Harald Seider
>
> when running Joe Harness's ivprob procedure I always get an
> error message
> (conformability error). I have no clue why this procedure
> won't work.
> I have about 250.000 observations which might be an issue.
> However, the
> error will also show up if we reduce the number of
> observations down to a
> few hundred.
> The next issue might be the fact that many of our variables
> are dummies
> (d*, transfer, diag* , dum*).
> Has anyone out there faced a similar problem before and
> know a solution for it?
>
> ivprob mort30 ,
> endog(sqrtvol)
> exog(procptca d0* d1* d2* d3* d5* transfer
> diagalkalos_x diagamisequ diagarterem
> diagasthma_x diagatheros
> diagbbblk_x diagcardiomeg_x diagchf
> diagchrglom diagchrlive_x
> diagchrpulh_x
> diagchrrena diagcnsdis diagcollvas
> diagcopd diagdbtcmp
> diagdbtuncm diagdeg2avb
> diagdement diagepilep diagfractur
> diaggihemor diaghismali_x
> diaghrsecma
> diaghtb diaghthrtf_x diaghyperli_x
> diagincorsy_x diaglrpmali_x
> diagmitvalv_x
> diagnutri_x diagostarth_x diagothcva2
> diagothvalv_x diagprcabg
> diagprevmi_x
> diagpsychos diagshock_x diagsknulcr
> diagsyncop_x diagthyroid
> diagunstangina
> dum*)
> iv (dum* pat3rad0to3 pat4rad3to7 pat7rad7to14
> pat4rad14to18
> pat4rad18to22 hosprad*) ;
>
Here is my wild guess (I don't profess to understand
the econometrics here):
Within -ivprob- one or more of your covariates
is being dropped because of collinearity when
it runs
probit `depvar' `fitted' `exog' if `touse'
When the matrix of means is being decorated
with column names, there are fewer columns
than names, and Stata squawks.
There is a possibility that there is
something for Joe to fix using -_rmcoll-.
Nick
[email protected]
*
* 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/
