Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: re: wald test across iv eqns

From   Kit Baum <>
Subject   st: re: wald test across iv eqns
Date   Fri, 22 Apr 2005 11:32:46 -0400

Johannes (he with the sticking caps-lock key) said

I have trouble achieving the following thing in STATA:
I have 4 linear regressions of the same type, and they differ in one
variable. Now I would like to test if the coefficients of this one variable
in the different estimations differ significantly (F-, or Wald-Test).
I figured out how to do it in OLS:
I typed "estimates store est_#" after every equation, afterwards I estimated
"suest est_*" (seemingly unrelated estimations), and afterwards I could test
"test var1=var2=var3=var4"
But now I would like to do the same thing after estimating IV (ivreg2). it
shouldn't be so hard but I cannot figure out how to store the estimation
results properly so that STATA can compare the coefficients over the 4

There is a bug in suest. Official Stata's ivreg (and our ivreg2) store a score variable, on request, as e(pscorevar). suest will not look there for it, demanding that it be named in e(scorevar). So the trick is...

ivreg2 ..., pscore(one)
est store one
est change one, scorevars(one)
ivreg2 ..., pscore(two)
est store two
est change two, scorevars(two)
suest one two

Kit Baum, Boston College Economics

* For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index