Thread by @kevinbonham

#ResolvePM federal ALP 42 L-NP 28 Green 12 ON 6 UAP 1 IND 9 (likely overstated) other 2. My 2PP estimate 61.5 to ALP (+2.1) Resolve's regular polls skew to Labor cf other polling (this is not true of their polls late in campaigns which use different methods.)

61.5 is the highest 2PP conversion I have for any poll in this term just beating 61.3 also from Resolve in Aug 2022. Nobody else has been above 57.4 on my conversions or 58,5 as a pubished 2PP. (Resolve does not normally release 2PPs and didn't for this poll.)

#ResolvePM Dutton net performance "in recent weeks" -28 down from -11 a month ago. Treat with caution because of the very low L-NP primary in this poll.

After adjusting for Resolve's regular series' massive house effect this is still a woeful poll for the Coalition but a 42-28 primary lead from Resolve is like a, say, 39-32 lead from anyone else.

Resolve does two things differently as elections approach. Firstly listing parties that are actually on the ballot in specific seats (this fixes the too-high IND vote) and secondly the last poll is a phone/online hybrid with a slightly larger sample size.

As I've said before Nine should either require Resolve to supply 2PPs for its regular polling or else have someone calculate them and publish them. That would make it obvious that regular Resolve has a major house effect.

I have received some correspondence that among other things claims that Resolve's high numbers for Labor are simply a product of forced choice on voting intention and that forced choice pushes undecideds towards the most popular option. I don't agree.

If that was the case similar numbers would have been seen in the former ReachTEL federal series which also employed forced choice on voting intention. (Respondent could be initially undecided but had to pick an option at the second pass). They were not.

It's also notable that Resolve *did not* show high numbers for Labor in its non-campaign-period polling before the 2022 election. It was generally closer than other polls.

I don't think the reason for Resolve behaving differently between elections and at elections (aside from the IND vote issue) is obvious. I certainly don't think it is deliberate. I don't believe the methods differences explain it.