This is so nice! Thanks for sharing :)

I think this question was probably the starting point of Dube, Lester, and Reich's (2010) “local estimators comparing all contiguous counties” approach...! At the very least, spatial heterogeneity would be less concerning than simply including state FE.

We use transfer learning, a machine learning approach that addresses the deviation-variance tradeoff.

It combines large source data with small target data to create a dementia estimation algorithm.

🔎 Replication R code 👉 github.com/TL-dementia/...

Funding: ERC #Episky #Dementia

Fantastic news before Christmas! :) Congratulations 👏👏👏

Here is our 9-minute YouTube video explaining the findings!
www.youtube.com/watch?v=s4bT...

@giorgiamenta.bsky.social has a fantastic paper on "Maternal genetic risk for depression and child human capital" using the Avon Longitudinal Study of Parents and Children (ALSPAC) survey. www.sciencedirect.com/science/arti...

Thank you so much. It means a lot. I just sent the WP, hope you enjoy reading them!

This is incredibly helpful, thank you for the transparency and the courage to share this!