G. Elliott Morris Profile picture
Mar 14, 2018 52 tweets 17 min read Twitter logo Read on Twitter
27% in now. Unclear if Lamb can pull this off. He’s running out of votes in Allegheny
Not sure if anyone else is seeing this or my data is messed up, but pretty interesting nugget — seems like Clinton precincts are voting a little worse for Lamb than her, but turning out much more. R precincts going more for Saccone than Trump, but voting far less.
This is a story of turnout
Sure looks blue…
FYI all: my model upload to the google doc broke somehow. trying to get it fixed but want to focus on the turnout estimates now. Will update you all if I can find a solution
Echoing @Nate_Cohn, hard to make a call on an election without live precinct data from Westmoreland county. I’m still going to run the numbers and see if they converge on a result result, but if the models can’t & we have no WML data, no way to call the race.
Still we get a pretty clear signal about the race when 51% is reporting. But so far it’s such a close contest that not having Westmoreland data really sends us back to the dark ages to depend not he AP and CNN.
*does* have some data from Westmoreland, however, so I *do* have a working model. The data just won’t upload to my google sheet. It’s Lamb +2 right now.
Model is now projecting a Lamb +0.1% victory.
With 72% reporting (including Westmoreland), Lamb vs his benchmark:

Allegheny -2%
Greene -8%
Washington EVEN
Westmoreland -5%

Trending toward that Lamb+0.1 estimate
The live model page has been updated and is now presenting data live:

87% in in PA-18. Benchmarks update:

Allegheny Lamb up by 0.8
Greene Lamb -4.8
Washington Lamb -1.4
Westmoreland Lamb -3.8
No model can tell which way this race will go it it’s projection 0.1% Lamb win (look at the confidence interval). You’ll have to wait until I have all data in. Reminder that with @DecisionDeskHQ I am providing live precinct data from Westmoreland county too.
Breaking: I’M CALLING THIS RACE FOR Satan, AKA the Westmoreland county clerk of elections
Forecast Lamb margin is up to 0.2% now and 60% confident that he wins. docs.google.com/spreadsheets/d…
Model update: hold onto your butts
#PA18 is not over yet, but every Republican elected official in the country should print this picture and frame it. This chart, showing the drop-off in GOP votes relative to Democratic turnout, is the single biggest issue for them going into November 2018

In the chat room ready to call this thing. patreon.com/notifications
As of 95% reporting, the worst Conor Lamb could do is lose by 0.9%. That would be a 19 percentage point shift from 2016. So GOP should panic either way.
Who knows recount procedure in PA? @rickhasen
We’re waiting on 16 (SIXTEEN!) precincts now.

Reporting Lamb Margin: +0.5
Projected Lamb Margin: +0.1
Worst Lamb margin: -0.5
Lamb Win Probability: 71%



1200 absentee ballots won’t be counted until tomorrow. If you want to stay up until a call, you’ll be awake a long time

12 precincts left in #PA18.

Lamb vs county-level benchmarks…

Target: +13%
Now: +14%

Target: -14%
Now: -17%

Target: -6%
Now: -6%

Target: -12%
Now: -15%
Share of votes cast in #PA18 by county (this is why lamb is winning):

Target: 43%
Now: 52%

Target: 2%
Now: 1.8%

Target: 22%
Now: 18%

Target: 33%
Now: 27%
Model is 99.9% sure of a Democratic win in #PA18 not, but (🚨) does not include absentee ballots (🚨). There’s no way to call this race without that data.

oh, that was quick
Absentee ballots just came in from Allegheny county. 800 vote lead for Lamb. I don’t know if there are enough heavy-Trump absentee ballots left in the other counties for Saccone to pull this off.
Reporting: Lamb up by 616 votes
Projection: Lamb up by 342 votes

There are 3 small precincts not yet in my data.

More importantly, Rick Saccone would need to win the remaining ~300 absentee ballots by more than 25 points to win. He only won by 13% in those three counties.
I’m not calling this for Lamb… but.. I don’t see how Saccone wins this (and my models don’t either, FWIW.)
(that’s 3k, not 300, sorry)
#BREAKING: Projection: Dem @ConorLambPA will win the #PA18 House special election by 0.2%. Trump won this seat by 20% in 2016.

Dems have flipped 39 state lege seats since 2016, 1 U.S. Senate seat, and now 1 U.S. House seat. Blue wave seems imminent.

You can see my model live here:


Why did Saccone lose? Because R precincts, despite still being pretty R, just did not turn out to vote for him.
FWIW I still give Saccone a 0.4% chance of winning. But Lamb passed the 99.5% threshold so… projection has him up.
Turnout stats & graphs incoming.
Thanks, Adam!
How did Lamb pull off a (probable) victory in a Trump +20 seat? Dems turned out at 79% of 2016 votes. GOP just 53%. Data:

Lamb votes as % of Clinton total votes:

Saccone votes as % of Trump total:

Lamb as % of Obama total:

Saccone as % of Romney:
The people who tweeted “See? No D wave.” after the TX primaries last week are very quiet tonight. Yeah. That’s b/c primaries don’t mean anything. There’s a blue wave coming, that much is clear. What is unclear is if it will be enough (it probably will).

Lamb call looking like it’ll hold steady
Here’s how Lamb’s 20.2% swing from Trump’s win in 2016 fits in with the House special elections since 2016
Thanks BenchmarkPol!
A great race tonight in #PA18. Let me just say that I wasn’t expecting 300 people on my google sheet and 1500 followers when the NYT needle died. Happy that the models went smoothly and grateful to @DecisionDeskHQ for the full precinct results!

Until next time!

Full thread:
Two things I missed tonight in the flurry or tweets:

This from @donnermaps:

This from @JMilesColeman:
I’m going to bed now, y’all.

It’s nights like these that I feel so incredibly blessed to have the best job ever. So very lucky to have been taught by stellar scholars and encouraged by my family & peers to pursue my passion. And thanks to my followers!

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with G. Elliott Morris

G. Elliott Morris Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @gelliottmorris

May 13, 2018
This is the right take (if I’m reading it correctly), but let’s talk about that article claiming the Democratic wave is receding (nymag.com/daily/intellig…). Oh boy...
Best evidence of REAL change in the wave is in actual voter behavior. If Democrats stop posting gains in special elections over time, they should worry more. Of course, they haven’t (yet?). These claims are almost all based on declining D margin in the generic ballot, which...
... is mostly an artifact of how wide you take your averages. Monthly Democratic margin in generic ballot polling has been between 7 and 9 percentage points since October of 2017. Combine with special elections, the wave looks nearly as tall as it did in March.
Read 7 tweets
Apr 25, 2018
There are TWELVE special elections tonight! One high profile election in #AZ08, a Trump +21 congressional district, and eleven (11!) in New York, where two R seats are competitive possible flips.

Coverage to follow here. <THREAD>
With results being dumped in for #NYSpecialElections, we have our first calls of the night… 3 holds for the Democrats.

NY #SD32: D hold (Clinton +88)
NY #HD74: D hold (Clinton +74)
NY #HD80: D hold (Clinton +66)
Democrats also hold NY #HD39 (uncontested) which Clinton won by 66 points in 2015. They are batting 4/11 tonight.
Read 32 tweets
Mar 13, 2018
TONIGHT is the much-hyped #pa18 race between @ConorLambPA and @Saccone4PA18. Who’s going to win? Predictions and results: [THREAD]

Before 8PM rolls around, I want to cover what the data say to expect in this contest. This thread will start off running through my predictions, will cover how I’m projecting the race, and will turn into a live blog with graphics when the votes start to come in.
There are multiple ways to predict an election. Polls, geographic data, demographic data, etc are all useful. I’m going to run through what all these measurements could predict in #pa18. Spoiler: it’s a mixed bag for both Lamb and Saccone.
Read 28 tweets
Feb 19, 2018
#New PA cong. map increases D chance of winning the House majority by 4%. They only need a 6.8% generic ballot margin now (down from 7.7) to be favored.

Diff. will be closer to 10% in Nov when uncertainty from polls is ⬇️. Redraw is a very big deal.

One major reason the new PA map is such a big gain for Democrats is that it solidifies, not just tips, PA districts in Philly burbs. Also creates one additional vulnerable R district. (Easily seen in this map made by @JMilesColeman)
My forecasting model will have new PA map in the near future. This is not just a simple switch; it’s a coding redesign, restructuring of input, & drawing new geographic files. Takes a while (for me, a student).

In the meantime, just add 3-4% to the model: bit.ly/2018_house
Read 4 tweets
Feb 19, 2018
New #PA Congressional map is pretty much the best one Democrats could have hoped for. At least 2 pickups in neutral environment, likely ability to win 11 in current D+6 environment. Post coming
PA redraw creates map that is almost entirely drawn from scratch. Huge shifts from 2016 plan all over the place. Dems favored to win 9 seats there in almost any national environment.
(big swings mostly from renumbering, but aggregate differences are clear)
Read 11 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!


0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy


3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!