Skip to content
Prev 19581 / 20628 Next

Using Robust Standard Errors lme4

Hi,

I have tried to do that with lme4 and the function tab_model() of pacakge sjPlot. That works for other types of models, but it did not worked for lme4. Try WeMix package, which mimic Stata?s mixed function.

However, I have substantive doubts about this. Do  robust standard errors make sense in multilevel models?. For instance, using clustered standard errors is a way of introducing level-two correlations in OLS, but in hierarchical models we are already modelling that correlation. Additionally, there are more general doubts about that. See:
https://www.nber.org/papers/w24003

I would like to listen more thoughts about robust and clustered standard errors in multilevel models, both from the philosophical point of view and the practical one.

Best,

Fernando Bruna
Department of Economics
Universidade da Corunha, Spain.
Message-ID: <PAXPR02MB7798D370AD8B329E178F9FBA92919@PAXPR02MB7798.eurprd02.prod.outlook.com>
In-Reply-To: <MWHPR0101MB31986D54DDD4BACBA61E4EBDA3919@MWHPR0101MB3198.prod.exchangelabs.com>