What's the Harm? Sharp Bounds on the Fraction Negatively Affected by Treatment

Authors: Nathan Kallus

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate our method in simulation studies and in a case study of career counseling for the unemployed.
Researcher Affiliation Collaboration Nathan Kallus Netflix and Cornell University kallus@cornell.edu
Pseudocode Yes Algorithm 1 Point estimate and CIs for AHEρ g0,...,gm
Open Source Code Yes Replication code is available at https://github.com/CausalML/BoundsOnFractionNegativelyAffected.
Open Datasets Yes Using data from [6] (BSD license), we compare three different assistance programs offered to French unemployed individuals: the standard benefits (sta), access to public-run counseling (pub), and to private-run counseling (pri). ... Luc Behaghel, Bruno Crépon, and Marc Gurgand. Private and public provision of counseling to job seekers: Evidence from a large controlled experiment. American economic journal: applied economics, 6(4):142 74, 2014.
Dataset Splits No The paper mentions using cross-fitting (Alg. 1, step 1) which implies data splitting, but does not specify the explicit percentages or sample counts for training, validation, or test splits. It does not provide details like '80/10/10 split'.
Hardware Specification Yes Experiments were run on AWS c5.24xlarge.
Software Dependencies No The paper mentions using 'R package grf' but does not provide a specific version number for this or any other software dependency.
Experiment Setup Yes We apply Alg. 1 to this data with K = 5, estimating e, µ using random regression forests and τ , τ+ using causal forests (all using R package grf with default parameters). ... We fit µ using linear regression and τ , τ+ using doubly-robust pseudo-outcome linear regression.