Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: r(1400): initial values not feasible with gsem command


From   Tim Kautz <tdkautz@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: r(1400): initial values not feasible with gsem command
Date   Wed, 31 Jul 2013 09:42:19 -0500

Hi,

I am attempting to estimate a factor model using the gsem command.
However, I keep receiving the error "r(1400): initial values not
feasible."

I am using the gsem command rather than sem command, because I would
eventually like to incorporate a probit as part of the model. However,
when I use a simpler version of the model without the probit step, the
gsem command still gives me an error. For this simpler model, the sem
command works fine.

I have tried some of modifications suggested in the manual with little
success. I have tried using the solution to the sem version as  a
starting point for gsem and that does not work either (see below). I
have tried all possible intmethods, adjusting the number of intpoints,
and scaling my variables so they are mean 0 and standard deviation 1.
Interestingly when I reduce the sample to about 20,000 observations,
it works fine (the full data set has 70,000 observations).

sem  (cvar1 cvar2 cvar3 cvar4 cvar5 nvar1 nvar2 nvar3 nvar4 nvar5
<-FC) (nvar1 nvar2 nvar3 nvar4 nvar5 <-FN)
matrix b=e(b)
gsem (cvar1 cvar2 cvar3 cvar4 cvar5 nvar1 nvar2 nvar3 nvar4 nvar5<-FC)
(nvar1 nvar2 nvar3 nvar4 nvar5  <-FN) , from(b)

Any suggestions would be greatly appreciated.

Thank you,
Tim
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index