Crémieux Profile picture
May 21, 2023 10 tweets 4 min read Read on X
Changes to the Bar exam so far don't seem to have made a dent in the race differences in performance on it.

How large are these differences? In the latest year, 2022, the Black-White gap in California Bar pass rates was 0.97 SDs.

And this matters because we expect a threshold Image
to practice (the Bar exam) to reduce subsequent performance gaps.

But how much should those gaps be reduced?

Let's simulate to find out!

@_twolfram recently provided an estimate for British lawyers' IQs of 110.31: sciencedirect.com/science/articl….

Let's go ahead and assume
there's a threshold that's 0.67 SDs (10 points) above the higher-performing of two groups with equal variances who are separated by 0.97 d.

With simulated group sizes of one million persons each, the mean differences decline, and the SDs do too. The new gap is 0.412 d. Image
But we know that the 0.97 d gap is an underestimate due to range restriction.

Using MBE scores, it looks like the unrestricted gap should be more like 1.22 d. That leaves us with a 0.537 d gap above the threshold.

Do we have subsequent performance measures?

Yes! We have three:
- Complaints made against attorneys
- Probations
- Disbarments

For men, the gaps, in order, are 0.576, 0.513, and 0.564 d. For women, the gaps are 0.576, 0.286, and 0.286 d.

Men fit expectations and women apparently needed less discipline. Source: https://board.calba...
These gaps probably replicate nationally.

For example, here are Texas pass rates from 2004 - a 0.961 d Black-White first-pass gap. The 2006 update to these figures raised the gap to 0.969 d. Source: https://ble.texas.g...Image
Those figures are basically in line with LSAC's national study of Bar exam pass rates. Image
And those are basically in line with New York's gaps. Image
And this should probably be expected, since tests measure the same things. Image
Since all of the people included in these statistics went to ABA-accredited schools, they all had the opportunity to learn what was required to perform well on these tests.

But just like the Step examinations for medical doctors, the gaps on the tests and in real life remain.

• • •

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

Keep Current with Crémieux

Crémieux 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 @cremieuxrecueil

Nov 25
After this article came out, several people responded, alleging that a cultural model made more sense.

Clark has a point-by-point response🧵

Let's start with the first thing: parent-child and sibling correlations in status measures are identical—hard to explain culturally! Image
The reason this is hard to explain has to do with the fact that kids objectively have more similar environments to one another than to their parents.

In fact, for a cultural theory to recapitulate regression to the mean across generations, these things would need to differ! Image
Another fact that speaks against a cultural explanation is that the length of contact between fathers and sons doesn't matter for how correlated they are in status.

We can see this by leveraging the ages parents die at relative to said sons. Image
Read 10 tweets
Nov 24
The idea:

The internet gives everyone access to unlimited information, learning tools, and the new digital economy, so One Laptop Per Child should have major benefits.

The reality:

Another study just failed to find effects on academic performance. Image
This is one of those findings that's so much more damning than it at first appears.

The reason being, laptop access genuinely provides people with more information than was available to any kid at any previous generation in history.

If access was the issue, this resolves it. Image
And yet, nothing happens

This implementation of the program was more limited than other ones that we've already seen evaluations for though. The laptops were not Windows-based and didn't have internet, so no games, but non-infinite info too

Still huge access improvement though
Read 4 tweets
Nov 22
What is the effect of having a parent get locked away on a kids' own risk of eventually committing crime?

As it turns out, basically nil.

Having a mother or a father locked away doesn't significantly increase risk, and indeed, may reduce it if it happens at an early age. Image
This is relative to no incarceration, so the result should be interpreted as... pretty shocking!

Similarly, we can look at the effects of longer versus shorter parental sentences.

There's seemingly little effect of the length of time parents are incarcerated for. Image
Compare those within-family estimates from above with these between-family results from the same study, period, cohort, etc.

Notice: the apparent 'effects' between families are significant stratified in the same way.

That's an important distinction! Image
Read 5 tweets
Nov 20
Why?

I doubt the answer is weight loss. Consider 2 other drugs for diabetes: DPP-4 and SGLT-2 inhibitors

GLP-1RAs are associated with less Alzheimer's vs. DPP-4is:

But not SGLT-2is:

Neither generates much weight loss, but SGLT-2is match GLP-1RAs on glycemic benefitsImage
Image
So, at least in this propensity score- or age-matched data, there's no reason to chalk the benefit up to the weight loss effects.

This is a hint though, not definitive. Another hint is that benefits were observed in short trials, meaning likely before significant weight loss.
We can be doubly certain about that last hint because diabetics tend to lose less weight than non-diabetics, and all of the observed benefit has so far been observed in diabetic cohorts, not non-diabetic ones (though those directionally show benefits).

Anyway, trials needed!
Read 4 tweets
Nov 20
El Salvador is now a safe country

The reason why should teach us something about commitment

The government there has previously attempted crackdowns twice in the form of mano dura—hard hand—, but they failed because they didn't hit criminals hard enough

Then Bukele really didImage
In fact, previous attempts backfired compared to periods in which the government made truces with the gangs.

The government cracking down a little bit actually appeared to make gangs angrier!

You'd have been in your right to conclude 'tough on crime fails', but you'd be wrong.
You have to *actually* enforce the law or policy won't work. Same story with three-strike laws, or any other measure

Incidentally, when did the gang problems begin for El Salvador? When the U.S. exported gang members to it

This was bad for El Salvador:

Read 5 tweets
Nov 18
Diets that restrict carbohydrate consumption lead to improved blood sugar and insulin levels, as well as reduced insulin resistance.

Additionally, they're good or neutral for the liver and kidneys, and they don't affect the metabolic rate. Image
Carbohydrate isn't the only thing that affects glycemic parameters.

So does fat!

So, for example, if you replace 5% of dietary calories from saturated fat with PUFA, that somewhat improves fasting glucose levels (shown), and directionally improves fasting insulin: Image
Dietary composition may not be useful for improving the rate of weight loss ceteris paribus, but it can definitely make it easier given what else it changes.

Those non-metabolism details may be why so many people find low-carb diets so easy!

Read 4 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!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(