web analytics
The Standard

Polity: Morgan’s poll, and Morgan’s commentary

Written By: - Date published: 10:54 am, June 7th, 2014 - 47 comments
Categories: election 2014, elections, polls - Tags:

polity_square_for_lynnReposted from here. Rob Salmond at Polity looks at the latest Roy Morgan poll.

Earlier this week the latest Roy Morgan poll came out. It was not good news for the left, with National up well over 50% and the Labour/Greens combination languishing below 40%.

This poll’s field period started just a few days after the Budget, and ended before the Banks stuff, so I think it is picking up a post-Budget spike for National and I do not think National will stay at this height for long.1 In addition, most of the poll was taken prior to the Laila Harre appointment was made, so I would not read much into Internet MANA’s poor joint showing.

Nonetheless, the poll shows that the centre-left parties have a big, big hill to climb in the remaining three-and-a-bit-months.

The poll also shows that Roy Morgan needs a new internal commentator on New Zealand politics. Building on the amusingly-naive claim a few weeks back that the Internet Party was busy building its support at the expense of the Conservatives, comes this week’s gem:

In addition the Greens (9%, down 4.5%) have slumped to their lowest level of support since before the last New Zealand election after announcing last weekend a proposal to introduce a Carbon Tax in New Zealand in place of the current Emissions Trading Scheme.

That announcement came out on the same day the poll closed. Eesh.

Unlike some others, I have a good amount of faith in the Roy Morgan polling method and in their estimates. But right now the commentary isn’t worth the pixels it is projected with.

1. This isn’t because soft voters are sitting around reading the Budget docs, believing them, then reading some Labour / Green / IM package later, and believing that instead. It is actually because soft voters remember a vague cue from the TV news or newspaper when they’re asked how they will vote. Just after the Budget, that vague memory is more likely to be of smiles and handshakes for a surplus. As those shots fade from the media, its effect fades from the polls.

lprent: Just by way of comparison, here are the equivalent figures for the first Roy Morgan poll taken fully after the budget in  2011, like the poll in the post above (ie to compare after two events)

Party 30 May-12 June 2011 Election 2011 19 May-1 June 2011
National  54.0  47.31  52.5
Labour  30.5  27.48  29.0
Green  5.5  11.06  9.0
Act  2.5  1.07  1.0
Maori  3.0  1.43  1.5
United Future 0.5  0.60  0.0
NZ First 3.5  6.59  4.5
Mana    1.08  0.5
Internet     0.5
Conservative    2.65  1.0

2011 poll: N+A+MP+UF = 60.0%, L+G+NZF = 39.5%

2014 poll: N+A+MP+UF+C = 56.0%, L+G+NZF = 43.5%

2011 election: N+A+MP+UF+C = 53.06%, L+G+NZF+M= 46.21%

The left block (assuming that percentages translate into seats) is in a far better position this June than they were in 2011 and this time there is no rugby world cup to distract voter attention. As Rob says; the government usually gets a significiant post-poll boost and that fades. Just like from the commentariat’ very short-term memory. hat-tip: colonial viper

47 comments on “Polity: Morgan’s poll, and Morgan’s commentary”

  1. Chooky 1

    “Bugger the polls…and bugger the pollsters” as Jim Bolger said.

    I think people are keeping their powder dry until they get into the voting booth.

  2. Tautoko Viper 2

    I agree with Frank Macskasy” view.
    Let’s be respectful to our allies and all face the common enemy.
    “United we stand, divided we fall, ……..

    • David H 2.1

      Yeah Franks real good at this stuff.

    • Ant 2.2

      Totally way to much time attacking allies……it also doesn’t help that the left’s most visible commentators spend about 50% of their time attacking the left.

  3. weka 3

    How much notice does the MSM take of RM’s commentary?

    • Lanthanide 3.1

      Pretty much nil. But MSM generally only pretend that their polling is the only polling that exists. Although two recent cases earlier this year they seemed to collaboratively break out and talk about ‘wider polling’, the first case around January was when two polls came out showing opposite scenarios, and again around March showing National strongly up in a range of polls.

      • swordfish 3.1.1

        No. National took a hit in the March polls and the MSM got it entirely wrong. They were down 5 points with the Right Bloc down 3.

        As I mention here the MSM employed contrasting modes of poll analysis in February and March: the common denominator being that in both cases the methodology underlying their analysis favoured National. In February, the MSM did indeed “break out and talk about wider polling”. In doing so, they rightly highlighted the clear swing to National and the Right in that month.

        Unfortunately, in March, the MSM employed an extremely restrictive mode of analysis which, for instance, saw Gower depict the 3 News Reid Research Poll as Good for National and the Right and Bad for Labour and the Left – when, in fact, it was no such thing. The corollary of what I would call both the Time-lag effect and the Media Exclusive effect, possibly with a subtle (and sometimes not so subtle) underlying pro-National MSM partisanship.

        RM results are covered on a fairly regular basis in the MSM Print Media – albeit very briefly in those little one-paragraph side-bar stories.

        The thing that really intrigues me about RM is their unusually small proportion of Undecideds – far lower relative to the other 4 public polls. I’d like to know their precise methodology and its impact on accuracy.

        • swordfish

          In terms of Lynn’s point that the Left Bloc is currently in a far better position / Right Bloc in a far worse position than they were in 2011, see my various comments and figures here and here and one of the probable reasons for the consistent over-stating of National and Right Bloc support here.

          • swordfish

            National – Average Monthly Poll support 2014 relative to 2011

            Jan 2011 52 % 2014 45 % (down 7 points)
            Feb 2011 52 % 2014 49 % (down 3 points)
            March 2011 52 % 2014 46 % (down 6 points)
            April 2011 54 % 2014 46 % (down 8 points)
            May 2011 52 % 2014 49 % (down 3 points).
            (In June and July 2011, National’s support averaged 53 %)

            National – Average Monthly Poll support at same point in electoral cycle 2014 relative to 2011

            8 Months out from Election – 2011 52 % — 2014 45 % (down 7 points)
            7 Months – 2011 54 % — 2014 49 % (down 5 points)
            6 Months – 2011 52 % — 2014 46 % (down 6 points)
            5 Months – 2011 53 % — 2014 46 % (down 7 points)
            4 Months – 2011 53 % — 2014 49 % (down 4 points)

            • swordfish

              The corresponding figures for the Right Bloc 2014 relative to 2011 are:

              Jan down 7 points
              Feb down 2 points
              March down 5 points
              April down 9 points
              May down 4 points

              8 Months down 6 points
              7 Months down 5 points
              6 Months down 6 points
              5 Months down 9 points
              4 Months down 5 points

              • Colonial Viper

                Excellent could you tell us what the figures are for 3, 2, 1, and 0 months are 😛

                • swordfish

                  3 Months Down 22 points
                  2 Months Down 28 points
                  1 Month Down 35 points
                  Final Month Down 43 points

          • lprent

            Cool. Worth reading even if it is incomplete..

        • Lanthanide

          Roy Morgan polled me, on my cell-phone no-less. I asked them how they got the number and they said they did random calling, and that’s why they asked me for my postcode at the start of the poll so they could get a fair sample across the country.

          Other than that it seemed like any other regular poll. They also sent me out a paper booklet with about 40 pages of other consumer survey stuff which I filled out and got a $20 petrol voucher for. I expect they probably used the phone poll results when reporting their political polling, although IIRC the same questions were also in the booklet (but since that took a week to arrive and then return via mail, they can’t be using that when reporting the results).

          • swordfish

            Interesting, Lanth. Roy Morgan do seem to be the only company polling cell-phones as well as landline.

            It’d be very interesting to discover whether or not they prompt those who are initially undecided on the Party Vote question with the follow-up question ‘ “Is there a Party you’re more likely to vote for ?” or something similar. Because that kind of very broad, inclusive approach where prompted respondents are included in the final party support results seems to be One News Colmar Brunton’s modus operandi.

            And yet One News consistently records a far higher Undecided (even after the follow-up) – and therefore excluded sub-set than Roy Morgan. And I mean far higher.

  4. Tracey 4

    One reason why some people (including young and new voters) stay away from the polls if they think their vote is irrelevant cos the result is a forgone conclusion. This is why polls are so frequent, it reinforces a certain meme.

    I am not a fan of them and would like to see them gone, forever, and definitely a few months before an election.

    But its purpose? What does it serve? Not the people. It’s only serving the people if you think a poll is a good substitution for thinking.

    • karol 4.1


    • Chooky 4.2


      • lurgee 4.2.1

        I love polls. Can’t get enough of them. Wish we had daily polls like YouGov d in Britain. though as I seem to be on some pollsters contact list, it would mean I’d be getting polled several times a week …

        I the left ‘bloc’ can not motivate people to vote for them or take an interest it is hardly the fault of pollsters. Perhaps the left ‘bloc’ should take a look at itself and wonder why people are not interested in what they have to say.

    • poem 4.3

      Totally agree Tracey, well said. I have no doubts that national is using polls as electioneering tools to hoodwink voters and demoralise the opposition.

  5. Chooky 5

    Martyn Bradbury on the Polls

    iPredict vs Roy Morgan Poll

    By Martyn Bradbury / June 6, 2014

    “I think ACT is over priced, Conservatives about right, IMP way too low and Greens I would buy up to 14%…..Two very different pictures are emerging between the iPredict market and the mainstream media polls.

    Roy Morgan’s this week was terrible for the Left, yet the iPredict markets are suggesting something very different.

    The iPredict Election show called the election closest last election and their read on National’s vote on election night suggests that punters acknowledge National are not doing nearly as well as the polls would have us believe. Punters willing to put their money where their mouth is say…


    • fisiani 5.1

      Current ipredict odds of National PM 74%

      You can short this for a layout of just $26 and win $100

      • bad12 5.1.1

        Cling to that wont you Fisiani, ipredict have the National Party polling at 44%, that’s one hell of a big Opposition,(and there is no position that i know of with the title of Prime Minister of the Opposition)…

        • fisiani

          you obviously don’t understand how ipredict works.
          Let me explain
          buy a 74c prediction ofPM Nat and if correct you win another 26c
          buy a 44c vote share and National gets 51% then you win another 7c.
          The clever money goes on PM Nat.
          If you are so clever and know that it will be The Cunliffe then you can make a tidy profit very easily.

          • bad12

            fisiani, your idea of ”risk” and ”tidy profits” obviously suffer a way lower thresh-hold to enable participation in a ”gamble” than mine,

            Believing that you are a bottom feeder obviously leaves me unsurprised by this, however, a tidy profit is a subjective term, i mean an obvious chimp like you will probably happily gamble 15 bucks and if you win blow your tiny little mind over the 5 extra bucks you have,

            At the same time tho,should you lose as all small timers eventually do you have an innate ability to provoke amnesia inside what passes for brain matter in your cranial cavity,

            i think i will decline your offer to play the odds casting a vote or two on the day seems a far more efficacious route to community prosperity…

            • fisiani

              why call it a gamble? If you really know politics then ipredict just means taking money off the misguided. That is not a gamble. For instance take the safe Labour seat of Palmerston North with a fabulous and well known incumbent. His ipredict price last month was 84%. Nat win was 16% Then Jono Naylor was chosen as the National Party candidate. I bought shares. Price now for Labour to win in 49%. Which one will win. My money is on Jono. A rising tide lifts all boats and Jono is the 3 time and current Mayor of Palmerston North and better known than the incumbent Labour candidate. Note how I am not disclosing who he/she is. Not many would know that nobody. If you know PN politics better than me then you can DOUBLE your investment in just 4 months.

              • bad12

                Hell fisiani when i get into investing my hard earned cash in anything i double the money i put in on a weekly basis,

                Off you go now and play with your little toy, you will be able to go to the ware whare after the election with a lot of luck and buy yourself a real soft fluffy one, with a squeak too…

                • dave

                  If you thank that’s true you should consider running a ponzi scheme!

                  • bad12

                    If i thank anything was true dave i reckon i would probably consider running a pinzu scheme,

                    Along with your inability to either spell or use the provided edit function i see the crux of your argument as really really THICK,

                    When i wish to double my money weekly on the original investement i just turn up the volume on the sounds and flick the switch to fire up my otherwise dormant double set of quad 600’s…

              • chris73

                Yes but that would mean putting money where ones mouth is whereas the left prefer to use other peoples money-

                • Colonial Viper

                  Yes but that would mean putting money where ones mouth is whereas the left prefer to use other peoples money-

                  So does the corporate right wing.

                  See also: finance companies, investment banks, fund managers and property developers. Also BHP, owners of the Bluff smelter.

  6. Sanctuary 6

    Polls should be banned for 100 days before the election. It might even force the media to talk about policies.

    • freedom 6.1


    • Bearded Git 6.2


    • Colonial Viper 6.3

      I was thinking no polling activity 21 days before an election (to balance freedom of speech considerations), but yeah.

      • Tempted to say they can poll all they want if they don’t publish the results, lol, but that does introduce the possibility of push-polling.

        • Colonial Viper

          The problem is that unpublished results can still be “leaked” and cause a media frenzy…best to ban commercial polling activity altogether for the 3-4 weeks before election day, then let the media rattle on about pointless 2 month old polls if they want to. (Also avoids the problem of censoring newspapers from printing news…)

    • poem 6.4

      +1Sanctuary Would like to see polls ditched permanently.

  7. bad12 7

    The data sets showing the actual numbers of all the Roy Morgan polls as opposed to the squiggly line one above is also an eye-opener,

    Sorry i haven’t got a link for it but it was linked to in ‘Open Mike’ the other day, what the actual numbers show tho is that the National vote keeps swinging wildly up and down, up,down,up down, over great swathes of the poll printed showing the actual numbers swinging not by a % or two but by 5% over 4–6 week periods,(like Roy cannot make up His mind),

    i find it hard to believe that the electorate as a whole engages in swings of that nature over such short time-frames…

    • lprent 7.1

      They have a particularly small sample size. But they are a lot more frequent.

      • lurgee 7.1.1

        I tend to assume individual RM polls are anything up to +/- 5%. So a bunch of them might give a fairly accurate idea, but by themselves pretty worthless. We can safely say National are more popular than Labour, but not much more than that.

        • lprent

          So a bunch of them might give a fairly accurate idea, but by themselves pretty worthless.

          Agreed for a single poll. However in this case I was at looking at the reaction immediately past an reproducible effect after an event that usually happens every year.

          If you look at the wikipedia link in that post you’ll find that there is a pretty clear pattern of post-budget polls for several years giving a significiant boost to the right bloc (especially National). If you went back further, you’d find that most post-budget polls usually strongly favour the main government party of the day (there are some obvious exceptions to that – the “mother of all budgets” being one of them).

          If you looked at a short rolling average you’ll see a consistent pattern showing a strong lift preceding and especially just after the budget for the main governing party. Which is exactly what you see for a number of years. I only bothered to look at it for 2009-2014 but it is there every year as a clear pattern.

          Not to mention that much of the variance between parties are within the overall right bloc and left blocs – which minimizes much of the ‘noise’.

          That is why I picked the way I presented it – as bloc votes after a recurring annual event. I could have dumped a frigging great badly organised spreadsheet or some of the messy R work that back the simple version that I gave as a example. But really why bother when most people won’t read it anyway.

      • swordfish 7.1.2

        Yep, Roy Morgan usually have a smaller sample (800-900) than 3 of the 4 other Public Polls. The exception is the Herald-Digipoll which has a sample of just 750.

        Roy Morgan, however, do vary their sample size and occasionally, just occasionally, they use a much larger sample (eg 1500 in early Jan 2014). Possibly part of an Omnibus survey.

        • lprent

          Interesting. I wasn’t aware of the larger sample sizes.

          It is a bit irritating the way that they don’t really state their margins of error for each poll. Trying to infer a non-linear value on just two points is ridiculously daft.

  8. Bearded Git 8

    Last 10 Roy Morgan’s. Nats v Lab/Green. Looks to me like the last poll was an outlier.

    43.5 46 +2.5
    47 44 -3
    48 42 -6
    48.5 41 -7.5
    45.5 45.5 –
    43 45 +2
    48.5 40 -8.5
    42.5 45.5 +3
    45.5 44 -1.5
    52.5 38 -14.5

    • geoff 8.1

      That’s a nice way to compare, BG. Cheers.

      I did a quick calculation, excluding the last outlier, and the average difference between the 2 blocs is Labour/Green -2.1. I did this by adding up the first 9 numbers and dividing by 9.

      If you include the last one you get an average difference of -3.35.

    • Lanthanide 8.2

      Possible outlier, but also budget effect.

Recent Comments

Recent Posts

Public service advertisements by The Standard

Current CO2 level in the atmosphere