Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Zero Inflated Poisson


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Zero Inflated Poisson
Date   Fri, 14 Dec 2007 11:53:37 -0000

If they are, then Stata will tell you, indirectly. There are many
strategies
for variable choice, which range from putting everything in and then
cutting 
out variables that appear to be unimportant, tor building up from a very

simple model until extra complications do not appear worthwhile. You 
could try to automate the process using -stepwise-, but most people I
know would advise against that and suggest that you should bring your
scientific and practical judgment to bear. 

francesca.modena

I run a zero inflated poisson model. The sample is 409 observations,
with 87
nonzero observations and 322 zero observations. There are 42 independent
variables. Long and Freese (2001) report a general rule of at least 10
observations per parameter for models for categorical dependent
variables,
but they suggest that the zero-inflated poisson model seems to require
more
observations.

Moreover, if there is little variation in the dependent variable (as
could
be this case, in which nearly all of the outcomes are 0), a larger
sample is
required. 

Can you suggest me if 42 parameters are too many for my sample?

*
*   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/



© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index