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]

Re: st: zero-inflated analyses: when do you decide that is zero-inflated?


From   David Hoaglin <[email protected]>
To   [email protected]
Subject   Re: st: zero-inflated analyses: when do you decide that is zero-inflated?
Date   Wed, 17 Jul 2013 06:49:52 -0400

Cris,

A GLM with a log link is what I had in mind.  I would explore first,
by using -regress- on the nonzero sizes, untransformed and in some
transformed scales, and making the customary plots of residuals.  The
log is often a good transformation to try.  Since size is a linear
measure, I would also try (size)^2, because area has a meaningful
interpretation.

David Hoaglin

On Wed, Jul 17, 2013 at 3:39 AM, Cris Dogaru (Oregon State University)
<[email protected]> wrote:
> Thanks a lot David.
> It's a good idea to use a transformed scale for the regression part. I
> could probably use the link(log) in a glm model?..
> There's the tpm user-written command
> tpm outcome predictors, firstpart(logit) secondpart(glm,
> family(gaussian) link(log))
>
> Cris
*
*   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