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
Josh Merfeld Profile picture
Josh Merfeld
@Josh_Merfeld
Development, labor, and nerdy measurement stuff | Associate Professor, @KDI_School | @iza_bonn | @jpal | insulin addict | dad of an awesome girl
Mar 12 • 21 tweets • 4 min read
New report just dropped. Time for a thread.

You can find the replication report and the author's response in the linked thread (in the second tweet). Let me start with a small point: I hate the title of the replication. It refers to a comment in a do-file but, to be perfectly honest, it's entirely possible that it was just poor word choice from an RA whose first language is not English. The data is much more damning.
Feb 24 • 24 tweets • 4 min read
Reported treatment effects are GIGANTIC. 2.5 sd in India! Never seen anything that large before. First damning thing: "Participants were drawn from two earlier studies and the treatment assignments... were simply reused." (p. 2)

Note the title of the paper though. Best-case scenario: they are layering another intervention but saying it's new.
Mar 28, 2024 • 6 tweets • 2 min read
Just to add on: It’s quite common in econ to just throw FE into a regression. Consider the two most common FE we use: unit and time.

For unit FE, you need the treatment variable to vary within the unit. If treatment is fixed for all units? This shouldn’t work. For time FE, you need changes in treatment within the time dimension. In other words, if all units are treated at the exact same time, the treatment dummy will be perfectly collinear with the time FE.
Aug 22, 2023 • 13 tweets • 3 min read
Take a look at these maps (from Liu et al, 2021). There’s something we don’t often expect in the US given how free people can move across states: very sharp discontinuities at the border.

So what’s going on? Image These estimates come from a statistical method called small area estimation (SAE). Suppose you have a survey in the US. Using only the survey, you may only be able to say something about states, given sample sizes.

But what if you want to say something about counties? SAE!