ME-Bangor Daily News/Digital Research: Gideon +1/+3 with RCV, Pingree +34, Golden +18

[TiltsAreUnderrated]

There are a couple of slight errors with this poll. "Will not vote" is listed at 1% despite being closer to 0% of the sample and 4 Gideon and Collins voters select "prefer to to say" as their second choice, yet only 3 voters' second choice is "prefer not to say."

https://drive.google.com/file/d/1rgOpdiCIcbEBjnt2pVvQO7ztkbfsNB90/view

Critical Insights
September 25 - October 4
Changes with July 28-August 9 poll; trends calculated pre-rounding

Statewide (Senate)
466 likely voters
MoE: 4.4% (apparently for likely voters, judging by the way the article reads)

Initial ballot
Gideon 44% (+1)
Collins 43% (+5)
Savage 3% (-2)
Linn 1% (not included in previous poll)
Prefer not to say 1% (n/c)
Someone else 1% (not included in previous poll)
Will not vote for Senator 0% (n/c)
Not sure 7% (-6)

With RCV (taking the second choices only of Linn/Savage/"someone else" voters):
Gideon 46%
Collins 43%
Do not have a second choice 1%
Prefer not to say 1%
Savage 1%
Linn 0% (but some voters)
Someone else 0% (but some voters)
Not sure 7%

ME-01 (House)
232 likely voters

Pingree 58% (+8)
Allen 24% (+1)
Someone else 3% (n/c)
Prefer not to say 2% (n/c)
Will not vote 1% (+1 from 0% with no voters)
Not sure 12% (-10)

With RCV
ME-02 (House)
234 likely voters

Golden 52% (+9)
Crafts 34% (+1)
Someone else 2% (n/c)
Prefer not to say 1% (+1 from 0% with no voters)
Will not vote 0% (but some voters)
Not sure 11% (-4)

[TiltsAreUnderrated]

Wait, they did RCV on the prez race but not Senate?

They did RCV directly on neither (in terms of published toplines). They just asked for if people did/didn't have second preferences (if they'd already picked a candidate) and you can work out from the cross tabs how non-Biden/Trump/Gideon/Collins voters split.

It took me a little while.

We're dealing with small sample sizes, but I found it surprising that Trump gained more than Biden from RCV and Collins gained zero second preferences.

[TiltsAreUnderrated]

Update: their last poll also had second preferences, but they did not publish the crosstabs for those amongst those who said absolutely or very likely, so trends would have to be compared to the larger previous sample.