mean regression vs. regression toward the mean

[Vatana Chea]

Dear All,

I just have a few dumb questions. Hopefully, someone here will help me understand econometrics better.

1. Mean regression and regression toward the mean, are they different or just one thing that is called by two different names? If they are different, what is Stata command for both of them?

2. The variance of the predicted value deriving from OLS is artificially reduced because it does not take into account the unobserved variables. Is there any way to fix the variance? I heard that we can add random error component to the predicted value to help fix it, but I do not think I understand how they do it clearly enough. Can somebody suggest something I can read? How about Stata command for that?

Thank you very much for your help in advance Statalist.

[Carlo Lazzaro]

Vatana:
1) Take a look at https://en.wikipedia.org/wiki/Regres...oward_the_mean and at -regress- in Stata.
2) Never heard about that point before. That said, the best approach to get informative predicted values, is to have a well specified regression model (see -estat ovtest- and -estat hettest-).