Dear Statalist,
I'm using metobit model to deal with the multilevel data structure and the censored dependent variable issues, and my data is panel.
I'm aware that it is important to test for assumptions of metobit model. But I don't know how to do it, especially for heteroskedasticity.
I used the following code to test for normality of the residuals,
reg GeoDispersion X1 X2 X3
predict r, resid
kdensity r, normal
pnorm r
qnorm r
swilk r
This is the plot graphs I got,
It seems not strictly normal but more or less normal, and the sample size is 600 firms and 3200 observations. Does it violate the normality assumption?
Thank you very much for suggestions!
I'm using metobit model to deal with the multilevel data structure and the censored dependent variable issues, and my data is panel.
I'm aware that it is important to test for assumptions of metobit model. But I don't know how to do it, especially for heteroskedasticity.
I used the following code to test for normality of the residuals,
reg GeoDispersion X1 X2 X3
predict r, resid
kdensity r, normal
pnorm r
qnorm r
swilk r
This is the plot graphs I got,
It seems not strictly normal but more or less normal, and the sample size is 600 firms and 3200 observations. Does it violate the normality assumption?
Thank you very much for suggestions!
