I'm trying to apply the Oaxaca-Blinder decomposition to the difference of two probabilities: Prob(y=1|A=1) for group A and Prob(y=1|B=1) for group B with y=(0,1) and Prob(y=1)=1/(1+e^-XBeta). My aim is separating the share of the probabilities difference due to cross-group differences in the explanatory variables from the part that remains unexplained by these differences (due to discrimination). I'd like to know if there is a package in R that perfoms that operation.
Thanks in advance for your help.
Details on the R package you want can be found at the link given below. Best wishes.
https://search.yahoo.com/search;_ylc=X3oDMTFiN25laTRvBF9TAzIwMjM1MzgwNzUEaXRjAzEEc2VjA3NyY2hfcWEEc2xrA3NyY2h3ZWI-?p=oaxaca+R+package&fr=yfp-t-563&fp=1&toggle=1&cop=mss&ei=UTF-8
Thank you for your reference paper and link. I see Fairlie developed different programs to achieve the decomposition of the differences with logit models (http://people.ucsc.edu/~rfairlie/decomposition/), but none for R.
Thank you for your help.
Check page 6 from the documentation for the oaxaca package in R here: https://cran.r-project.org/web/packages/oaxaca/vignettes/oaxaca.pdf
They say that you can choose a different regression function, although the decomposition will be based on the linear systematic component of the regression method:
"Last but not least, users can use a different regression function (argument reg.fun) to estimate the regression coefficients used in the decompositions. Note that, if a non-linear function such as glm() is chosen, the decomposition will be based on the linear systematic component – usually associated with the estimation of the corresponding latent variable – of the regression method."
There is an R package named 'GeneralOaxaca' that can perform Blinder-Oaxaca Decomposition for Generalized Linear Models: https://cran.r-project.org/web/packages/GeneralOaxaca/GeneralOaxaca.pdf.
I am surprised nobody mentioned it before, the package is around since 2015.
- Badges
- Science topic
- Similar topics
- Psychology