Software
Inference for an algorithmic fairness–accuracy frontier
TestAlgFair is the replication package for “Inference for an Algorithmic Fairness–Accuracy Frontier,” with Yiqi Liu. It provides the data and R code used to reproduce the Monte Carlo simulations and empirical application, together with functions implementing the paper’s statistical procedures.
MATLAB: confidence intervals for projections of partially identified parameters
The ZIP file contains MATLAB programs that replicate the simulation results for the Monte Carlo exercises in Kaido, Molinari, and Stoye (2017). The code is “portable”: modifications for other applications, such as different studentized sample moment functions or estimators of population gradients, are minimal and documented in the user’s manual.
If you use the software, please cite Kaido, Molinari, and Stoye (2017) as the source of the theoretical results and code.
Fortran and MATLAB: sharp identification regions
Download the Fortran and MATLAB programs
The ZIP file contains programs that replicate the simulation results in Beresteanu, Molchanov, and Molinari (2011). If you use the software, please cite Beresteanu, Molchanov, and Molinari (Econometrica, 2011) as the source of the theoretical results and the online supplement as the source of the code.
Stata: best linear prediction with interval outcome data
The ZIP file contains .ado and .sthlp files that replicate results in Beresteanu and Molinari (2008) and implement their theoretical results. The code supports best linear prediction with any number of perfectly observed covariates and returns estimates and confidence sets for the sharp identification region.
If you use the software, please cite Arie Beresteanu, Francesca Molinari, and Darcy Steeg Morris (2010), “Asymptotics for Partially Identified Models in Stata,” and Beresteanu and Molinari (Econometrica, 2008).
The programs are distributed in the hope that they will be useful, but without any warranty, including implied warranties of merchantability or fitness for a particular purpose.
This research was supported in part by the National Science Foundation through grants from 2006–2014 and 2018–present. Any opinions, findings, conclusions, or recommendations expressed in these materials are those of the authors and do not necessarily reflect the views of the National Science Foundation.