MI-PPP: Biden +7

[Hammy]

Comparing their poll to this time in 2016 (June 28) was Clinton 46-40 with 17% other/undecided, final result was essentially 47-47 with Trump having a quarter point win.

[Hammy]

Looks good to me. A lot less undecideds than 2016 and people seem much more solid in their choices right now.

Of course, people won't be happy unless Biden has a 15% lead here come November. And even that would be discouraging because only Trump & Republicans benefit from outrageous polling errors

Well sh**t.... considering the polls were wrong by like no less than +5 for Clinton AND the MoE for MN, WI, MI and PA and then followed up by completely screwing the pooch in polling FL both in 2018 where Gillum/Nelson both consistently led by at least that margin and STILL freaking lost.... It's less giving Trump and Republicans that much leeway and more like I DO NOT trust the polling.

I am personally doubtful of the polls to some extent (especially if you have a leading candidate under 45%) but you also have to look at how and why the polls were wrong. You can't simply look at margins because that only tells you a small part of the story.

Most of the polls in 2016 did in fact underestimate Trump, and yes by quite a lot. But you had an absurd number of undecideds that you simply aren't seeing this year--not to mention they also underpolled Hillary by a few points--aside from Florida (moreso in 2018 because Gillum ran such an inept campaign) you didn't see voters going D on the polls and then going R in the election.

Republicans picked up most of the undecideds--and we're not seeing enough of that this year to make the polling difference that it made in 2016, when 44-38 leads were the norm most of the election.

[Hammy]

Neither of you are wrong about anything you said. We can point to the undecided vote of 2016 and what happened a few days leading up to the election. It still doesn't add up that Trump shattered turnout even in Democratic areas.

In a lot of cases Dems simply didn't show up. In Wisconsin, both parties were down from 2012. Michigan, Hillary lost five votes for every two Trump gained. Even Iowa and Ohio--IA saw Hillary lose three votes for every vote Trump gained, Ohio same as Michigan. In Minnesota, despite not quite flipping, Hillary lost four votes for every vote Trump gained.

If Hillary had maintained Obama's voters who didn't flip to Trump, only PA and FL would've flipped and we would've seen a 283-255 Dem win. All Biden has to do, supposing we have proper turnout, is keep all of Hillary's voters, push a few more to show up, and flip some of Trump's.

Really the only two exceptions were Pennsylvania where Republicans swamped rural areas, and Florida, where both parties were up from 2012.

[Hammy]

The polls were mostly fine in 2018, so I think firms fixed a lot of the issues that caused inaccuracy in 2016. The Michigan polls in particular were basically spot on for both the governor and senate races.

It does help that the massive number of undecideds weren't present as well--I think that more than anything likely messed up the 2016 polling.