README.md

In telliott27-research/KOBdecomp: KOB Decomposition

KOBdecomp

Package for running customized KOB decompositions.

This package was developed for the academic article Softer Policing or the Institutionalization of Protest? Decomposing Changes in Observed Protest Policing Over Time by Thomas Elliott, Jennifer Earl, Thomas V. Maher, and Heidi Reynolds-Stenson. Please refer to that paper, specifically the methodological appendix, for more details on how the model works.

The package provides a method for applying the modified Kitagawa-Oaxaca-Blinder (KOB) decomposition used in the article to binary and multinomial outcomes with continuous grouping variables (e.g. time).

Installing the package

The easiest way to install the package is to use the install_github function from the remotes package:

remotes::install_github("telliott27-research/KOBdecomp")

For binary outcome models, the primary function for analysis is gen_decomposed_results_probit.

For multinomial outcome models, the primary function for analysis is gen_decomposed_results_multinom.

telliott27-research/KOBdecomp documentation built on March 19, 2022, 7:27 a.m.