The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.

NonlinearDiD 0.2.0

• Added support for repeated cross-section staggered DiD designs.
• Added data_type = "panel" and data_type = "repeated_cross_section" options.
• Made idname optional for repeated cross-section designs.
• Added support for sampling weights through weightsname.
• Added support for clustered inference through cluster_var.
• Added examples for binary repeated cross-section outcomes with staggered treatment timing.
• Preserved all panel-data functionality from version 0.1.0.

Implementation notes

• Repeated cross-section ATT(g,t) uses the Wooldridge (2023) pooled QMLE with a treatment-by-period interaction. The doubly-robust variant augments this with inverse probability weighting on the estimated propensity score.
• Sampling weights (when supplied via weightsname) are used throughout: the outcome regression, the propensity score model, and the pooled QMLE. They are multiplied with the IPW factor in the doubly-robust path.
• Analytical SEs for the RCS path use sandwich::vcovCL when cluster_var is supplied, sandwich::vcovHC (HC1) otherwise. Panel SEs continue to use the influence-function approach from v0.1.0; set boot = TRUE for fully clustered panel inference.
• The bootstrap automatically resamples whole clusters when cluster_var is provided, units when data_type = "panel", or individual rows when data_type = "repeated_cross_section" without clustering.
• The compiled C++ helpers from v0.1.0 have been replaced with equivalent pure-R implementations, eliminating the Rcpp dependency. The package no longer requires compilation.

Bug fixes

• None.

NonlinearDiD 0.1.0

• Initial CRAN release.
• Panel-data ATT(g,t) estimation under logit, probit, Poisson, negative binomial, and linear outcome models.
• Doubly-robust estimator combining outcome regression and propensity score weighting.
• nonlinear_aggte(): event-study, group, calendar, and simple aggregations.
• nonlinear_pretest(): joint chi-squared and individual pre-trend tests.
• nonlinear_bounds(): Manski and parallel-trends-constrained bounds.
• binary_did_logit(), binary_did_probit(), binary_did_dr(): 2x2 binary DiD estimators.
• count_did_poisson(): Poisson QMLE DiD.
• odds_ratio_did(): scale-free odds-ratio DiD.
• sim_binary_panel(), sim_count_panel(): simulation utilities.

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.