I'd like to use the "mrobust" command by Young & Holsteen (2017) for estimating model robustness. I investigate the effects of duration in preschool on school success. My logistic regression contains a squared term, the independent variable, and looks as follows:
Code:
logit success c.preschool##c.preschool c.loginc c.logwealth i.educ i.mig i.aspiration, vce(cluster region)
Code:
gen preschool_squared = preschool*preschool
Code:
mrobust logit success (preschool preschool_squared) c.loginc c.logwealth i.educ i.mig i.aspiration, vce(cluster region)
Does anyone here have experience with the mrobust command and the possibilities for solving the polynomial problem? Or knows a similiar way to check on model robustness with another command than mrobust but similiar content?
Thank you very much,
Susanne
[References: Young, C. & K. Holsteen, 2017: Model Uncertainty and Robustness. Sociological Methods & Research 46: 3–40.]
0 Response to Using mrobust with polynomial independent variable
Post a Comment