Dear all,

I'm trying to compare the coefficients of an OLS regression with a multilevel regression. I tried to use the Hausman command but got an error saying that
Code:
 
 "model fitted on these data fails to meet the asymptotic assumptions of the Hausman test; see suest for a generalized test"
I'm trying to use the Suest command as a proper solution, but this command doesn't work after Mixed command. I understand that there's an option to use
Code:
vce(cluster clustervar)
in order to cluster 1-level models and with that use this command for mixed models.

However, in my case there's only one model with clustering (the other is just a normal linear regression), so if I try this syntax -

Code:
regress sqrt_Relative_time_days White Hispanic Other Age Sex i.Seriousness_of_crime_7 Death_Penalty i.year_of_conviction_groups

estimates store fixed


mixed sqrt_Relative_time_days White Hispanic Other Age Sex i.Seriousness_of_crime_7 Death_Penalty i.year_of_conviction_groups  || states_combined:

estimates store random

suest fixed random, vce(cluster clustervar)
the command create two identical clustered regression, where I want just the second one to be clustered (the model I call "random").

Is there a way to cluster just 1 model in the suest command?

Thanks!
Eran