01-18-2012 05:42 PM
In my theises I run a fixed effekts model with Proc GLM on trade data spanning over 30 years (variable=x6) and 150 countries.
My GLM program is the following:
proc glm data= data_a;
m1: model y=x1 x2 x3 x4 x5 x6 /solution;
My question is how I obtain and use robust standard errors adjusted at 'country' (eg variavle x7) level, if this variable is not included in the model?
All help is highly appreciated!!