Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: why does stata creates missing values when running zinb model


From   Ekaterina Hertog <ekaterina.hertog@sociology.ox.ac.uk>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   Re: st: why does stata creates missing values when running zinb model
Date   Wed, 19 Oct 2011 11:48:05 +0100

Dear Maarten,
You are right, I have a big dataset of marriage customers which includes both men and women of ages up to 70. Thank you very much. I experimented a bit and it seems Stata is just telling me how many observations it is not using in this particular model.
Ekaterina

On 19/10/2011 10:35, Maarten Buis wrote:
Given that you used an if condition I assume that your dataset
actually contains more observations than just women who are less than
40. If you type after your -zinb- model -count if !e(sample)-, do you
get 31478? If that is the case than there is probably no problem. (In
that case someone at StataCorp forgot a -quietly-.)

-- Maarten

On Wed, Oct 19, 2011 at 10:44 AM, Ekaterina Hertog
<ekaterina.hertog@sociology.ox.ac.uk>  wrote:
Dear All,

I am trying to run a zero inflated negative binominal model in Stata 11. My
dataset contains 13440 women below 39 who are marriage agency customers. The
customers of the agency HAVE to provide information on all the major
demographic variables so there are no missing values on key variables (age,
height, income, education etc). My dependent variable is nbidder which
stands for the number of offers a given customer received over the past
year. About 25% of women did not receive any offers (nbidder=0)

The command I typed is:

zinb nbidder prsexrat20441539 ciunmsexrat2039 guntotot age agesq bmi divwid
havechildren lninc university smoker if age<=39&  sex==2, nolog inf(age
agesq bmi havechildren lninc university) exposure(duration)

prsexrat20441539 ciunmsexrat2039 are measures of local sex ratio for each
customer and guntotot measure how rural the area is.

Once I run the command I get a message (31478 missing values generated). I
will be very grateful for any advice of why this may be happening,
Ekaterina

On 19/10/2011 00:28, Nick Cox wrote:
You fail to tell us anything whatsoever about your data (except that
the number of observations presumably is at least 3000) and you do not
tell us even the exact command you typed. This is quite contrary to
all the advice given to you in the FAQ and in particular repeated on
the list within the last 24 hours.

-set trace on- will show you where what you think is a problem occurs.
If the resulting log is long you may need to edit it for Statalist.
Otherwise why not liaise with Stata technical support directly? As
above, they will want to know about your dataset.

Nick

On Tue, Oct 18, 2011 at 11:20 PM, Ekaterina Hertog
<ekaterina.hertog@sociology.ox.ac.uk>    wrote:
Dear All,
I am trying to run a zero inflated negative binominal zinb model in Stata
11
and the first output Stata gives me is: 3000 missing values generated. I
am
not sure why this is happening and it sounds wrong, but I cannot find out
what exactly this means and hence how to solve the underlying problem.
Any
advice will be greatly appreciated,
*
*   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/
*
*   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/




*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index