Hi,

I'm assessing the disability wage gap at two points in time using the Oaxaca-Blinder decomposition.

Regarding robustness checks, I was planning to assess the exclusion of one or two variables, as well as running preliminary OLS regressions and assessing the linearity of these via Breusch-Pagan in order to indirectly assess the imposition of linearity by the O-B decomposition.

I was wondering whether there were more optimal robustness checks and/or more direct checks for the O-B decomposition, because reading previous literature and online resources does not seem to answer this?

Any advice would be greatly appreciated, thank you.