Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Ryan Matsumoto Profile picture
Ryan Matsumoto
@ryanmatsumoto1
Director of Data Engineering @dcccㅣpreviously political analysis @InsideElections / engineering @google | bylines @FiveThirtyEight @UpshotNYT | Stanford '16
Puneet Kollipara Profile picture 1 subscribed
Nov 10, 2022 • 4 tweets • 2 min read
NEVADA SENATE Update:

Laxalt leads by 22,595 votes statewide
~145,000 mail ballots left to count (Clark + Washoe)

Cortez Masto needs to win these mail ballots by ~15.6% to pull ahead.

(This doesn't include mail ballots arriving through Saturday).

Positive signs for Cortez Masto:
- Apparently other mail ballot batches in Clark had been going 2-1 Democratic.
- The Culinary Union was encouraging its members to drop off mail ballots at drop boxes, so the margin with these drop box ballots may be especially Democratic.
Nov 9, 2022 • 4 tweets • 2 min read
Lots of ballots left to count in Arizona (beyond this list there's also ~180k from Pima County, plus other counties with unconfirmed totals).

If you recall from 2020, the "late early vote" favored Republicans.

Masters is still in the game.

#AZSEN

With the above confirmed county vote totals (645505) and his deficit statewide (83,262), Masters would need to carry remaining ballots by 12.9% to take the lead. It's certainly possible.

Nov 9, 2022 • 4 tweets • 3 min read
Per @Taniel's spreadsheet, Republicans have either won or are leading in 220 House districts, with Democrats winning or leading in 215 districts.

The path for Democrats to win the majority could be California - #CA13 #CA22 #CA27 #CA45 are districts to watch. You may recall that in 2018, Democrats gained considerably in competitive House districts once late mail ballots were counted.

If a similar "blue shift" happens again this year, then Democrats could win some of the seats where Republicans are currently leading.
Nov 7, 2022 • 9 tweets • 2 min read
My final 2022 Senate prediction: Republicans win the majority 51-49.

Republicans win PA and NV
Democrats win NH, AZ, and GA

Just my own personal guess - obviously many of these races could go either way! Pennsylvania: Tilt R

PA was 3 points more Republican than the nation in 2020 and this is an open seat. If the national environment is several points GOP-leaning, the fundamentals point to a GOP win. The candidate quality gap doesn't seem large enough to me to overcome that.
May 17, 2022 • 5 tweets • 1 min read
Final RealClearPolitics Pennsylvania GOP Senate Primary polling average:

Oz 26.8%
Barnette 24.2%
McCormick 19.6%

Oz is the slight favorite, though Barnette has a decent chance and McCormick has a small chance. Image The case for Oz: he's led in all 9 of the latest non-internal polls. While his favorability ratings are mediocre, that hasn't stopped him from increasing support in the last week from the low-20s to high-20s/low-30s. And the Trump endorsement could help with late-deciders.
Nov 3, 2021 • 6 tweets • 2 min read
NEW JERSEY GOVERNOR UPDATE (per @CNN results):

Phil Murphy (D) - 49.64%
Jack Ciattarelli (R) - 49.57%

1667 vote lead.

Despite the narrow lead, Democrats should be considered the heavy favorites here.

THREAD 🧵(1/6): First, the most obvious place where Murphy will pick up votes is Mercer County. It's currently at ~74% of its 2017 total votes, a clear outlier from every other county that's at 100%-135%.

This is a Biden +40 county and its remaining votes are probably absentees.

(2/6)
Nov 2, 2021 • 4 tweets • 2 min read
By all accounts, it seems like turnout is pretty strong in the heavily Democratic northern Virginia suburbs.

A good sign for Democrats when ~30% of total votes typically come from Fairfax, Prince William, Loudoun, Arlington, Alexandria, Manassas, Manassas Park, and Fairfax City. The bigger question mark is probably relative turnout in counties with high shares of Black voters.

This report out of Newport News is certainly not as strong as NoVa, but may be alright for Democrats - depends on how the next few hours go.

Nov 2, 2021 • 9 tweets • 2 min read
FINAL VIRGINIA GOVERNOR ESTIMATE (my personal opinion):

McAuliffe by 2.5 points

This is a very competitive race and both McAuliffe and Youngkin have solid chances of winning. But I think McAuliffe may have a slight edge and will win by a similar margin to his 2013 race. My guess: the northern Virginia suburbs are good enough for McAuliffe for him to survive.

2013 -> 2017:
Loudoun Co: D+4 -> D+20
Fairfax Co: D+22 -> D+37
Prince William Co: D+8 -> D+23

McAuliffe's margins here may be closer to 2017 even if the topline margin is closer to 2013.
Oct 25, 2021 • 4 tweets • 1 min read
It's important to be careful with Virginia early voting analysis, both on turnout + partisanship:
- Early voting was much more restricted in 2017. And in 2020, people were more worried about COVID.
- 50%+ will vote on Election Day
- Early voters lean D, Election Day voters lean R At its best, early voting analysis can tell you some information about *who* voted, but it won't tell you *how* they voted.

This is an important distinction because elections are not 100% about turnout, they are also about persuasion - swing voters do exist!
Oct 25, 2021 • 4 tweets • 1 min read
Virginia Governor polls (nonpartisan only) in October:

Emerson - McAuliffe +1
Christopher Newport University - McAuliffe +4
YouGov/CBS News - McAuliffe +3
Fox News - McAuliffe +5
Monmouth - McAuliffe +1.5

Average = McAuliffe +2.9% Decided to put the Monmouth poll in as D+1.5 because the director of the poll says that the preponderance of their various likely voter models gives McAuliffe a 1 or 2 point edge (they had registered voters tied and likely voter models from R+3 to D+3).

Oct 23, 2021 • 5 tweets • 2 min read
On Newsom / Latinos:
- I'm not sure that Newsom did particularly poorly with Latinos, especially comparing to 2018. Much of the underperformance in Central Valley was differential turnout IMO.
- But estimating Newsom's Latino % with precinct data has ecological fallacy problems. Differential turnout by race can make a big difference.

Kern County, for example was Trump +10 -> Yes +24.

But the early vote went from 63% White / 30% Latino to 69% White / 26% Latino. That makes a big difference, especially with high racial polarization. ImageImage
Oct 19, 2021 • 4 tweets • 2 min read
Big picture takeaways from the Virginia early vote data:
- Early turnout is lower than 2020 but higher than 2017
- Democrats are outvoting Republicans ~2:1

Neither changes my priors on the race (~3 in 4 odds D victory, median outcome McAuliffe by 4-6 points).

THREAD 🧵(1/4): Early turnout is lower than 2020 but higher than 2017.

But this makes sense - early voting rules were more restrictive in 2017, people are less worried about COVID than in 2020, and turnout is lower off-cycle.

The Fox News poll said ~50% plan to vote on election day.

(2/4)
Oct 16, 2021 • 4 tweets • 2 min read
Even if you assume the sole purpose of vaccination is to prevent death (and not cases), there's an argument for boosters on the basis that they reduce transmission, which would in turn reduce the raw number of breakthrough deaths.

Per CDC data, the rate of COVID-19 deaths among vaccinated people increased from .11/100k the week of June 20 to 1.20/100k the week of August 8.

This was mostly because the rate of overall transmission increased, not because of any waning vaccine effectiveness against death. Image
Sep 14, 2021 • 7 tweets • 1 min read
FINAL CALIFORNIA RECALL ESTIMATE (my personal opinion):

Keep Newsom - 59%
Remove Newsom - 41%

(1/7) Here are my county-by-county estimates for top 25 counties by population (hopefully some of them pan out!):

Los Angeles: 68% Keep 32% Remove
San Diego: 54% Keep 46% Remove
Orange: 48% Keep 52% Remove
Riverside: 46% Keep 54% Remove
San Bernardino: 49% Keep 51% Remove

(2/7)
Aug 31, 2021 • 4 tweets • 2 min read
California Returned Ballots Update (with a comparison to the equivalent point in 2020):

Aug 30, 2021 - 3,796,496 votes
54.48% D / 23.51% R / 22.02% I
D+30.97%

Oct 19, 2020 - 3,988,502 votes
56.21% D / 21.07% R / 22.72% I
D+35.14% Takeaways from CA returned ballots data:
- Turnout is high, almost keeping pace with returned ballots at an equivalent point in 2020.
- GOP may have a small enthusiasm advantage (4% change in D-R right now), but will need to flip many Biden voters to win.

Aug 29, 2021 • 4 tweets • 1 min read
Current breakdown of returned ballots in CA is 54% D / 24% R / 22% I.

At an equivalent point in 2020 (Saturday 10/17), the breakdown was 57% D / 21% R / 22% I.

Consistent with a small R enthusiasm advantage relative to 2020 (note: this is NOT me saying the recall wins). This is not to say that the recall will prevail (we currently have it at Likely D), but just that current returns suggest an electorate that may ultimately be a few points more favorable for Republicans than November 2020.

To win, Republicans need to flip a LOT of Biden voters.
Aug 25, 2021 • 4 tweets • 1 min read
Although initial returned ballots data in California (57% D, 22% I, 21% R) seems decent for Democrats, I think there's a possibility that current data may be a bit skewed and may actually be consistent with Republicans having a bit of an enthusiasm advantage.

THREAD (1/4) 🧵 It looks like there are many counties in California that are either not reporting or underreporting returned ballots data - these counties tend to be more rural and Republican-leaning than the state.

For example, Kern County (Trump +10) has barely reported anything yet.

(2/4) Image
Aug 23, 2021 • 4 tweets • 2 min read
The current partisan breakdown of returned ballots in the California Recall is about the same as it was at an equivalent point in 2020:

2021 Recall (8/23): 58% D, 20% R, 22% Other
2020 General Election (10/12): 57% D, 21% R, 22% Other

Of course, important to note that Republicans in California are both (i) more likely to return absentee ballots later on and (ii) vote in-person on Election Day.

Here's what the data trend looked like leading up to the 2020 general election:

Jul 7, 2021 • 4 tweets • 2 min read
Overall I'm pretty happy with my final call that Adams had a slight advantage!

Originally had Garcia favored, but glad I flipped when I looked at the data and realized that the pro-Garcia absentee skew was probably not going to be as strong as my priors.

My baseline assumption that the absentee vote would be roughly equivalent to the early vote panned out really well.

Adams led Garcia for first round preferences among early voters by 3.58%.

For absentee/provisional voters it was Adams +2.53%.

Jul 6, 2021 • 5 tweets • 2 min read
Here's where I see the NYC Mayor race heading into tomorrow:
- It's a toss-up between Adams and Garcia
- Garcia is very likely to stay ahead of Wiley
- Garcia should win the absentee vote (final round) - the question is whether or not her margin is big enough (she needs +16-17%) There's a lot of uncertainty, but I'd say Adams may have a very slight advantage.

Potential Final Round Vote Shares:
Adams 50.2%
Garcia 49.8%
Jul 4, 2021 • 7 tweets • 2 min read
Gonna take a moment to revisit this scatterplot I made a while back and explain why it probably *overstates* Garcia's strength with absentee voters.

There are some statistical / demographic quirks that make this interesting.

THREAD 🧵(1/7)

Initially I figured that Assembly Districts would be a good comparison point because they are supposed to be evenly populated.

But there are actually *gigantic* turnout differences by Assembly District.

ADs 52/57/67 have 30k+ votes while ADs 47/49/62 have <6.5k votes.

(2/7)