I am running a general linear model in gamlj. The model runs without errors, but has a number of assumptions violations. When I address the assumption violations by using robust standard errors I get this error message:
1-Non-standard coefficient covariance matrix may not be used for model with aliased coefficients.
When I bootstrap the CI I get this error:
2-Replacement has x rows data has y.
The surprising thing is that without bootstrapping or using robust SE there are no error messages. Any idea why or how to correct this?
Bootstrapping and robust SE cause errors
- mcfanda@gmail.com
- Posts: 501
- Joined: Thu Mar 23, 2017 9:24 pm
Re: Bootstrapping and robust SE cause errors
Make sure the model is correct, check that all effects are estimable and all coefficients are estimated. If the model cannot be estimated, robust standard errors cannot be computed