends to avoiding releasing results that are not tactically desirable because they are bad.

    The second problem is the assumption that those involved are actually tactically rational. They’re often not – there can be a desire to boost the morale of the support base with messages that they’re on the right path. This behaviour continues no matter how many times it is proven to be ineffectual in actually lifting vote share. (The Greens are world champions at this but I doubt they’re alone.)

  38. A few of the posts re skewing are missing important points.

    Yes if you only use landlines that will skew your sample towards older voters. But you can get around that to a degree by either waiting until you have enough younger voters, or weighting the results you have so the young voters count for more. The first method costs more and the second makes your result more volatile. But you can do it and Newspoll have been doing it very successfully at recent elections.

    The idea, put about by the Bob Ellises and so on of this world, that if a polling method is more likely to be answered by older voters that therefore its results will skew right is just nonsense. The pollsters aren’t idiots, unlike Ellis. They know about this issue and they work around it.

    Whether mobile-phone-only polling or SMS-only polling can be fixed up the same way is something we’re not yet clear on. It may be that these technologies in Australia have skews to the Greens that run so deep that nothing you can do short of fudging the figures will get rid of the skew (eg even within any given demographic, those who use a mobile phone are more likely to be Greens voters). But I’m sceptical of this. I think it’s not that the method is hopeless but that it is still very new and that those trying exclusively mobile polling just aren’t doing it very well.

  40. davidwh@139

    Kevin my question was how can you identify the demographic when using SMS polling?

    As I understand it Morgan uses a panel so they’ve pre-recruited respondents, they know their age data, and then they sample a random portion of the panel each week. They already have the age data for each respondent who answers on file against their phone number. (This method has a lot of limitations if your panel is not representative.)

    If you do random SMS polling – which I’m not aware of anyone doing – you’d need to ask them their demographic details via SMS as well.

  42. I thought those polls might have had some money come back onto Labor, but everywhere that’s moved has gone the other way. You’re doing well if you can still get on 1.12 for the LNP.

  43. With the polls firming up in the 51 or 52% to LNP 2PP range, I think the real X factor here is going to be the extent to which the preference flow from GRN/others to ALP differs between this election and the last election – especially if the report I heard that slightly less than 50% of Greens votes made it through to Labor as preferences last time.

    I would expect that number to be at least 60% this time, even taking into account OPV (under full preferential, I would expect it to be at least 80%, if not more than 85%). As others have noted, this would cause polls quoting 2PP figures based on previous-election preferences to be off by one or two percent.

    Still, I think even the most generous adjustment to Labor only just barely puts them in minority government territory.

  47. Wiiliam 145 – was that “Who’s got the wheel?” ad really 2001? I remembered it as 1998 but if the CCA were in there it would have to be 2001. It’s like all the people who remember “It’s the Economy, Stupid” as the slogan of Clinton’s 1992 campaign. When in fact it was just an internal “meme” made famous in the campaign film “The War Room” and was not actually used as a public slogan.

  48. In

    [As I understand it Morgan uses a panel so they’ve pre-recruited respondents, they know their age data, and then they sample a random portion of the panel each week. They already have the age data for each respondent who answers on file against their phone number. (This method has a lot of limitations if your panel is not representative.)
    ]

    this is true for Morgan, and they use their Face to Face sessions to categorise people demographically and sign them up for ongoing commercial product and electoral polling. They also offer incentives.

    I suspect the A/B demographic they cherish is underrepresented on Morgan panels, as if you are well off to start with, you wouldn’t waste an hour of your time with his recruiters

    I only wish that the U.S. Practice of some pollsters to publish the refusal rate would be replicated here

  49. Roy Morgan polls didn’t show any particularly strong pro-ALP bias in the Victorian 2014 election, and the final poll actually had a pro-Coalition bias.

    The 7-10 November Roy Morgan poll was 53.5-46.5 to ALP compared to 56-44 in the 6-9 November IPSOS and 54-46 in the 27-30 Oct Newspoll.

    The ALP vote-share rose to 55-45 in the 19-20 November Morgan poll, but then reduced to 52-48 in the 21-24 November Morgan poll, which was in the same range as other polls at that time.

    The final Morgan poll was, of course, 50-50 between ALP & Coalition, with a 2 percentage point overestimation of the Coalition vote & 2 percentage point underestimation of the ALP vote.

    So, it seems to me that there aren’t any strong reasons to outright reject the Morgan QLD poll because “Morgan polls have a pro-ALP bias”.

Comments Page 3 of 4
1 2 3 4

Leave a Reply

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