Re: st: Question concerning Bootstrapping in Stata

[Stas Kolenikov <skolenik@gmail.com>]

This means that what Stata applied bootstrap to produced an error --
most often, a convergence failure or perfect prediction (in binary
dependent variable models). To see what's going on, specify -noisily-
option of the -bootstrap- command. If you have an insanely heavy model
with 150 variables, then you may simply have  run out of degrees of
freedom for these bootstrap samples.

On Mon, Jul 23, 2012 at 11:11 AM, Anna-Leigh Stone
<alstone@crimson.ua.edu> wrote:
> Hello,
>
> I am running a fractional probit model in STATA 12 using the glm
> command. I have a panel data set of almost 200 companies over 50
> years. I am then bootstrapping 500 times to get the average partial
> effects. I am bootstrapping using the -bootstrap- command and
> following options: reps(500) seed(123) cluster(gvkey)
> idcluster(newid). Where gvkey is my company code. When STATA is
> bootstrapping, sometimes I have red x's appear. Then at the end of the
> bootstrapping I get a message like the following:
>
> Note: one or more parameters could not be estimated in 81 bootstrap replicates;
>       standard-error estimates include only complete replications.
>
> I was wondering what this means and if it invalidates my results?
>
> Thank you,
> Anna-Leigh
>
> --
> Anna-Leigh Stone
> Finance PHD Candidate 2014
> MA/EC 2011
> MSF 2010
> alstone@crimson.ua.edu
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



-- 
---- Stas Kolenikov
-- http://stas.kolenikov.name
---- Senior Survey Statistician, Abt SRBI
-- Opinions stated in this email are mine only, and do not reflect the
position of my employer
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/