Monday, March 15, 2021

AsiNine: NSW Labor Not At Hundred-Year Low

With much attention on the Western Australian election and also a rather interesting Newspoll this weekend, unfortunately there was a third story that saw some of the most lamentable poll reporting that has been seen from a major network for a while.  I've put the other matters aside for an hour or two now, because what has happened here needs to be strongly condemned.

The first I saw of this poll was a news item tweeted by the Nine News Sydney account, concerning a disaster poll for NSW Labor, one supposedly finding that the party's primary vote was lower than at any election since 1904, and worse even than the 2011 debacle.  The Twitter video of the news report was three minutes and 18 seconds long.  I watched it three times in disbelief that a report of such a momentous poll did not even name the pollster involved, let alone state who had commissioned the poll.  Then I ranted about it on Twitter.  Then I watched it a fourth time just to make sure that I hadn't just made a real fool of myself.  But it was real - 198 seconds complete with interviews with ex-Premier Morris Iemma, but no mention of the pollster's name.  This alone was quite ridiculous.

(I've been told there is a longer version that does name the pollster, but haven't seen it, and don't know how widely versions with and without the pollster named were broadcast.)

Indecision Clouds My Vision

But things got worse than not naming the pollster, much worse.  Firstly, I found an SMH report of the poll by Chris O'Keefe (also the presenter of the TV item).  This report declared that:

"The poll predicts Labor’s primary vote has crashed to 23.9 per cent, a figure lower than the landslide loss of 2011 under former premier Kristina Keneally (25.6 per cent) and an 8.4 per cent drop in primary vote on the 2019 election. It would be the lowest primary vote recorded by NSW Labor since 1904."

This did mention who the poll was by (RedBridge) and where it came from (the AWU, though it only mentioned said union as the supplier and not the commissioning source).  However, the report was still very thin on detail.  The only voting intention results it reported were the Liberal Party on 36.5%, Labor on 23.9% and intriguingly "18.9 per cent of voters remain undecided."  No statement of how these undecided voters were treated - were they redistributed according to a leaning they expressed when prompted, were they excluded and effectively redistributed (a practice Newspoll employs), or were these simply raw figures with the undecided voters left in the sample?

It turns out it is actually the latter.  The 23.9% figure for Labor is in a sample where only 81.1% expressed an actual voting intention.  If an election were held, it is simply bonkers to suggest that none of the 18.9% "undecided" would vote Labor, especially as Labor voting intention is often softer than for Liberals and Greens when voters are actually asked how sure they are of their votes.  uComms polls that prod initially undecided respondents for a preference habitually find that those undecided voters who do express a preference go all over the place.  Comparing 23.9% out of 81.1% decided to 25.6% out of 100% at an election is not comparing oranges with oranges.  It shows both an embarrassing level of innumeracy concerning polls, and also an embarrassing level of compliance with the wishes of the suppliers of this story.  It is an extremely severe case of the unhealthy synergy blighting Australian poll reporting across most if not all networks, in which a free and sexy polling story is somehow nearly always rewarded with uncritical and gushy reporting of dubious and often false polling claims, with nary an independent polling commentator in sight.

(As an aside, it's also wrong to say that a poll "predicts" anything at this time.  Very close to an election, a poll can be regarded as a prediction of sorts of the result, but two years out from one there's no event; the poll is but a snapshot.)

Never minding all this, there is also the issue that an 18.9% undecided rate is just too high.  Most pollsters find undecided voter rates in the high single figures, though sometimes they have to prod initially undecided voters as to which way they are leaning to to get there.  A very high undecided rate usually doesn't mean that huge numbers of voters have no idea who they would vote for - it usually means that respondents are not being pushed hard enough to express an opinion.  The undecided voters might break strongly one way or the other - for a number this large, if, say, 50% of them leant to one major party and 20% to another, that would make a few percent of difference to the 2PP.  

What Is The Actual State Of Play?

RedBridge have been notable for issuing several polls of NSW federal seats that show very gloomy narratives for Labor.  While it is always possible that RedBridge just happen to have found a bunch of seats where Labor is doing particularly badly, there is no support for the idea that Labor are doing badly in NSW federal polling generally, so if they are doing badly around the Hunter Valley they would be doing well elsewhere in the state (where?).  Of course it could be that Newspoll are wrong and RedBridge are right, but RedBridge have never been tested at an election where they have posted public results in advance, while Newspoll's record is well known and mostly good aside from the 2019 failure.  RedBridge have also been known to pollwatchers for some very unusual approaches to measuring voting intention (such as including leader names in the party description, which is likely to favour incumbents), though I understand this time they didn't do this.

The RedBridge primaries (thanks to Kos Samaras of RedBridge for this data and no thanks to the SMH for not publishing it themselves like they should have) were "Libs 37, Labor 23.9, Nats 3.1, 4.3 PHON, SFFP 0.8, Greens 6.7, Indi 5.3, 18.9 not sure."  With the "not sure" redistributed proportionally that would come to Lib 45.6 Nat 3.8 ALP 29.5 Green 8.3 ON 5.3 SFF 1 Inds 6.5.   I estimate the 2PP off these numbers at 59-41 to Coalition, a 7% swing which would see Labor lose about 9-10 seats out of 37 if the draft redistribution boundaries are confirmed.  

For sure that would be a dire outcome, but it would not be 2011.  And we don't know if it's accurate - though we also cannot yet know that it isn't.  We have the usual warning signs here: poll commissioned by sources with skin in the game (specifically, the desire to replace Jodi McKay as leader, most likely with Chris Minns), poll by pollster without a publicly testable track record, and breathless media reporting without sufficient details of question wording, orders of questions and other important methods information.  

Unfortunately, I can't benchmark this RedBridge poll against other NSW voting intention polls, because there aren't any, and haven't been since the 2019 election.  For what it's worth, I think the Coalition would win a NSW election held right now, and would increase its majority.  The NSW Government has been remarkably effective in keeping a lid on COVID in recent months without the level of restrictions seen in Victoria and Gladys Berejiklian continues to poll excellent ratings despite scandals that would have destroyed or damaged political mortals, or anyone perhaps in normal times.  I am confident most voters would see her government, despite its flaws, as deserving re-election.  

What we have here is a collaboration between a poll sponsor seeking to achieve a political goal and a media source that was at best willing to put sensationalism ahead of the interests of its readers in obtaining a sober and realistic view of this polling.  This is not good enough from Nine, especially not when it dumped Ipsos following the last federal election.  At that time, Nine justified its decision (in my view really just cost-saving) thus:

"As chief political correspondent David Crowe says, accurate polling can be an invaluable reality check when journalists are faced with relentless spinning by political parties, interest groups and think tanks."

Now it seems that Nine has thrown in with the relentless spinners. It's disappointing when we are starting to see some signs of better poll reporting in the media (for instance the West Australian's reporting of the YouGov Dawesville poll that correctly forecast the demise of Zak Kirkup was excellent, providing an unusual depth of coverage of the details of this seat poll).

"However, we have a responsibility to put our finite reporting resources into journalism that best serves our readers."

No, you are now serving faction hacks playing still more of the whatever-it-takes internal ALP games that have made NSW Labor a bad joke for rather more than "ten wasted years".  You are betraying your readers and the interests of voters in the process. Whatever it is isn't journalism.

I can do no more than to award Nine this.

Porcupine Fish Award for Ultra-Fishy Poll Reporting (image credit)

4 comments:

  1. There's one more factor: polling this far out from an election tends to be much less accurate than usual. It's much easier for respondents to give affective responses ("oh I feel like Labor isn't doing much, I'll say undecided") when an election campaign isn't on.

    At the same time, kudos to RedBridge for putting out a poll during this time. It may not be perfect (in particular I suspect they didn't push voters to select a party e.g. "if undecided, to which party are you leaning?"), but I personally believe some indication of voting intention is better than nothing.

    ReplyDelete
  2. I do find myself wondering if this is an attempt at damage control following the shocking WA results for the Liberals and the consequent interest in the Federal implications - the spin, I mean, not the poll itself. After all, this particular media group has always let a pro-liberal bias show, or so it seems to me.

    I live in NSW and I'm not even sure who the state Labor leader is right now, which is a telling point in terms of the cut-through for Labor's message, and suggests to me that state Labor have very little chance at the next ballot. It will take a massive turnaround in approach to alter that expectation.

    So, I suspect the poll is more right than wrong. But I could be more wrong than right!

    ReplyDelete
  3. This; "...you are now serving faction hacks playing still more of the whatever-it-takes internal ALP games that have made NSW Labor a bad joke for rather more than "ten wasted years". You are betraying your readers and the interests of voters in the process. Whatever it is isn't journalism".

    What short term thinking for Labor. Achieving the leadership aims of a small minority, damaging the wider party brand for the much longer term.

    'Reporting' like this all too common these in the age of sensationalism for clicks. What a shame that Labor is unable to capitalise on "scandals that would have destroyed or damaged political mortals".

    Another 10 years in the wilderness not unlikely.

    ReplyDelete
  4. Thanks for this very informative article.

    I sometimes call out bad science reporting, and you calling out bad poll reporting is much more important.

    I have expressed my thanks in the hopefully helpful way of dollar notes (virtual of course).

    ReplyDelete

The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.