Kit and Austin are correct (except that I am not clear why,
apart from speed, Austin is recommending -gr7-).
However, there is also a package -modeldiag-, which is
less restrictive than official Stata's offerings under
this heading, and includes other plots that are less
trivial to do by yourself. I've not tested that any
work after -ivreg- or -ivreg2-, but I am hopeful.
. search modeldiag
will indicate the Stata Journal archive source.
An update will appear via Stata Journal 8(1) 2008.
Nick
[email protected]
Kit Baum
========
-rvfplot- is nothing more than
ivreg2...
predict resid,resid
predict xb, xb
scatter resid xb
whereas -rvpplot- is just
scatter resid xvar (where xvar is one of the regressors)
so you can readily generate either type of plot after -any- estimation
command from which residuals and predicted values are available.
Austin Nichols
==============
Marck Bulter <[email protected]>:
One can always construct such plots using -predict- and -graph-
use http://fmwww.bc.edu/ec-p/data/wooldridge/card, clear loc x "exper*
smsa* south mar black reg662-reg669"
ivreg2 lw `x' (educ=nearc2 nearc4), first qui reg educ `x' nearc2 nearc4
predict xhat qui reg lw xhat `x'
predict ehat, resid
predict yhat
gr7 ehat xhat
gr7 ehat yhat
See also sec 4.7 of http://pped.org/stata/ciwod.pdf
Marck Bulter
============
Just wondering, is there a post estimation command like rvpplot or
rvfplot available after ivreg(2). I have searched through the list,
but could not find something useful.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
