Monday, March 6, 2017

Patchy Polling In WA As The Final Week Begins

Going into the final week of the WA campaign, not much has changed from where it started a month ago.  The relatively scant and yet surprisingly diverse nature of polling data available leaves poll-watchers free to choose their own adventure, from a comprehensive win for Mark McGowan's Labor opposition to a very close and messy race in which Colin Barnett's Liberals might even cling on in minority by the skin of their teeth.  About the only thing the statewide polls agree on is this: following a shambolic and confused campaign, One Nation may have tanked.

The following table gives all the statewide polling known to me.  The poll marked as "Essential" is known to me only from a Poll Bludger exclusive that describes it as Greens-commissioned Essential robopoll.  That description places it at least two degrees of separation from anything we can place verified trust in but I mention it all the same.  I've also given an average and a time-weighted (but not performance-weighted) aggregate, in each case excluding the Essential.

In this case though, I don't place that much confidence in aggregation methods. There is a major house-effect difference between the ReachTELs and the Galaxy/Newspoll stable, worth at least four points on each major party's primaries, and coming out at two or three on the 2PP only because ReachTEL use respondent preferences (which have been skewing massively to Labor).  There's a pretty good chance here that someone's right and someone's wrong.  We're not just seeing margin of error issues here - there would not be such large and consistent differences in the major party primaries by pollster if we were.  (And no, voting intention doesn't bounce around this much in reality through a campaign either.)

The few commissioned electorate polls purport to show Labor getting slightly larger swings than the 7-11% implied in the polls above where it matters:

ReachTELs for The Parenthood have Labor winning six marginals, two of them narrowly, with an average swing of 12.9%.

Multi-marginal ReachTEL for Tourism Council showing the swing at around 12%.

Some more notes on the Galaxy.  Firstly it has Mark McGowan preferred Premier by a massive margin of 46-33. No real inroads have been made by Barnett on this score and retaining office with such a deficit on a score skewed to incumbents would be amazing.  Barnett leads 70-10 among Liberal and National supporters, McGowan 81-7 among Labor supporters, and I infer that McGowan leads about 42-21 among supporters of minor and micro parties.  A further sign of trouble for the government is that it leads by only one point (38-37) on a question about likelihood to reduce state debt, a natural conservative strength area as with other such economic management questions.

One Nation

As noted above, it is looking like One Nation may have tanked a bit, although their vote is still significant, may still be understated, and would still be enough to give them a good shot at Upper House seats.  ReachTEL's measurement of their vote is complicated slightly by having them on the readout in seats they aren't contesting. I'm unsure if this is also true of Newspoll/Galaxy.

The One Nation campaign has been a shambles, featuring many cases of disendorsements and apparent "resignations" (some of them after the ballot papers were printed, meaning that they have no actual effect.)  It's been obvious that One Nation's candidate vetting like PUP's in the Tasmanian state election, has been practically non-existent.  This is scarcely surprising given that One Nation's vetting was done by a married couple with a combined age of 165, in these days when finding red marks against candidates requires advanced social media search skills above all else.  The vetting has been so bad that nasty piece of work Richard Eldridge was outed as a liability by the West Australian two weeks before nominations closed and nothing was done about him.

But the party has had bigger problems, with its deal with the Liberals causing substantial disunity and it seems a degree of brand damage.  It's increasingly unclear - both state and federal - whether One Nation is trying to market itself as the ultimate protest party or simply as a substitute for the Nationals at the Liberal Party's side.  This is complicating its leader's reputation as a straight shooter.

One Nation preferences in the Lower House remain a source of mystery that could have a point or so's impact on the final 2PP result.  Respondent-allocated preferencing tends to suggest they will split pretty evenly, but the party is preferencing the Liberals and may have Liberal Party assistance in handing out cards if needs be.  How much difference that might make if it comes about is hard to say, because One Nation voters tend to be distrustful (including of their own party's cards.)  A 60-40 split to the Liberals could happen; any projection above that seems adventurous.

The other key point is that the impact of One Nation preference directions on individual seats will be patchy, because One Nation are only running in 35 of 59 seats, and are notably absent from Labor target seats in the 4-10% range.  Even in the 10-15% range they are only running patchily.


I think that what benefit the Liberals might gain from any preference shifting involving the One Nation how-to-vote cards will be modest and could be partly cancelled out by shifts in the preferencing patterns of other parties. (Especially, Labor's share of Green preferences in WA has a lot of room for improvement.)  At this stage I am inclined to weight the historic track record of Galaxy somewhat higher compared to ReachTEL in resolving the stoush between the two, and there is also some reason to suspect the damage could fall more heavily where it counts.  If the 2PP is about 53-ish then the ABC's uniform swing calculator credits Labor with only the slimmest of victories (30-29), but this is again premised on the Coalition retaining a ridiculously high proportion of close seats.  My variable swing calculator converts 53% 2PP to a most likely seat result of 33-26 to Labor, followed by 32-27.  If the 2PP is more like 54 and the seat swings fall in the right place, a significantly higher seat tally is on the cards.  The likelihood remains that Labor is winning and the Barnett government's chances still appear slight, though we still don't have the quality of data needed to write the contest off.

Once again, a note that I won't be online much over the weekend and will unfortunately be unable to provide live coverage.

Update: Liberal Internal Poll Rumour

Yet again a Liberal source has released talked to a journalist about internal polling saying they are losing 57:43. This would reduce the Liberals and Nationals combined to around 19 seats if it happened on election day.  In this case we don't even know if this polling exists and assuming it does we don't know if the release of these figures is deliberate or is being done by a rogue element.  However it is common for parties that believe that they are going to lose to release such dire internal polling for expectation management reasons so that a moderate loss is considered an acceptable result.  It's very unlikely we'd be seeing this stuff if the party thought it had any chance whatsoever of holding on.

Final polls

Both Newspoll and ReachTEL have come out at 54:46 to Labor, though the ReachTEL would be more like 53:47 by last-election preferences.  On this basis my model has about 34-35 Labor seats.  For the Liberals to retain they would most likely need to depend on both the polls being slightly wrong and a lucky distribution of seat swing. As noted above I don't alas have time to cover the election in detail this weekend but I will have some post-commentary in a few days.


  1. Kevin, why would "The ParentHood" and the Tourism Council commission voter-intention polls? I can imagine the Tourism Council might have done it as a proxy for the Illiberals, but hardly the other lot. Is it more that they commissioned a poll on issues of interest to them, and ReachTEL hung the voter-intention questions on to that? Or what? [And why oh why do they put the TEL in caps?]

    And if you want to start a not-a-poll on the 2PP, I'll vote for 53/47

  2. There were extra questions on the Parenthood poll (example of reporting here: Ditto the Tourism Council poll ( It's standard for ReachTEL to ask the voter intention question first on all polls especially so they can do party breakdowns on subsequent questions. Commissioning sources seem to be realising that providing the voting intention results to journalists makes it much easier to get a publicity hook for getting results of the other questions published. [As for the caps, I do not know.]

  3. PHON candidates in the WA lower house are still leaving and being disendorsed in relatively large numbers. There will still be many candidates listed on the ballot as being PHON endorsed. I imagine the Liberals will have their volunteers hand out for PHON, it will be interesting to see how preferences flow and how PHON supporters will respond to Liberals telling them how to direct their preferences.

    1. I suspect a lot of One Nation supporters, if they cannot see their own party handing out, will tell the HTV card brandishers from the other parties where to go and how fast to go there. So the Liberals will probably need to have people who are handing out specifically for One Nation to have the full desired effect.