The 20% is sourced/cited...One in five of early votes came from new voters (those that did not vote in 2016), per WaPo link. I'm not sure what readers would have to assume?
Combine that with the majority of early votes arriving from Dems (per Election Project), and that's the correlation which is inferred. Of course, a household registered as Dem isn't necessarily required to vote blue, which is also conceded. I was not attempting to surreptitiously hide from stats, but may not have thread the needle clearly.
And of course this is all speculative. The entire piece is. As is every poll. We will all find out what happens sometime after Nov 3.
I'm not sure I'd agree with all your stated Trump accomplishments, and the Marist poll suggests the same. Disapproval outweighs approval rating, overall. But that's just one poll. Sure, it is likely to source polls with findings that counter Marist.
And his approval rating within the party? - 95% is to be expected. I'd be more shocked if he didn't. But, that's how we roll here, unfortunately. Extremely partisan.
Thanks for the feedback and discussion.