Publication Code for Balkus et al. (2026)

Code to reproduce the results reported in the JRSS-B paper "The causal effects of modified treatment policies under network interference" by Salvador Balkus, Scott Delaney, and Nima Hejazi.

• data-cleaning contains R scripts used to harmonize California ZEV data at the ZCTA-level (treatment), gridded NO2 pollution forecasts (outcome), socioeconomic Census data and EPA Smart Mapping land use data (confounders), as well as the interference network from Census LODES data.
• analysis contains Julia and R scripts used to perform the data analysis using the ModifiedTreatment.jl and Condensity.jl packages.
• simulations contains Julia code to generate synthetic data using CausalTables.jl and evaluate operating characteristics of the model using ModifiedTreatment.jl and Condensity.jl to fit the models.
• data contains the NO2/ZEV data example, as well as simulation results. See also the Harvard Dataverse entry for the data example.
• results stores the final manuscript images and .csv files used to create plots for the data analysis.

Note that analysis and simulations depend on Julia packages still under development. If any issues or concerns arise, please file an issue describing the problem at hand.

If you wish to cite this work, you may use the following Bibtex:

@article{Balkus2026,
  title = {The causal effects of modified treatment policies under network interference},
  ISSN = {1467-9868},
  url = {http://dx.doi.org/10.1093/jrsssb/qkag052},
  DOI = {10.1093/jrsssb/qkag052},
  journal = {Journal of the Royal Statistical Society Series B: Statistical Methodology},
  publisher = {Oxford University Press (OUP)},
  author = {Balkus,  Salvador V and Delaney,  Scott W and Hejazi,  Nima S},
  year = {2026},
  month = mar 
}