Lightnews — Scholar-powered news

Pietro Biroli

@pietrobiroli.bsky.social

3.2K followers 2.6K following 500 posts

Applied econ @unibo interested in genes, health, human capital. Dadx2. Lactose and fascist intolerant. he/him 📉📈 https://sites.google.com/site/pietrobiroli/

sites.google.com

Posts Media Videos Starter Packs

Pinned

Pietro Biroli @pietrobiroli.bsky.social · May 20

🧵New paper out:

“Beyond Time: Unveiling the Invisible Burden of Mental Load”

with Barigozzi, Monfardini, Montinari, @elenapisanelli.bsky.social @svevav.bsky.social

🔗 arxiv.org/abs/2505.11426

/1

Beyond Time: Unveiling the Invisible Burden of Mental Load

This paper introduces a novel, scalable methodology to measure individual perceptions of gaps in mental load -- the cognitive and emotional burden associated with organizing household and childcare ta...

arxiv.org

2 13 32

Pietro Biroli @pietrobiroli.bsky.social · 3d

Look I don't think we're going to agree. There are many reasons why in practice, for certain settings, RCT can be worse than other methods.

But I'm any case, lots of messy reality and not much idealization in the following:

gen treat = 1 if runiform()>0.5
replace treat = 0 if treat==.

Pietro Biroli @pietrobiroli.bsky.social · 3d

I believe there's a difference between creating your own random variable which you know to be orthogonal _by construction_ to everything else, and looking at a cutoff and arguing that things on average are similar on one side and the other

1 2

Reposted by Pietro Biroli

D. Earl Stephens @dearlstephens.bsky.social · 4d

440 3.2K 11K

Reposted by Pietro Biroli

Ruth Ann Crystal @drruth.bsky.social · 4d

The HPV vaccine is doing what had been promised.
17 years after it became available, HPV infections decreased significantly in vaccinated people and unvaccinated people because of herd immunity. buff.ly/WvjQ1BS

h/t @boghuma.bsky.social

#medsky #pedsky 🛟🧪

Light blue background. Young woman with dark hair makes a sign of strength with her arm and she has a vaccine bandaid on her upper arm. It says "The HPV Vaccine Works!"

7 140 470

Pietro Biroli @pietrobiroli.bsky.social · 4d

I mean, your actual RCT is *not* that perfect trial. For sure. But mostly in (semi) testable ways: non compliance, attrition, lack of adherence to the treatment content, etc.

Pietro Biroli @pietrobiroli.bsky.social · 4d

Okay but did they have IRB approval to conduct experiments in plato's cave?

I mean were they trespassing??

(I'm joking mostly because I agree with you)

1 1

Pietro Biroli @pietrobiroli.bsky.social · 4d

Is ciccio _||_ age an assumption? Or a property of that random variable?

Cov(ciccio,age) will never be exactly zero. But that doesn't tell me that ciccio is endogenous

Pietro Biroli @pietrobiroli.bsky.social · 4d

Adherence is never perfect, so you and never have the ideal experiment, but that's a different discussion.

Let me rephrase. Let's say that I open my favorite data set, which contains a variable called age, and I generate a random uniform variable which I call ciccio.

/1

Pietro Biroli @pietrobiroli.bsky.social · 4d

Ok but now you're changing the goalpost.

What I said I didn't agree with in your previous post was "You (a) have to assume that the experimental protocol was followed perfectly"

That's different from Y1 and Y0, which are unobservable.

Adherence to experimental protocol is observable.

Pietro Biroli @pietrobiroli.bsky.social · 4d

Very true.

But the messiness of RCTs or other data analysis are very empirical and applied in nature. No need to invoke differential hypothesis of counterfactuals for those.

One argument could be that rcts are "better" in theory but "worse" in practice, but I thought your claim was different

Pietro Biroli @pietrobiroli.bsky.social · 4d

Sorry I thought you were arguing that RCT were worse because they had to assume SUTVA

Certainly true that you need assumptions to interpret any estimate, RCT based or otherwise

1 3

Pietro Biroli @pietrobiroli.bsky.social · 4d

Regarding (a), more than an assumption it's something you need to check ex post. Like the relevance of IV: you calculate the f statistic in the first stage.

In an RCT you know if you randomized or not. You can check whether there's differential attrition.

Pietro Biroli @pietrobiroli.bsky.social · 4d

I agree that RCT are not the gold standard, but I don't agree with a and b

isn't (b) a problem common to every method? Actually to every statistic? That's always the statistic for that sample. If you want to draw more general conclusions, you always have to make assumptions about extern validity

1 3

Pietro Biroli @pietrobiroli.bsky.social · 4d

But virtually every standard causal method assumes SUTVA.

Whenever we write down Y1i Y0i we are assuming SUTVA

2 3

Pietro Biroli @pietrobiroli.bsky.social · 4d

I assume that epsilon and X are always orthogonal.

Not always and everywhere, of course, I'm not stupid.

Just in all the regressions that I decide to run. Totally plausible assumption.

Chelsea Parlett @chelseaparlett.bsky.social · 5d

It’s not the method that makes you causal it’s the assumptions

Reposted by Pietro Biroli

Eladio Bobadilla @eladiobobadilla.com · 5d

22 450 1.8K

Reposted by Pietro Biroli

Philip Bump @pbump.com · 5d

smart people: this is a violation of the constitution
me: the antecedent of “you” is the ice officers and cops

David S. Bernstein @dbernstein.bsky.social · 5d

Oh he can just unilaterally pass unconstitutional laws now huh ok

84 250 1.8K

Reposted by Pietro Biroli

Realities 🌐 @insuret.bsky.social · 4d

3 21 100

Reposted by Pietro Biroli

Chelsea Parlett @chelseaparlett.bsky.social · 5d

It’s not the method that makes you causal it’s the assumptions

5 18 87

Reposted by Pietro Biroli

Julia M. Rohrer @dingdingpeng.the100.ci · 5d

Today is a public holiday in Germany to celebrate reunification. Seems like the perfect day to once again recommend this amazing book on the rather incredible chain of events leading up to the opening of the Berlin wall!

Julia M. Rohrer @dingdingpeng.the100.ci · Aug 14

Just finished "The Collapse" about the precise events that led up to the opening of the Berlin Wall -- struggle within the Soviet Bloc and demonstrations in Saxony (most prominently Leipzig), bureaucratic miscommunication, and the chaotic and improbable events in the night itself.>

Cover of "The Collapse: The Accidental Opening of the Berlin Wall" by Mary Elise Sarotte. Cover shows a person peering through a crack in a wall covered in graffiti

2 7 50

Pietro Biroli @pietrobiroli.bsky.social · 7d

but they have outrageous claims. click click click

Reposted by Pietro Biroli

Hongyu Zhou @zhou-hy.bsky.social · 7d

🚨Thrilled to share our paper🚨
Does science focus on the diseases that hurt people the most?
When deadly outbreaks hit, how does research respond?

Our work with:
@prashantgarg.bsky.social
@trfetzer.com

shows how medical research worldwide responds to both endemic burdens and emergencies.
🧵👇

1 5 14

Pietro Biroli @pietrobiroli.bsky.social · 7d

I think that we need to value and consider different journals with different objectives but still a high bar in terms of scientific rigor

Pietro Biroli @pietrobiroli.bsky.social · 7d

I really like restud (and jeea) and they do have less bias in favor of American data or applications

But, correctly so, they're playing the same game as qje aer ecta etc: novelty, cool, path breaking research yada yada

That's not a healthy objective for ALL PhD students.

1 2

Pietro Biroli @pietrobiroli.bsky.social · 7d

Let's hope we go beyond counting the number of top5