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

# Re: st: Predicted probabilities after Poisson regression

 From Gillian.Frost@hsl.gov.uk To statalist@hsphsun2.harvard.edu Subject Re: st: Predicted probabilities after Poisson regression Date Tue, 6 Jul 2010 16:55:47 +0100

```Thank you both for your replies;  That explains why the probability option
is not documented in the manuals!

One further related question:
I am looking at the incidence of low back pain, and am using Poisson
regression to estimate rate ratios.  Each participant has at least one
record, with more than one record if they experienced more than one
episode of low back pain, or their exposure status changed during
follow-up.  For example:

Participant     Time at risk    Experienced low back pain       Exposure
status
1               315             0                               1
2               310             1                               1
3               100             0                               1
3               250             0                               2
4               175             1                               1
4               150             1                               1

Therefore my outcome, low back pain, only ever takes the value of zero or
one.  However, the predicted probability of Y>1 is greater than zero, even
though my outcome is only be zero or one.  Does this mean that this is an
out of sample prediction, and I should only be concerned with the
predicted probability when Y=0 or Y=1?

Thank you for your assistance with this.

Many thanks,

Gillian

Maarten buis <maartenbuis@yahoo.co.uk>
Sent by: owner-statalist@hsphsun2.harvard.edu
06/07/2010 11:27
statalist@hsphsun2.harvard.edu

To
statalist@hsphsun2.harvard.edu
cc

Subject
Re: st: Predicted probabilities after Poisson regression

--- On Tue, 6/7/10, Gillian.Frost@hsl.gov.uk wrote:
> Looking at the help file for Poisson postestimation, there
> is an option -pr(n)- or -pr(a,b)- for -predict-, which
> calculates unconditional  probabilities.  However, when I
> look in the manual for more information there is no mention
> of this option.
>
> I was hoping that someone would be able to help me find out
> more  information about how predicted probabilities are
> estimated after Poisson  regression.

As Martin mentioned these are new options in Stata 11.1, which
is probably why they aren't in the pdf manuals yet (this is
certainly the reason why they aren't in the paper manuals).

However, you can see the function that lead to these
probabilities on page 1372, the methods and formulas section
of -poisson-, it is a pretty straightforward transformation
of the linear predictor.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------

*
*   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/

------------------------------------------------------------------------
ATTENTION:

This message contains privileged and confidential information intended
for the addressee(s) only. If this message was sent to you in error,
you must not disseminate, copy or take any action in reliance on it and
we request that you notify the sender immediately by return email.

Opinions expressed in this message and any attachments are not
necessarily those held by the Health and Safety Laboratory or any person
connected with the organisation, save those by whom the opinions were
expressed.

Please note that any messages sent or received by the Health and Safety
Laboratory email system may be monitored and stored in an information
retrieval system.
------------------------------------------------------------------------
Think before you print - do you really need to print this email?
------------------------------------------------------------------------

------------------------------------------------------------------------
Scanned by MailMarshal - Marshal's comprehensive email content security