Early & Absentee Voting Megathread - Build the Freiwal

[RI]

Turnout in WA-06 and WA-08 seems pretty high as well. Not sure what the story in WA-06 is.

[RI]

A mixed county/LD turnout map:

[RI]

OK still has a ton of registered Demosaurs (although a declining number).

[RI]

I made some maps comparing the 2018 Clark County EV+MB vote to the 2016 EV+MB:

2018 Partisanship:


2016 Partisanship:


2018 Turnout Relative to 2016:


Partisan Swing, 2016-18


Turnout is clearly lagging in the Hispanic NE part of the Vegas metro (NLV, Sunrise Manor) while up in the more Republican suburbs, although these suburban areas appear more Dem than 2016.

Precinct Type	Rel. Turnout	Swing
Majority White	77.8%	R+1
Majority Black	62.2%	D+1
Majority Hispanic	57.8%	R+8
Majority Trump	85.0%	R+2
Majority Clinton	68.0%	R+4
Above Avg. Less than HS	63.9%	R+5
Above Avg. HS Grads	71.5%	R+3
Above Avg. Some Coll	76.0%	R+2
Above Avg. Bach Deg	79.3%	R+1
Above Avg. Grad Deg	77.0%	R+1
Below Avg. Income	67.0%	R+4
Above Avg. Income	79.2%	R+1

[RI]

The Hispanic share of the early/absentee electorate is up from 2014 but down from 2016 in most places, it seems.

[RI]

Latinos do not typically vote early.

What I said takes this into account as I was considering only the early/absentee vote.

[RI]

As said a million other times please don’t compare this EV to a presidential year

2016 is more relevant than 2014. Turnout is much closer to 2016 than 2014; you're asking us to compare the highest turnout midterm in living memory to one of the lowest! Not to mention 2016 was a small-scale realignment. Everything after Trump is different than before Trump.

[RI]

TargetSmart EAV swing, day before 2014 to today:

[RI]

TargetSmart EAV swing, day before 2014 to today:

How are they able to determine Indiana's swing when we have nonpartisan registration?

It's based on "modeled partisanship" which is a proprietary measure which is likely based on demographics, vote history (e.g. which primaries you vote in), social media data, consumer data, etc. I don't know their secret sauce exactly. As long as their methodology is consistent and at least somewhat accurate, it should still tell us something.

[RI]

Target Smart is VoteCastr 2.0. Only thing I trust from them is data from states with party registration.

The problem with VoteCastr was that it relied on polls which ended up being wrong. AFAIK, TargetSmart doesn't use polls.