I am currently using rxLogit models in MRS as an alternative to standard GLM models in MRO (~300,000 rows, but 3 factors with 200, 400, and 5000 levels). This cuts my computing time from 26 to 7 hours on a 2x6 core Xeon with 128 GB RAM.
After the estimation I need/want to calculated clustered robust standard errors. In standard R I can do this in multiple ways, e.g., via cluster.vcoc() from the multiwayvcoc package.
I tried converting the rxLogit object with as.glm() to a glm object and perform my analyses subsequently, but then I am getting the following error:
Error in qr.lm(object) : lm object does not have a proper 'qr' component.
Rank zero or should not have used lm(.., qr=FALSE).
I also get this error when calling, e.g., summary(as.glm(rxLogitObject)).
Has anyone encountered a similar/the same issue and can offer a solution?
Thanks a lot,
Niklas
