Poll Roundup: No, Labor's Honeymoon Is Not Over Or Anything Like It

2PP Cross-Poll Average 56.1% To Labor
Labor would win an election "held now" with an increased majority (approx 92 seats)

Today's quick Poll Roundup is provoked by one of the more curious and spurious interpretations by an actual pollster that I've seen.  Today's Resolve poll had Labor leading 40-31 on primary votes, which while down on December and January is still the same 9-point lead they held in September and twice in October and off a one point higher primary vote than those.  It would come out to a 2PP of around 58-42 to Labor*, which if repeated at an election would result in Labor gaining about 20 seats from the Coalition and winning in one of the biggest landslides ever seen with about 97 seats to the Coalition's 38.  Yet Resolve's Jim Reed was quoted as declaring that “It looks like Albanese and Labor’s election honeymoon is over,”.  The main justification for this was that Labor had come down slightly from even greater highs in polls taken over the holiday season.  The other was movements of a few points from Labor and unsure to the Coalition across a raft of attribute and leadership polling measures, many of which were within the poll's in-theory margin of error and many of which still showed Labor way ahead anyway.  (The accompanying article says jobs was one of Labor's biggest setback areas, but on jobs the Coalition's gains came entirely from "someone else" and "undecided").  

A polling honeymoon doesn't end just because there are slight wobbles up and down as a result of statistical noise or minor issues, especially not when the wobbles come from implausible highs.  If Resolve actually published regular two-party preferred estimates it would be immediately obvious that this poll was showing Labor way in front and that they were continuing to record honeymoon polling.  The debate about the need to publish 2PP estimates regularly was run and won in the case of Newspoll over 20 years ago when Newspoll finally started routinely publishing 2PPs after Peter Brent pointed out that its focus on primary votes was mismeasuring the contest.  It is past time for The Age and SMH to require Resolve to produce credible 2PP estimates for every poll, or alternatively to obtain them from elsewhere, instead of blatantly misleading readers about the meaning of the results of polls that said mastheads are covering.  Heck, I can do it for $5 per poll, or $500000 per poll if I have to go through their onboarding system again. 

It's also time for wider agreement on rough standards for declaring that a polling honeymoon (i) exists and (ii) is finished.  In my view a honeymoon exists if there is a largely continuing post-election run of aggregated polling with a government 54-46 or more ahead, or for less impressive baselines than 2022 with a government at least 2% above its election result.  In the case of a leadership change honeymoon, the latter standard might apply with respect to the run of polling preceding the change, provided that the new leader is competitive.  A honeymoon ends when this sort of polling stops, not on the basis of a single reading but over multiple polls from more than one company over at least a month, and taking house effects into account.

A further minor point, but one that has been made so many times before, is that when changes between two polls are within in-theory margins of error, there is no evidence that voters have changed their mind about anything.  This applies especially to the two-point shifts in primary votes back from a 42-29 lead that no sane person believed was real anyway.  And it applies especially in February, since there is strong evidence (at least in Newspoll) that there is a degree of seasonality in polling; both government 2PPs and Prime Ministerial ratings often go down a bit at this time of year.  (That's an old article, but I've checked since and it is still a thing.)

As for Albanese's ratings, his net rating of +25 (56-30; NB rounding) is his lowest from Resolve so far, but not by that much; he was 57-28 in late October.  Albanese's lead over Peter Dutton on the skewed Preferred Prime Minister metric was also his lowest so far but it was still 55-23 (new leaders take a while to get established on this indicator.)  If this is the end of the honeymoon, Anthony Albanese will be hoping his honeymoon remains this much over forever.  

* 57.9 if lumping IND and Others, 58.6 treating IND as IND.  I have gone off doing the latter as I believe the excess IND votes in Resolve between election campaigns are likely to come mostly from divisions where Independents often do not run, and where their preferences would flow more weakly to Labor if they did.

Essential

Essential also finds Albanese's net ratings starting to come down; they have him on net +19 (53-34) compared with net +24 last month and net ratings mostly in the low +30s since taking office last year.  However, their other measure of Albanese's approval, their rather odd net favourability score, saw him improve slightly from +23 to +25 since November.

Labor holds a strong 52-41 2PP+ lead (equivalent to about 55-45 the way other pollsters measure 2PPs) off raw primaries of Labor 33 L-NP 30 Green 14 One Nation 6 UAP 3 others 8 undecided 8.  I get this as 56-44 to Labor by last-election preferences, down from 57.4 a fortnight ago.  Late last year Essential made methods changes that included switching from weighting by reported party identification to weighting by reported past vote.  What difference this would make is unclear to me but for whatever reason their Greens reading has jumped since they did it, including an implausible raw 17% a fortnight ago.

Morgan

Morgan has lately been releasing 2PPs and basic primary figures (Labor/Coalition/Green/Others) on a weekly basis, most recently 56.5 to Labor (37-33.5-12.5-17).  However there hasn't been a clear statement of how they are now calculating their 2PP estimates.  In the leadup to the 2022 election they switched to using last-election preferences, but their recent 2PPs show a persistent skew to Labor compared to what I estimate off 2022 preferences.  Off recent readings of 58.5,57,58.5 and 56.5 I got last-election 2PPs of 56.6, 55.5, 56.1 and currently 55.6.  In any case, Morgan too continues to show a large lead.

(Update: Morgan has a fresh poll with a 2PP of supposedly 58.5 but which I get as 56.0 off the primaries (37-33-13-17).

Newspoll

The last Newspoll from two weeks ago is relatively old news now but in any case it also showed Labor doing fine with a 55-45 lead off primaries of Labor 38 Coalition 34 Greens 11 One Nation 6 UAP 1 others 10.  It also had Albanese on his lowest netsat so far (but still a healthy +24) while Peter Dutton was also on his lowest so far (net -10).  

(Incidentally the non-transparent Morning Consult tracker is not showing much contraction in Albanese's ratings but has had him a little lower lately than at most times late last year.)

Overall it is absolutely clear that the nine-month-old Albanese Labor Government continues to be recording honeymoon phase polling, despite some easing in leadership ratings, and that any statement to the contrary is absurd.  Long honeymoons have been recorded by four of the last six new federal governments.  This one will end sometime for good and interesting reasons like they all do.  Impatience to declare that point to have come might make a more dramatic story in the short term, but in the long term will simply damage the credibility of the outlet claiming it, and in the long run won't advantage anybody.