GA-6 Special election discussion thread

[GeorgiaModerate]

Some thoughts from Nate Cohn on Twitter:

You must be logged in to read this quote.

You must be logged in to read this quote.

You must be logged in to read this quote.

[GeorgiaModerate]

For reference:

In April (all votes), Fulton was 45%, Cobb 32% and Dekalb 23%.

In early vote, Fulton was 53%, Cobb 26% and Dekalb 21%.

Early vote through yesterday:

Cobb 13199 (15.8%)
DeKalb 20622 (24.7%)
Fulton 49581 (59.4%)
Total 83402

Through today:

Cobb 15283 (16.3%)
DeKalb 23334 (24.8%)
Fulton 55319 (58.9%)
Total 93936

Keep in mind that Cobb has fewer early voting locations during the runoff than the other two counties.

I've also heard rumors that (some of?) the counties have a big backlog of mail ballots that haven't yet been processed.

[GeorgiaModerate]

It was likely at some point Cobb would pick some up but being in thrid will hurt Handel

Cobb will probably pick up from here.  They've had only one early location open; a second was opened today and will remain open.  DeKalb has had 5 locations for the whole period except last Saturday.  Fulton has 6 locations, all open for the whole period.

[GeorgiaModerate]

Good thread from Nate Cohn about the "no party" early vote: https://twitter.com/Nate_Cohn/status/873938254793510912

[GeorgiaModerate]

538: Too close to call, but Ossoff has a slight edge: https://fivethirtyeight.com/features/democrats-have-a-slight-edge-in-the-georgia-6/

[GeorgiaModerate]

EV through yesterday.  As expected, Cobb is picking up a bit now that they have a second location open.

Cobb 18097 (17.6%)
DeKalb 24998 (24.3%)
Fulton 59695 (58.1%)
Total 102790

I played around some with the EV data through Saturday and made a few observations.   A big caveat: this is the first time I've ever tried to look at voting data down to the precinct level.  Here's what I found:

Obviously EV is way up in the runoff compared to the first round.  But it's not consistently up across the district: some precincts are way up, some are a little bit up, and a few are still lagging behind the first round.  (This is comparing the runoff EV through Saturday to the total EV in the first round, so the lagging precincts are still likely to exceed their first round totals.)

So can we say anything about which precincts are up more, versus those up less?  Perhaps.  I plotted the increase in EV against the D share of the total D/R vote in the first round (ignoring the few votes for independent candidates).  This showed a definite trend such that a greater increase in EV was correlated to a degree with higher D% in the first round.

BUT! you say...D-leaning DeKalb has more EV locations this time, while R-leaning Cobb has fewer, so the above trend is only to be expected.  To which I respond: well, this is a very good point.  However, these counties do not uniformly lean toward their dominant party.  All three counties have some precincts that voted very D, some that voted very R, and some in the middle.

So I looked at each county individually.  And I found that considering DeKalb only, the trend held up: within the county, the precincts with the higher D% in the first round tend to have higher increases in EV.  Cobb has the same trend, but to a lesser degree.  The Fulton data is messy; there doesn't appear to be much of a correlation, perhaps a slightly R-favoring one.

To further complicate this, not all precincts are created equal; some are very large and some are tiny.  And since overall vote total is the only thing that counts, it's more beneficial for a candidate to get 55% of 1000 votes than 90% of 100 votes.  So I tried replotting the data as a bubble chart with sizes representing the total vote in the first round.  This didn't really make anything clearer.

What I think this is telling me is that the increase in EV turnout so far is greater in D-leaning areas than R-leaning areas in DeKalb; the same may be true to a lesser degree in Cobb; and Fulton is anyone's guess.  But my confidence level in this is not at all high.

[GeorgiaModerate]

Still Lean D, Ossoff wins by 2-4 points.

[GeorgiaModerate]

Miss Enormous Tentacles is not going to win. It's important to remember that Ossoff over performed polling in the first round.

It's also important to remember that this is a tossup, and anybody who claims to know a definite and airtight result has had one line too many. This is Tilt D at best, and Tilt R at the worst. That is why I think this will be a close 52-48 for Handel, BUT I am not exclusively and definately calling it for her. This is very much in play, and either side could win.

This is spot on, although I think it's Ossoff who will win 52-48.  But I wouldn't be surprised with a win by either.

[GeorgiaModerate]

Early vote through today:

Cobb 20088 (18.1%)
DeKalb 26394 (23.7%)
Fulton 64748 (58.2%)
Total 111230

[GeorgiaModerate]

This morning I heard an anti-AHCA radio ad from AARP.  At the end of the ad, it mentioned the GA-6 election and encouraged listeners to check out the candidates' positions on AHCA at aarp.org/gavote,  which redirects to http://action.aarp.org/site/PageNavigator/Georgia6Vote2017.html. 

[GeorgiaModerate]

Tempers flaring among early voters: http://www.myajc.com/news/state--regional-govt--politics/voters-6th-district-losing-manners-election-becomes-aggressive/qzRKcKC1RaMy2iJQdD5aOM/

This kind of behavior is just sad.

[GeorgiaModerate]

GOP pollsters Trafalgar Group have it: Ossoff 50 Handel 47

There polls were super Republican in 2016, but most ended up being pretty accurate.

Can we spot calling them super Republican when they were the most accurate state level polls?

Also, *their

Well, the actual results were pretty super for Republicans.

[GeorgiaModerate]

GOP pollsters Trafalgar Group have it: Ossoff 50 Handel 47

There polls were super Republican in 2016, but most ended up being pretty accurate.

Ossoff's support also seems to be a tiny bit firmer than Handel's; she has a larger percentage of leaners than he does.

[GeorgiaModerate]

Can anyone please tell me how margin of error is calculated? Thank you in advance.

http://www.dummies.com/education/math/statistics/how-to-calculate-the-margin-of-error-for-a-sample-proportion/ is a pretty good explanation.

[GeorgiaModerate]

They're sending me to Georgia tomorrow until the election, this should be interesting.

Please continue to send us dispatches from the front lines.

[GeorgiaModerate]

FWIW, that poll also has Ossoff winning the early vote 56-44.

Would this be...good? Dems won the early vote in April by 30 points, but presumably a greater share of the electorate has voted early this time.

I think it's believable, and Republicans seem to be cannibalizing a lot of their election day vote this time

I do too.  With the large number of Republican candidates in the first round, many of their voters probably waited until Election Day to make up their minds.  That factor isn't present in the runoff.  There was a comment by one of the well-known analysts (sorry, don't remember who) a few weeks ago that he thought Handel needed at least 46 or 47% of the early vote.

[GeorgiaModerate]

GOP pollsters Trafalgar Group have it: Ossoff 50 Handel 47

There polls were super Republican in 2016, but most ended up being pretty accurate.

Can we spot calling them super Republican when they were the most accurate state level polls?

Also, *their

There's a difference between being right coincidentally and actually being right. If I correctly calculate the odds of a quarter landing on heads three times in a row, then that makes me right. If I correctly guess that the quarter will land on heads three times in a row, then that makes me lucky.

That's why there are so many "pollsters" in each election cycle who get bestowed with the title of "[one of the] most accurate pollsters of the election cycle", and then proceed to being hilariously wrong in future ones (just like they were in the past). In most cases, their methodologies haven't changed; they just lucked out. If a pollster predicts in every election that the results are going to be more GOP/DEM than other pollsters are showing, and the results in one election cycle do in fact end up being more GOP/DEM than most expected, then voila: "most accurate pollster"!

TL;DR: broken clock is right twice a day; Trafalgar is a sh**t pollster

Great point Adam. I was just going to ask isn't Trafagular a bad pollster in general earlier.

538 gives them a C rating based on 12 polls.

[GeorgiaModerate]

For reference:

In April (all votes), Fulton was 45%, Cobb 32% and Dekalb 23%.

In early vote, Fulton was 53%, Cobb 26% and Dekalb 21%.

Through today:

Cobb 22479 (18.8%)
DeKalb 28159 (23.6%)
Fulton 68907 (57.6%)
Total 119545

[GeorgiaModerate]

For reference:

In April (all votes), Fulton was 45%, Cobb 32% and Dekalb 23%.

In early vote, Fulton was 53%, Cobb 26% and Dekalb 21%.

Through today:

Cobb 22479 (18.8%)
DeKalb 28159 (23.6%)
Fulton 68907 (57.6%)
Total 119545

Also for reference: the distribution of registered voters in the district is approximately Cobb 29%, DeKalb 23%, Fulton 48%.

[GeorgiaModerate]

Tom Bonier did a breakdown of EV an Ossoff seems to be getting more black and female irregular voters then first round https://mobile.twitter.com/tbonier/status/875341387763519488

I'm a little skeptical of his analysis because he says 124K votes have been cast so far, but as of last night's absentee voter file update the actual number of accepted ballots was 119545.

[GeorgiaModerate]

Crystal Ball keeps it a tossup, but says that all things considered, one would rather be in Ossoff's position than Handel's.  http://www.centerforpolitics.org/crystalball/articles/ga-6-special-still-on-the-razors-edge/

[GeorgiaModerate]

Suspicious package shuts down street in Handel's neighborhood (ongoing): http://www.wsbtv.com/news/local/north-fulton-county/breaking-suspicious-package-shuts-down-street-where-karen-handel-lives/533890723

And on the other side:

You must be logged in to read this quote.

This behavior is unacceptable in a civilized society!

[GeorgiaModerate]

Current update from WSB:

You must be logged in to read this quote.

[GeorgiaModerate]

For reference:

In April (all votes), Fulton was 45%, Cobb 32% and Dekalb 23%.

In early vote, Fulton was 53%, Cobb 26% and Dekalb 21%.

Also for reference: the distribution of registered voters in the district is approximately Cobb 29%, DeKalb 23%, Fulton 48%.

Through today:

Cobb 24870 (19.4%)
DeKalb 29783 (23.3%)
Fulton 73316 (57.3%)
Total 127969

[GeorgiaModerate]

For reference:

In April (all votes), Fulton was 45%, Cobb 32% and Dekalb 23%.

In early vote, Fulton was 53%, Cobb 26% and Dekalb 21%.

Also for reference: the distribution of registered voters in the district is approximately Cobb 29%, DeKalb 23%, Fulton 48%.

Through today:

Cobb 24870 (19.4%)
DeKalb 29783 (23.3%)
Fulton 73316 (57.3%)
Total 127969

Am I wrong or doesn't in-person Ev end tomorrow? That means Cobb will be in third 

You are correct.