treatment effect
The causal impact of an intervention (the treatment) on an outcome, the quantity causal analysis aims to estimate.
As defined in Susan Athey on Machine Learning, Big Data, and Causation
Where it comes up · 4
- Susan Athey on Machine Learning, Big Data, and Causation2 claims
- The Disruptive Potential of GenAI for Emerging Markets | Rem Koning | Leading with AI2 claims
- Jonathan Rodden on the Geography of Voting1 claim
- Jeremy Giffon - Special Situations in Private Markets - [Invest Like the Best, Replay]1 claim
Most important claims · 6 of 6Search all claims for “treatment effect” →
- For causal inference, you must distinguish the treatment variable (the minimum wage, the drug) from the control covariates and model them differently: apply machine learning to flexibly select controls, but give a causal interpretation only to the treatment variable, since data-mined controls are uninterpretable proxies that capture correlated information parsimoniously.Susan Athey on Machine Learning, Big Data, and Causation
- Elite organizations generate value either through selection effects, where prestige comes from picking people as with Harvard or modeling agencies, or through treatment effects, where the program changes the person, as with special forces training; the best organizations do both, and the resulting polish is visible in how people write and think.Jeremy Giffon - Special Situations in Private Markets - [Invest Like the Best, Replay]
- Whether cities make people Left-leaning (a treatment effect, e.g., urban residents valuing public transit or wanting drug problems addressed) or whether already-Left-leaning people sort into cities (a selection effect) is the central analytical question, and disentangling it would ideally require randomly assigning people to urban, suburban, and rural environments.Jonathan Rodden on the Geography of Voting
- The starting point for causal inference is to treat the causal variable of interest (e.g., minimum wage, the drug) differently from all the control covariates: use machine learning to data-mine the covariates but model the treatment variable carefully and separately, rather than putting everything into one regression with the same statistical treatment.Susan Athey on Machine Learning, Big Data, and Causation
- In a randomized controlled trial of 640 Kenyan entrepreneurs (311 control, 306 treatment), providing access to a GPT-4-powered WhatsApp business advisor had no statistically significant effect on revenue and profit in the full sample (approximately 5% increase but within error bars), contrary to expectations.The Disruptive Potential of GenAI for Emerging Markets | Rem Koning | Leading with AI
- The speaker expects that with continued refinement of the prompt and more data over another year of work, he can improve the negative effects observed for low-performing entrepreneurs and 'get that up' (improve their treatment effect from negative toward positive).The Disruptive Potential of GenAI for Emerging Markets | Rem Koning | Leading with AI