Poll Roundup: 2022 Seat Poll Sludgefest

This is the second part of what was at first intended as a single roundup, and deals with seat polls and seat betting.  Any new seat polls will continue to be updated on this page until election day while a seat betting update will probably be added to a final roundup tomorrow night.

As a general comment, 2022 has seen a major change in the election seat polling landscape.  At previous recent elections, published seat polls were dominated (in public attention if not always in numbers) by major national players - YouGov, ReachTEL, the old Newspoll and so on, or at least by specialised pollsters who were distant from campaigns (JWS).  Often the seat polls still sucked.  In 2013 they skewed to Coalition on average (some firms more than others), in 2016 they were under-dispersed and in 2019 they did well at picking winners but badly when they sat on the fence (and they also skewed to Labor).  But at least they were neutral attempts by pollsters with skin in the national game.  

At this election YouGov has so far given seat polling the flick and instead switched to its MRP model (discussed too briefly in a previous edition).  The seat poll landscape has been dominated by uComms (a union-connected pollster with simplistic weightings and an ordinary recent track record), Redbridge (another campaign-focused pollster with often weird methods decisions and a remarkable ability to detect UAP voters), and to a lesser extent Utting Research and KJC/Telereach, neither of whom have had much public testing and the first of which is not an APC member, with publication of details rarely exceeding a single media article.  The overwhelming method of seat polling has been robopolling of often already saturated seats (one voter in Swan this week told me they'd been polled seven times).  Moreover, the two most commonly seen pollsters have been mostly conducting internal and campaign-adjacent polls rather than media-commissioned polls.   The seat polling landscape has been dominated by strategic or incidental releases of polling for campaigning purposes - mostly fed to journalists to get publicity and (in the case of teal independents in some seats) try to exploit strategic voting arguments.

In a tweet on Wednesday, Redbridge director Kos Samaras explicitly said "We do not poll to predict.", claiming that the purpose of his polls was to measure voting intention so that campaigns could change it.  But if campaign seat polls even in the final weeks are not predictive then is there any reason for newspapers to report them other than them being an easy story to write?  Whether pollsters poll to predict or not, surely their ability to measure the mood of the electorate at any time can only be judged by polls they put out for judgement at the end so they can show clients that they are accurate.  If there is no evidence a pollster is capable of accuracy when it counts, there is no evidence its numbers are right and no reason to believe they represent even a snapshot.  And so I intend to measure how accurate all these seat polls are or are not as per normal, and this article lists some recent ones so people can see how they go on election night.  However as usual I won't be writing up the accuracy findings in full until after all the House of Reps votes have been counted.  

North Sydney

North Sydney saw one of the more curious seat polls with an entry by Compass Polling, hitherto infamous as a source of right-wing culture war fodder for Sky News.  The remarkably detailed local newspaper report nonetheless didn't say who commissioned the poll (if anyone) or what the polling method was.  Anyway Trent Zimmerman (Lib) came out with a respondent-allocated 54-46 off primaries of Zimmerman 40.5%, Renshaw (Labor) 21.6, Tink (IND) 13.6, Greens 12.9 and nobody else above 3%, with TNL's Victor Kline on his way to a 20-seat majority, er sorry, on 0.8%.  Whether or not the poll is accurate we can see here a path to victory for Zimmerman off a low primary vote - the vote for opponents is too spread for any to climb the mountain (it especially helps him if Tink is eliminated.)

As if the teal team wasn't going to take that lying down, there soon emerged a Climate 200 Redbridge poll of North Sydney.  In this one it's Zimmerman 36 Tink 25 Renshaw 19 Greens 9.  The 2CP given is 54.5 to Tink, though I'd think more like 53.5.  The most interesting thing in this poll for me was a question about voting above or below the line in the Senate, which revealed that the polled voters rampantly over-reported voting below the line.  (18.3%, or 22% with undecided removed - the real figure was 9%).  Moreover, Liberal supporters, who actually did vote BTL quite a bit for Jim Molan in this seat in 2019, remembered which side of the line they were on reasonably well, while Labor and Greens supporters claimed rates a few times the actual figures.  An unrepresentative sample or poor voter recall?  Hard to say.

Redbridge also had a poll of Wentworth (Sharma (Lib) 37 Spender (IND) 34 Labor 13 Greens 7, Spender getting 70% of respondent preferences which would come out to 54-46 - I think she might get more than 70%).  Here we can see how it is harder for the Liberal to survive if the vote against them is concentrated.  Also of interest was table 4.7 which is the first polled evidence I've seen that unvaccinated voters do indeed like the UAP (tiny as the sample is).   And there was one of Goldstein with Wilson (Lib) 39 Daniel (IND) 29 Labor 15 Greens 9 and 80% of respondent preferences to Daniel (perhaps a bit high but Daniel would also win 54-46 even if it was a bit lower on these numbers.)  Another Goldstein poll, by uComms for GetUp! also had Daniel winning around 56-44.  With Liberal internal polling also being said to have Tim Wilson on only 37, I am not sure what I will ever do with my time if I no longer need to have a separate column in TweetDeck to see if Wilson has been ratioed again.

How many of these indepedents are really winning?  We will find out soon enough.

Utting WA Polls

Utting Research released robopolls reported in the West Australian of four previously polled seats, finding movement back to the government in all of them since March, but not enough to avoid losing Swan (47-53) or Pearce (48-52).  The government was ahead in Tangney (54-46) and Hasluck (55-45), for primaries (which in the UAP's case seem too high in a few seats) see Poll Bludger for full non-paywalled primaries.  This was followed by another one with Kate Chaney said to be leading Celia Hammond (Lib) 52-48 (Hammond 38 Cheney 32 Labor 13 Green 9).  The Utting polls were impressively close to a couple of results in South Australia recently, but my past experience is that seat robopolls with n around 500 can't keep hitting marks within a couple of points for very long.

Others

news.com.au reported uComms seat polls for GetUp! showing Labor holding Gilmore 57-43, trailing 49-51 in Eden-Monaro (a pair of results that taken together are difficult to both believe), and gaining Ryan 55-45 (with the Greens not far off nabbing it instead) as well as holding Macquarie 56-44 and being competitive in Page (51-49 to Nationals).  Unfortunately while methods details are up on uComms' website, they don't post full results, though the numbers for Gilmore and some numbers for Ryan are in the article.  The Eden-Monaro sample had Labor ahead of the Liberals on primary votes but supposedly being overhauled on One Nation/UAP preferences.  The seat is hardly fertile ground for UAP and in the absence of a Nationals candidate this preference performance seems implausible.

An unknown pollster Laidlaw Campaigns was the source of a rather dubious sounding Fowler seat poll of unknown source with Kristina Keneally leading Dai Le (Liberal-associated IND) 45-38 2CP with way too many undecided.  The report also references various subsequent anti-Keneally priming questions reminiscent of the "aided vote" nonsense I've seen way too much of from the left this campaign and pending full details I don't want to go too close to it.

In another game of seat-poll Cluedo (or is it whack-a-mole), something supposedly called the Industry Association (which industry?) commissioned polls by an unknown pollster reported by Sky with the following 2PPs for Labor: Robertson 58 (!) Reid 53 Parramatta 54 Gilmore 56 Shortland 57 Hunter 51 Lindsay 43.

An Australia Institute seat poll of Higgins that I missed until checking summaries on Poll Bludger had both Labor and Greens supposedly winning the 2CP easily (Labor 55-45, Greens 54-46) but with Labor clear enough on the primary votes to win the seat.  If punters thought commissioned uComms polls were any kind of signal, Labor would probably be a lot shorter than $3.10 in Higgins.

There have also been many rumours of internal polling, most of them self-serving tactical drops by parties to tame journalists (who seem to have outnumbered the wild population this election).  There is generally no way to verify claims that any polls are genuine leaks, or in some cases that they ever existed.  Of note here was the oddly shaped 2PP collection donated to Peter van Onselen of Reid (Liberals trailing 48-49), Bennelong (43-50), Parramatta (41-50) and Warringah (47-53).  

A claim by an unnamed Liberal to have a Werriwa internal at 50-50 may have been intended to bother Labor, but it is notable that the YouGov MRP has this as a potentially relatively close seat.  There are also any number of other internal polling claims without numbers attached.

Seat Betting

Seat betting is not especially predictive but I like to keep an eye on it so as to comment on how predictive it is or isn't. Here's the current situation by my usual method of breaking down seats into groups ("close" = two candidates $3.00 and under or no candidate under $1.50):

Projected Labor gains from Coalition (not close): Reid, Boothby, Swan, Chisholm

Projected Labor gains from Coalition (close): Longman, Bass, Braddon, Pearce, Robertson (Longman and Robertson each tied on one market.)

Mixed market: Brisbane

Projected Coalition holds vs Labor (close): Bennelong, Lindsay, LaTrobe,  Leichhardt, Sturt, Deakin, Hasluck, Casey, 

Projected Coalition holds 3-cornered vs Labor/IND (close): Page, North Sydney

Projected Coalition holds 3-cornered vs Labor/Green (close): Ryan

Projected IND gains from Coalition (close): Goldstein

Projected Coalition holds vs IND (close): Mackellar, Nicholls, Curtin, Kooyong, Wentworth, Hughes

Projected Labor holds vs Coalition (close): Corangamite, Werriwa, Gilmore

Projected Labor holds vs IND (close): Fowler

Labor is favourite in a majority of seats (78.5-66.5-7).  With close seats adjusted for, the number of close Coalition seats as the government fends off opponents both red and teal makes the numbers more lopsided (79.4-62.4-9.2).  So the seat markets are not really differing much from the models.   Also, since yesterday a Labor majority has gone back to very narrowly odds-on, and Labor's median expected seat tally on seat totals markets is back to the high 70s. 

But none of the markets (not that they should), the models, the MRP or the national polls can stop Hung Parliament Club.  In the latest example of the genre, we have another Ninefax piece that sees Labor and the Coalition struggling to get to 74 seats let alone a majority, on the back of nothing whatsoever beyond nameless Liberals boosting their campaign and their polling and generously "conceding" that their opponent is unlikely to win outright. 

Updates Friday:

* YouGov Pearce: Just when I was saying there were no real pollsters in the lagoon, along comes YouGov with a 53-47 to Labor in Pearce, off raw primaries reported as Labor 43 Liberal 40 Green 4 UAP 2 One Nation 1 and 5 undecided.  This would be a big increase in the combined major party primary which would be surprising in that One Nation got 8% there last time (but its vote has crashed since in the WA state election).  These YouGov seat polls use targeted telephone polling with sample sizes in the 400s but using careful respondent selection to avoid the reweighting problems with seat robopolls.  They have had a decent record so far.