Nuvan Rathnayaka
banner
nuvanrath.bsky.social
Nuvan Rathnayaka
@nuvanrath.bsky.social
Brighton and Hove Albion fanatic. Statistician. Dog lover.
Reposted by Nuvan Rathnayaka
I was NOT expecting to win this award and I’m so so so honored. Grateful to the New Yorker, the Beard Foundation, and of course most of all Padma.

I wasn’t able to be at the awards ceremony in person — here’s my acceptance speech.
June 15, 2025 at 2:45 AM
2) What are the workflow tradeoffs? With IPW, one can revise the treatment model until adequate balance is achieved without biasing the treatment effect estimate or needing to account for the multiple looks in the variance estimation. Would the same hold true for g-computation?
February 2, 2025 at 9:19 PM
Really great write up! Clear and concise explanation with a tidy example in code. Two things I'm wondering about:

1) How does the computational complexity compare with that for IPW methods?
February 2, 2025 at 9:16 PM
Reposted by Nuvan Rathnayaka
a relevant starter pack:

go.bsky.app/2Fq4P6e
January 15, 2025 at 7:15 PM