Mincer zarnowitz regression

Demographer — Tue, 12 Sep 2017 12:43:03 GMT

Hello,

I would like to perform a Mincer zarnowitz regression, which is used to assess if there are systematic errors in predicted values (P) compared to observed values (O).

Basically, the equation is simple linear regression, such as

P=a+b(O)

However, to see if there are systematic errors, I need to test if b is statistically different from 1, while the p value of proc reg indicates if it is different from 0. Is it possible to change this? Otherwise, what would be the best way to test if my b are statistically different from 1?

Thank you.

Re: Mincer zarnowitz regression

Rick_SAS — Tue, 12 Sep 2017 13:49:50 GMT

Use the TEST statement:

test b=1;

topic Re: Mincer zarnowitz regression in Statistical Procedures

Mincer zarnowitz regression

Re: Mincer zarnowitz regression