Stanford releases TROP, a triply robust panel estimator with official Python and Stata packages
asusarla · x · 2026-10-13
Susan Athey's Stanford team released official Python and Stata packages for the Triply Robust Panel (TROP) estimator. TROP relaxes the parallel trends assumption by combining unit weights, time weights, and a low-rank factor-model regression adjustment; its bias is the product of the three imbalances, so it is asymptotically unbiased if any one removes the underlying bias. It performed well in 20 of 21 simulations calibrated to widely used datasets like CPS and the Penn World Table.
Related event: Stanford releases triply robust panel estimator TROP for Python and Stata(2 posts)→
More from Research
- Noise: a speedrun for pretraining transformers without backprop — madhavsinghal_ · 2026-10-13
- Same model scores 24% vs 59% across harnesses, Chamath declares leaderboards worthless — eliano · 2026-10-13
- Harvard releases freeinference_agentic_trace, a million-scale agent trace dataset — harvardMadsys · 2026-10-13
- UCSD's Mental-Models framework gives agents explicit Theory-of-Mind for multi-agent systems — UCSanDiego · 2026-10-13
- Vals AI: Agent Teams Cost 1.8-5.1x More, Only One Score Gain Was Significant — rseroter · 2026-10-13
- DINOv3 vision transformers learn word meaning from images alone, study finds — udmrzn · 2026-10-13