Reversal of fortune

Remarkable poll results from McNair Ingenuity Research in today’s Sunday Herald-Sun give the Coalition a commanding primary vote lead of 46 per cent to 39 per cent one week out from the Victorian election. From this it is somehow inferred that Labor still holds a 50.5-49.5 lead on two-party preferred, suggesting an extraordinary 77 per cent of minor party and independent preferences will go to Labor. McNair managing director Matt Balogh reckons the figures "may reverse on the day a little bit", a clear indication that he himself does not believe them. The sample size is an almost but not quite respectable 609.

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.

31 comments on “Reversal of fortune”

  1. If the pollster thinks it’s a ‘surprise’ poll isn’t that a good indicator that it’s not that accurate? A large change like that sounds unreliable.

  2. I simply don’t believe the swing will be that big. There is nothing out there that would justify such a massive swing and massive it would be. besides it goes against the Newspoll.

  3. Latest Morgan Poll has ALP 55 – LIB/NP 45 unchanged from last time

    But has a 13.5% Green Primary Vote (which imho is way overstated and won’t be anymore than 8-9% at best in reality)

  4. I think there is an explanation for it but I need to do some more research. Problem is that when the electorate is feed information and believes that one party is dominating and will win an election they feel over secure and begin to take risks that under a close election they would not take. This is what did in Jeff Kennett to a large extent. The main difference between 1999 and this election is that Jeff was hated by many and the same degree of hatred does not apply to the Bracks government which has not abused its majority in both houses. It has overall be a responsible government.

    My main concern is that should a Lieberal State Government cross the line then there will be nothing to stop John Howard from seeking to overstep his responsibility and unleash a raft of issues that attack civil rights in Australia. We have seen it in the past and we will no doubts see it again.

    A labor State Government is the best means of ensuing checks and balances in Canberra. Now is not the time for complancency.

    I have updated my upperhouse predictions to reflect the McNair poll.

    Click on my name above to view.

  5. Bort where did you get your Morgan Poll from. I think that it might be more realistic. But there is an error in what you have reported. The 55 45 is the two party ALP/Lib preferred. I think there is a consolidation of the Green vote in the inner city as many voters are of the belief that labor will win the election hands down. It is possible that this might effect the seats of Kew and even hawthorn although it is hard to see the liberal Party not going beyond 50% on the primaries it is never the less a theoretical possibility. A lot of voters will decide this week or even minutes before casting their vote.

  6. Channel Nine news tonight commissioned the Morgan poll

    Those were TPP figures – the 13.5% Green primary was tacked onto the story without giving the primary vote for the other 2

    I still reckon the Green vote is grossly overstated and they won’t come close to winning any lower house seat

  7. Yes the age and Channel 9 are reporting the Morgan Poll which sound more realistic. It was a telephone poll of 778 people ALP primary vote down 3.9% whilst Greens up 3.8% Whats interesting is the high percentage of complacency in with over 70% of victorian believing the ALP was going to win. I might wait for the Newspoll which is expected to be published tomorrow.

  8. Details of the Morgan poll can be found on their web site

    Extract below..

    If a Victorian State Election had been held this weekend the ALP would have won, but with a reduced majority from the 2002 State Election (ALP 57.8%; Liberal-National Party 42.2%).

    Primary support for the ALP was 44% (down 3.9% since the 2002 election), Liberal-National Party 38% (Liberal 34%, National 4% — about the same as the 2002 election result), The Greens 13.5% (up 3.8% since the 2002 election), Family First 2.5% and Independents/Other 2% (down 2.1% since the 2002 election).

    A key contributing factor to the final election outcome will be where the relatively high 7.5% (up 2.5% since October 24/25/26) that remain undecided direct their vote.

    Analysis by groups of seats shows that with Liberal preferences, the Greens have a good chance of winning some inner city Melbourne seats particularly Melbourne and Richmond. The Liberals could gain seats in the fairly safe ALP country seats in the Ballarat and Geelong areas, with the outcome being determined by preferences of Nationals, Greens, Family First and Independents.

  9. Another factor worth noting is that minor parties other then the Greens and NP are down in percentage points confirming my belief that family First was not going to do better then in the 2004 Senate. Basing my upper-house research on the Senate vote is proving to be valid and a good way of profiling the electorate.

  10. Melbcity… remember Victoria is the only state that has given its system of awards under Federal Jurisdiction whereas all the other states have keep their own. I think that’s right for memory. But even with it under Federal Boundaries – I agree – a State Labor Govt could pull out any moment should workers rights really be threatened.

    William – a couple of queries about the Vic 06 Election guide.
    1: Is the Nepean page down because it’s the only electorate I can’t
    get into it to read about.
    2: For South-West Coast you rightly put it as Liberal… however under
    candidates list you list the current sitting member as (Labor) where
    it should say (Liberal).

  11. The Morgan poll is also showng a 2% drop in the liberal/National Party vote (Mainly form the National Party side). Stephen Maynes People power is not even on the radar yet they have spent over $35,000 in non refundable nomination deposits. peoples power;’s only hope of attacting voters i this late stateg is the possibility that it becomes a single issue party ‘Anti Pokies” failing that it is game over for People Media players.

  12. The Age today reports a massive swing towards the Greens in inner city seats with the Greens obtaining 40% of the vote. If reports of a 40% vote are true then the seat of Melbourne will change hands but it is not all comforting news for the Liberal Party whose vote in the inner city has also been eroded.

    In theory the Liberal Party could suffer in Kew and in Hawthorn where the increase in Greens support has pegged back the Liberal Party in what may very well become a Green/Liberal contest.

    IN 2002 the Liberal vote Hawthorn and Kew fell below 50% forcing the Liberal party to preferences. It is expected that Family First will also also help peg back the Liberal Party vote keeping it below 50% and forcing the Liberal Party to preferences for the second time. Family First are the only other minor party that has preferenced the Liberal Party ahead of the Greens. Whilst Hawthorn is expected to be safe the same can not be assured for Kew. Green preferences are not expected to alter the outcome in the lower house as most Green voters have already made up their minds which party to support second and will not be influenced by the Greens HTV card. Party names are published on the ballot papers and Green voters will decide for themselves. But in situations such as Melbourne and Richmond the Greens will most likely out-poll the Liberal party and as such those voters who have parked their votes on the Greens will be locked in and Green preferences will not be distributed. the same possibility could occur in Key with the Greens outselling the ALP.

    Upper-house results do not change as a result of the Greens surge.

  13. The Green vote in Kew was 17% and 19% in Hawthorn.

    The Opposition leader is member for Hawthorn (imagine the political fallout if he lost his seat to the greens).

    Did that inner city poll cover Richmond, Northcote and Brunswick as well?

  14. Melbcity… outselling the ALP? how?

    by telling voters to make up their own minds? andn ot preferencing either major party?

    the ALP has some nerve. you guys can preference anyone you like and get any conservative elected. But the Greens have to be a preference feeder to you guys?

    Seriously, the ALP is deluded.

    And stop feeding talk of a deal. If it was a deal, the Liberals got a pathetic deal. No preferences. 3 year old can figure out that that is a not a deal.

  15. Only if they stay below Labor who`s preferences are tightly anti-liberal.
    A 7 percent swing form Labor to the Greens in Hawthorn without an increase in the Lib primary vote (wich will almost definatly increase) would send the seat Green.

  16. but only 20% tops of greens preference lib before alp, and in victoria less than 20% prefrence lib. so there has been no deal between them.

  17. The problem is, Elbow Room, you’re basing your assumption on a flawed poll. No other poll predicts a 46% primary vote fo the Liberals.

  18. Two things, the federal Liberals may be on 46% not the state Liberals.

    Hawthorn will swing toward its local member whom has run a good campaign, his name is Ted, I believe his Liberal leader.

  19. Actually bmwofoz not even the Fed Libs have been on 46% in polls for some time now. The’ve been lowish 40’s. I agree with you about Hawthorn.

  20. The McNair Gallup Poll was conducted over 3 evenings, with a very high response rate, which is more important than sample size in determining an accurate result. I am Confident the poll conducted last week was accurate. The main third party are the Greens, and the majority of Greens preferences will go directly, or indirectly to Labor.

  21. So Matt if the McNair poll was accurate that makes all other polls inaccurate by a good margin doesn’t it? I’m not sure that the 3 nights or high response rate makes any difference to be honest. Why would they?

  22. If Matt Balogh believes that response rates are more important than sample size in determining accurate results then he needs to hand back his membership of the society of pollsters.

    In 2nd year physics we did experiments on radioactive decay of atoms – the ultimate random sample. The message that was drummed home to us again and again was that even with a totally random sample you could, and indeed would, get anomolous results if the sample size was not big enough. A perfect “response rate” ie equipment which measured every decay, is no substitute for a sample size that lacks statistical significance.

  23. You said the Herald Sun poll gives “the Coalition a commanding primary vote lead of 46 per cent to 39 per cent one week out from the Victorian election. From this it is somehow inferred that Labor still holds a 50.5-49.5.”

    As Matt said in his comments, that seems logical to me. Look at the example of Prahran district in 2002. From memry it was about 46% to 37% Liberal well ahead of Labor on first preferences. But almost 11% for the Greens of which 9% went straight to Labor ending up with a 2 party preferred vote of 54 to 46 in Labor’s favour!

  24. Well, I wouldn’t say that sample size counted for nothing – a bigger sample is better. But even a huge sample won’t compensate for a sample bias or the bias due to poor response rates.

    And yes, the McNair Gallup Poll, over it’s 40 year history, has often been more accurate that other polls, and this been well documented. We realise that our poll points to a closer election than the others, but that doesn’t make our poll wrong.

    Oooh and I am glad to meet another person that likes to compare radio-active breakdown to sampling theory.

  25. Matt where is the detail in the results. Given that you poll is swidely out of kilter withy the other pollster it is incumbent on your company to back up your cliams by publishing more detailed breakdown and analysis. In furether review of polling results published yesterday and the surge of the Greens Eatsren metro’s last seat could readily fall to the Greens giving them a shot at four upper-house seats. I am srill working thorugh the results and playing with the margins. What is clear is that there is little “other” room for other minor parties. Morgan clearly made a mistake in his reulst assessment as Western Metro will deliver at least 3 ALP one Liebral and possibly a second. The other aspect of the consolidation was that it looks lioke preferences will not be distributed at all in Northern Metro which could save the VEC some money or out a number of tally clerks out of work.

  26. I’ve just returned from voting in London in the election. It was surprisingly busy, and officials were running out of ballots for inner city seats. They were estimating that over 2500 people will vote in person, with many more sending in postal ballots —- not sure of the exact effect though, as the ballots aren’t counted unless the result in the seat is close! In Geelong in 1999 it was of course overseas votes which ‘decided’ the result

  27. This particular poll was commissioned by The Sunday Herald Sun, so we are not permitted to release anything other than what they have published – I am sure you will understand that they really own the results. Sometimes we conduct polls off our own bat, and these are published more openly on our website.

    The actual voting intention was only part of that particular poll, and we only asked about the lower house, although there was an interesting question about how well people who represented them at the regional level.

  28. That’s the perfect example of how sample size doesn’t make up for sampling issues. It’s not just that not everyone can respond to an online survey, but far more importantly, there are clear patterns to the profile of respondents and non-respondents, with a strong bias to younger respondents. Well, in 48 hours all will be told! 🙂

Comments are closed.