The current implementation of the regression simulators generates the covariates internally as normals or uniforms. This makes it impossible to simulate a regression with dummy variables.
Solution: add the input argument X, which is the whole matrix of covariates (which can then be simulated in whatever way one likes) with a default as NA. If X == NA, the code then generates the X internally as done currently.
The current implementation of the regression simulators generates the covariates internally as normals or uniforms. This makes it impossible to simulate a regression with dummy variables.
Solution: add the input argument X, which is the whole matrix of covariates (which can then be simulated in whatever way one likes) with a default as NA. If X == NA, the code then generates the X internally as done currently.