Dear all,

I learned from Statalist that we could use the following formula to calculate icc after meologit regression (by Stata 15.1):

ICC= var_u/(var_u + (pi)^2/3)

But my question is what if the second level control has more than one variable? How should this formula be?


My code is :meologit urbani A1b A0 marriage i.educat4 occu_prof hhwealth hukou_cat32 hukou_cat33 hukou_out_town_impute1 [pw=weight_temp_use] || gbcode: sample_urbaning pilot1 , nolog

As can be seen that are three second-level variable control: country id (gbcide); urban sample(sample_urbaning); and policy pilot(pilot1)
and the variance of second-level data are shown in the table:


Array

Array


Then how could I calculate icc?

Thank you for your help!