Share this page!

John Pfaff Profile picture
Twitter logo
Jan 28 • 4 tweets • 1 min read
Since the text of the Constitution doesn’t say that you need to be textualist, I think the next nominee should reject GOP questions abt “sticking to the text” by saying textualism requires her to reject textualism and then claim the title of invincible Metatextualist. Image
Ok, two things.

1. I thought I was going to need a root canal, but turns out I need two wisdom teeth yanked. Both procedures sound far far better than what I imagine SCOTUS discourse will be for weeks or months to come.
2. Apparently Taika Waititi is directing a cinematic The Incal. Can’t wait. Also can’t wait for Luc Bresson to sue them for copyright violation of the 5th Element.

theverge.com/platform/amp/2…
(A bunch of you are wondering if I just wrote a Dune-ripped-off-Star-Wars tweet, aren’t you? It’s… not clear, is it?)

• • •

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

Keep Current with John Pfaff

John Pfaff 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!

PDF

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 @JohnFPfaff

John Pfaff Profile picture
Jan 30
I think—and, honestly, I’m not really joking—we’ve hit the point where major criminology journals need to declare a one (two? five?) yr moratorium on papers using admin data, and focus on a host of papers laying out how, exactly, we can use that data.
And I mean the major journals, not the lower-ranked ones. This isn’t some second-tier methodological issue. This gets at the heart of every quantitative paper using admin data. It’s a first-most toppest-tied issue.

Our admin data is… just a giant minefield.
Like, the UCR flaws are well known, but as you dig down into them, they just get worser and worser, like some fractal of fail.

We need to identify what data seems more reliable, which is less, how we can carefully impute across the gaps, how we can cross-validate, etc etc etc.
Read 5 tweets
John Pfaff Profile picture
Jan 30
This looks like an interesting paper, except….

It is REALLY HARD to use admin race data.

There’s a reason the BJS report never breaks out racial data by state.

This isn’t the only paper out there doing this, so it’s an important point!
The BJS gets its national racial estimates by combining admin data at the state level with several national-level surveys that gather self-reported race data. It then runs all these things through a fairy complex algorithm to get a NATIONAL estimate.
The challenge is that admin race data isn’t based on self-reports, but rather on how correctional staff choose to record things.

In this case, I’d bet how race is recorded correlates with North vs South… or at least the risk is really high.
Read 6 tweets
John Pfaff Profile picture
Jan 28
Has anyone come across a paper that rigorously does either of these:

1. Compare the various marginal contributions of crime interventions ($1 for police cuts crime by x, $1 for drug treatment by y)?

We can't think abt opportunity costs without this.

And...
2. Compares how marginal effects of a crime intervention vary across ... other stuff? Like: $1 on drug treatment cuts crime by x in a city with a big per capita police force, and by y with a small one.

None of this stuff is one-size-fits-all, and feel like baselines are huge.
I feel like 95% of the debate abt crime is about which is *the* right response.

But it's all about marginal opportunity cost tradeoffs against different background conditions in different places.

Which, I know:
Read 5 tweets
John Pfaff Profile picture
Jan 12
This is a really intriguing article, though not for the reasons I think it was intended.

I think it demonstrates how impoverished our thinking about what accountability is or can be, and how that inhibits efforts to push for deeper, more substantial changes.
The crime here is serious, and that no one died or was seriously hurt is fortuitous.

Trent set fire to his ex's apartment at night while she slept, out of anger. She and her roommates escaped, but... that doesn't always happen.

Ends up serving 6 months.
Here's what makes it so intriguing.

Trent himself admits he did something wrong. Trent himself feels like he needs to atone more, take more accountability.

But our options (conceptually) are basically: prison, or nothing.
Read 6 tweets
John Pfaff Profile picture
Jan 11
Unless I'm way off, the instances where Bragg plans to (systematically) plead around the "unless required by law" mand-mins for all viol felonies higher than Class E are:

1. A narrow set of commercial robberies where no one is at risk of real injury.

That's it.
(b) isn't a violent crime (not how we count weapons charges), and doesn't involve guns. Anyway, in 2019 a grand total of 55 ppl were sent to prison from Manhattan for 265.02.

(c) and (d) lower the felony to D--still mandatory prison time.

(e) is about drug possession.
"He's not going to send anyone to prison for violence anymore!"*

* For one narrow category of case, under some situations, and supervisors can override for "extraordinary reasons," which includes prior criminal history.
Read 5 tweets
John Pfaff Profile picture
Jan 10
Gotta say, the more I think abt it, the angrier I get abt the CDC saying life expectancy at birth fell by 1.8 years from 2019 to 2020.

We had a huge spike in mortality, yes. And one in morbidity that will def shorten lives in the future.

But the number here is total gibberish.
To start, Covid might have ~0 impact on expectancy AT BIRTH. Maybe it will—maybe kids exposed at birth or in utero will be sicker. Or maybe it made them longer, by jumpstarting mRNA therapies.

But a static accounting based on 2020 data is … beyond meaningless.
It’s faux quantification of a health shock not easily reduced to a single headline number. Easily repeated, seemingly meaningful… and completely meaningless—at best.
Read 5 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

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

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

Like this author's thread?

Get emails when @JohnFPfaff has new unrolls!

Check out other threads from @JohnFPfaff!

Read all threads by @JohnFPfaff

:(