1. I am asked to test heteroskedasticity on the Tobit model by using Stata. I know how to construct this test but I dont know how to generate a second order generalized residuals.

2. When I estimate an equation using Wtihin-Group estimator, why is one of the variables redundant?

3. What's the meaning by " the shape of the estimated relationship between dependent variable and explanatory variable"

I have just started to learn econometrics, so probably those question are too esay for you. I really appretiate that you can give me the answers or hints. Thank you.