Nov 6, 2009

Morgan Adds Outlier Weight to Newspoll

Morgan has cheekily come in a week early, giving us the first week of what is usually a two

Possum Comitatus — Editor of Pollytics

Possum Comitatus

Editor of Pollytics

Morgan has cheekily come in a week early, giving us the first week of what is usually a two week Face to Face poll with the primaries running 51 (down 1) /32.5 (down 2) to Labor, washing out into a two party preferred of 61/39 the same way – a half a point increase to Labor. The Greens are on 9.5 (up 2) while the broad “Others” are sitting on 7 (up 1). This comes from a sample of 1050 giving us an MoE that maxes out around the 3% mark. This poll was taken over the period of October 31 to November 1 – the same period as the now infamous Newspoll that has caused so much hyperventilation by the usual suspects over the last week.

If we plug this Morgan result in with the other ultra-regular polls of Newspoll and Essential Report, we now come up to date as of November 1st.

ALPprimary lnpprimary


Outliers are simply part of life for a pollster – they happen and there’s absolutely nothing they can really do about the odd one popping up. So the first thing to point out is that if the last Newspoll was an outlier – the probability of which has just increased – there is no Newspoll conspiracy here. We actually expect pollsters to give us a WTF moment every now and again.

What catches my eye about these three pollsters is the way the ALP primary vote most likely did actually drop a tad between the 25th of October and the 1st November. Both Essential and Morgan had the Labor primary dropping a point over that period – after having tracked each other virtually identically over the period since late September.  Newspoll showed a drop too, but it looks like it overshot the true underlying movement of the public by 3 or 4 points. The direction of the change was consistent with the other 2 pollsters, only the magnitude of the size of the change differed.

The Coalition primary however is where it get’s interesting, with both Essential and Morgan showing a slight drop (1 point for Essential, 2 points for Morgan) while Newspoll gave them a 7 point increase. Not only did the magnitude of the Newspoll change differ from the Morgan and Essential, but the direction of the change differed as well. One thing the pollsters did all pretty much agree on though was the Greens vote:


So if Newspoll was an outlier, it all came down to a direct substitution between ALP and Coalition supporters on measuring the vote estimates – with the Greens vote effectively staying consistent across all three pollsters. However, the satisfaction ratings and preferred PM figures didnt move in Newspoll as much as a 7 point change in the primary would ordinarily suggest. If it was just sampling error responsible, we’d have expected those non-vote metrics to move more since effectively all that would be happening is a larger proportion of Coalition voters was sampled by chance.

So it’s probably something more complicated – some form of non-sampling error. However we know that Newspoll runs a kick arse sampling frame so it’s not some structural problem, but rather most likely a human issue, or more particularly, a respondent issue. That gets us into navel gazing areas of what makes a significant number of ordinarily Labor voting people keep their preferred PM and satisfaction ratings the same as if they were voting Labor, but tell the pollster instead that they’re actually intending to vote for the Coalition? That seems to be what happened to Newspoll.

The alternative is that Newspoll picked up a change of voter sentiment earlier than the other two polls. If that was the case, we’ll know on Monday with Essential Report.

(Visited 2 times, 1 visits today)


Leave a comment

58 thoughts on “Morgan Adds Outlier Weight to Newspoll

  1. Sam Bauers

    Yet [the ute-gate conspiracy] was all true, not a theory at all, but fact.

    I’m not so sure that those you are saying are involved actually were. No mud has stuck to anyone except the one public servant who supposedly concocted the emails (and obviously that would always be the case). I’m not saying what you propose isn’t the truth, but there is an equal level of plausibility to the “lone gunman” theory because of the inherent nature of the way News Ltd. works.

    If it was the one person working alone, it is entirely possible that their “revelations” could whip some News Ltd. journalist/editor/etc. up into a frenzy just as we saw. No conspiracy required. That’s what you should be worried about, not some supposed fiddling around the edges on one or two occasions. Their modus operandi is entrenched and does not require specific intervention for conspiracy-like chains of events to occur. They happen naturally. That’s why I’m saying that as a concerned individual you should save your energy to fight against the entrenched biases of the organisations you oppose, rather than risking being painted as something you are clearly not.

    I respect your passion for this topic, but there is nothing to be won even if you turn out to be right, maybe except for the scalp of some peon at NewsPoll and a strongly worded public statement.

  2. Possum Comitatus

    Paul – it may just be a sub-sample issue. Maybe something was going on last week that particularly affected the types of people that were home answering surveys last weekend.

    It’s pretty much impossible to tell, or even really guess without the raw respondent level data.

  3. Possum Comitatus


    Newspoll isnt really the culprit here – it’s News Ltd journalists talking out their arse *about* the Newspoll figures.

    Newspoll would cease to exist if they started fiddling with their figures because they earn their biscuits from commercial research. We need to remember that News Ltd only own 50% odd of Newspoll. If Newspoll was some in-house News Ltd job then, considering the sordid history of some News publications here and around the world, there would be a case to at least watch carefully what comes out of such polling.

    But Newspoll isnt an in house job and the integrity of the people who work there is beyond exceptional. Some of them would resign before they’ be party to deliberately fiddling figures.

    Outliers BB just happen. The fact that people are telling me that they’ve been phoned up over the weekend for another Newspoll tells me that far from being some conspiracy to bring down Rudd, they were concerned enough at the results of their previous poll to spend considerable money to run another one to make sure that what they are producing is as accurate as possible.

  4. Bushfire Bill

    [Wailing about imagined conspiracies allows others to pigeon-hole you …]

    “Conspiracy theories” are always depicted as crazy, until they’re proved correct. It’s not really an argument. It’s a cheap sledge. Considering what News Ltd journalists have gotten up to in the past, a conspiracy is a much better bet than it should be.

    Who’d have thought that solemn testimony, complete with embarrassed reticence, and coughing reluctance delivered to a Senate Inquiry by a trusted public servant about the PM being corrupt could possibly be completely manufactured? Who could believe that this public servant would later be found to have been running a comprehensive, but secret scheme, in collusion with the Leader of the Opposition and a senior Senator, designed to make the PM and Treasurer look like cheap crooks? Who could believe that a copy of an email published, in its original format, as proof of the public servant’s evidence, would be a fake? It’s all so “conspiracy theory”, and would allow the perpetrator of it to be pigeon-holed as paranoid.

    Yet it was all true, not a theory at all, but fact.

    That it really was a conspiracy doesn’t prove that everything is, of course. But it proves that such scams can and do exist. It’s just that they’re not usually discovered as quickly as this one was, if they are discovered at all. When a news organization has form, consistent form, for bodgying up important news stories, you start to get leery about anything it publishes as “fact”.

    Some forgive each individual action as “part of the culture” or “simple stuff-ups”. Taken one by one, in isolation, it’s a sort of half-plausible explanation. But when you consider the stuff-ups as a body of work, it’s not. News Ltd repeatedly “stuffs up”, and the object of their stuff ups nearly always happens to be the Labor government. That indicates more than clumsiness to me.

    [The emails were fake, that’s not a controversial point. The idea that the designer of the layout was somehow knowledgeably complicit is ridiculous.]

    I never said that. I wouldn’t expect a busy graphic artist to know all the names involved (it’s possible, but I’m not assuming that). However, the very fact that the name was wrong (a mixture of two of the “players'” names in the general story, but any name other than “Godwin Grech” would do to make my point) proves beyond doubt that the “email” graphic was concocted, and it had to be concocted by someone: man, woman, journalist, editor or hanger-on.

    That was my point.

    [No doubt if the other numbers did change it would also mean that there was a conspiracy.]

    Cutting sarcasm, but the fact remains Newspoll used the consistency of the PPM figures as “proof” the poll was accurate. I never mentioned your other scenario. That’s a Straw Man all out of your own head Sam.

  5. Sam Bauers

    Sucked in big time, Sam.

    Actually, no. Wailing about imagined conspiracies allows others to pigeon-hole you and in the meantime the real influence-peddling keeps occurring, you should save your good intentions for more important fights.

    The very fact that all other numbers were unchanged was used by Newspoll to prove the poll wasn’t wrong.

    That’s pretty contrived. No doubt if the other numbers did change it would also mean that there was a conspiracy.

    You said someone gave the graphic artist the bodgy text.

    The emails were fake, that’s not a controversial point. The idea that the designer of the layout was somehow knowledgeably complicit is ridiculous. It’s a wasted point in your argument, that event was a scam by one or more people but there is scant evidence to suggest that it was a “top-to-bottom” or “organisational” conspiracy.

    The graphic was meant to look like they had a copy of an actual, real email, one that was sent from one person to another.

    Clearly it was a graphic treatment. The text in the “email” flows around the guys head. This sort of “dum-dum” literal layout is par for the course at News. I’m not doubting that the graphic treatment is meant to lend weight to a story’s intent, I’m saying that it doesn’t require a conspiracy for that to occur.

    Luckydave talked about the entrenched culture in News Ltd. that perpetuates this treatment of news. In such an environment overt conspiracy is not required. I recommend reading “Manufacturing Consent”. It’s an important academic work on entrenched media organisation bias that points to a much larger problem than any individual conspiracy.

  6. Paul from Berwick

    Hey, is this outlier problem really a sampling issue?

    As has been stated elsewhere, the “collapse” in the vote was for the under 50’s. Now, if Newspoll simply had a higher number of under 50’s who were Liberal voters (ie, the sample was either “biased” toward electorate’s with a high under 50’s weighting, or there were more of under 50’s home due to exams, away due to Victorian public holiday, etc).

    Could it simply be say 200 people more than ususal who had the following characteristics: were both under 50 & Liberal voters?

    Just a thought – neither conspiracy or stuff-up, just Australian life on Newspolling day!

  7. Bushfire Bill

    [Surely if they were going to fiddle the primaries they would fiddle a few other numbers to make it more convincing overall.]

    Sucked in big time, Sam. The very fact that all other numbers were unchanged was used by Newspoll to prove the poll wasn’t wrong. Keeping the other numbers the same gave it veracity (until another poll came out and blew it away).

    And your faith in stuff-up versus conspiracy is touching. You said someone gave the graphic artist the bodgy text. Exactly.

    And what David above says too. If this wasn’t an attempt to influence the casual reader, I don’t know what was. As far as I can see from the graphic as posted (link on previous page) there is no reference to “Artist’s Impression” or similar exculpatory words. The graphic was meant to look like they had a copy of an actual, real email, one that was sent from one person to another. Except there was never an actual email to “Godwin Grant”.

  8. David Richards

    Sam – to the technical illiterati, a cobbled together graphic looking the same as an email they might get from cousin Dweezle would look pretty real and add gravitas where there wasn’t any.

Leave a comment

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details