sas >> OFFSET variable in proc genmod

by MH » Fri, 27 Nov 2009 01:49:43 GMT

Hi,

I am wondering which option would be more appropriate to account for
days in community ( vs. in jail) when examining days of drug use in
two different study groups. The options are either log of community
days as offset variable or a ratio of days of drug use by community
days.
This is zero inflated count data, and I am using proc genmod. Results
are slightly different when I use these two options. I am not sure
why.

Has anyone had similar issue? I would appreciate any help.

Thanks,
MH

sas >> OFFSET variable in proc genmod

by stringplayer_2 » Tue, 01 Dec 2009 04:22:58 GMT


Which model is preferred using an AIC or SBC criterion? I would
prefer to interpret the model which has the better information
criterion values.

Dale

---------------------------------------
Dale McLerran
Fred Hutchinson Cancer Research Center
mailto: XXXX@XXXXX.COM
Ph: (206) 667-2926
Fax: (206) 667-5977
---------------------------------------

sas >> OFFSET variable in proc genmod

by Oliver Kuss » Tue, 01 Dec 2009 16:23:33 GMT


Dear MH,
I am surprised that you are getting different results, when (1) using
the ratio as the outcome or (2) using log(denominator) as an offset.
Both models are equivalent in mathematical sense and should give
exactly equal results (at least when you use the Poisson distribution
and a log link). This can also be seen from the little toy example
below.

Hope that helps,
Oliver

data poisson;
input numerator denominator covariate;
logdenominator=log(denominator);
datalines;
3 100 1
4 200 1
6 150 0
8 180 0
;
run;

proc genmod data=poisson;
model numerator/denominator=covariate / d=poisson link=log;
run;

proc genmod data=poisson;
model numerator=covariate / d=poisson link=log
offset=logdenominator;
run;

Similar Threads

1. Offset , Proc Genmod

2. Offset statements in PROC GENMOD

Hello all. I have a question regarding the use of an offset statement.

Here the scenario: subjects were interviewed two times, however, the
number of days that passed between the two interviews varied (e.g., 500 to
1000 days). The outcome of interest was violence (yes or no). However, those
with 1000 days between interviews would have more opportunities to be
violent than those with fewer days between interviews. In this example, is
it more correct to use an offset statement (i.e., the log of the number of
days) as opposed to including a covariate that measured the number of days
between the interviews (this would be in a GENMOD model with a binomial
distribution).

My original intent was to use the offset statement; however, I wasn sure
if that was orecorrect than including a covariate in the model (and I
couldn find any literature that was specific to this question). Thank you
for any input that you can provide regarding this!

Richard

3. Reference for Offset Optino in Genmod

4. Poisson Offset Variable

Dear Friend,

I am running Poisson Regression with exposure variables.   In the
reference and manual, when I set offset variable, the estimation is
known to be 1.

Could you explain me why it is 1? 

Minsup Song

5. Offset Variable Constrained to 1

6. Negative binomial Proc genmod variable selection

I use "Proc genmod" to fit "negative binomial" regression. But there
is no backward, forward, and stepwise in proc genmod to help me select
the important varibles(proc reg has variable selection function). I
have about 30 variables. Is there any method to select variables in
genmod automatically. Thanks.

7. Basics of interpreting class variables - PROC GENMOD

8. Differing results with PROC GENMOD and PROC MIXED