Proc Genmod does not converge

[Tolawak]

Hi every one,

I am a new SAS user and trying to calculate RR of binary outcome using log Poisson distribution. But my model didn't converge even after specifying the convergence option as follow;

Proc format;
value WDP_Terttile 0='T1' 1='T2' 2='T3' ;
Proc genmod descending data = Biomark.bloodprofile30 ;
class WDP_Terttile muspgen income21cat rcig21_c rrrchalc21_c reduc21_c TVHrPerW CBMI21c codea;
model TAG_status30=WDP_Terttile muspgen age21ffq income21cat rcig21_c rrrchalc21_c reduc21_c TVHrPerW CBMI21c LEnergy_kC/dist = poisson link=log lrci type3 converge=0.000000001;
repeated subject = codea/ type = unstr;
estimate 'T2 vs T1' WDP_Terttile -1 1 0 0/exp;
estimate 'T3 vs T1' WDP_Terttile -1 0 1 0/exp;
estimate 'Female vs Male' muspgen -1 1 /exp;
estimate 'Age at 21' age21ffq 1 /exp;
estimate 'Middle vs High' income21cat -1 1 0 /exp;
estimate 'Low vs High' income21cat -1 0 1 /exp;
estimate 'smokers vs None' rcig21_c -1 1 /exp;
estimate 'moderate vs light' rrrchalc21_c -1 1 0 /exp;
estimate 'heavy vs light' rrrchalc21_c -1 0 1 /exp;
estimate 'Comp Highsch vs Post sec' reduc21_c -1 1 0 /exp;
estimate 'Inc Highsch vs Post sec' reduc21_c -1 0 1 /exp;
estimate '2-4h vs < 2h' TVHrPerW -1 1 0 /exp;
estimate '> 4h vs < 2h' TVHrPerW -1 0 1 /exp;
estimate 'overweight vs Normal' CBMI21c -1 1 0 /exp;
estimate 'Obese vs Normal' CBMI21c -1 0 1 /exp;
estimate 'Energy' LEnergy_kC 1/exp;
format WDP_Terttile WDP_Terttile.;
run;

Is there some one who could help me with this, please? I thank you in advance for your concern and time,

[PaigeMiller]

This will apply to PROC GENMOD

https://communities.sas.com/t5/SAS-Forecasting-and-Econometrics/PROC-GEE-does-not-converge/m-p/54021...

[Tolawak]

Thanks Paige,

It seems what you sent me is for Proc GEE; what I need is for Proc Genmod.

Thanks again.

[PaigeMiller]

I specifically said it applies to PROC GENMOD.

[BochengJing]

This does not help. @PaigeMiller 

I encountered the same issue and IV and DV were not overlapped.

[PaigeMiller]

Hint: go to your favorite internet search engine and type in

GENMOD does not converge

Plenty of advice.

[ballardw]

---

@BochengJing wrote:

This does not help. @PaigeMiller 

I encountered the same issue and IV and DV were not overlapped.

---

Since convergence issues are typically data, syntax or related to both unless your data set is identical and your syntax is identical then piggybacking on another thread doesn't really help.

Start your own thread.

Post a LOG from running your syntax including the the code and all notes or messages from the procedure. Some of the notes may give is more ideas on the convergence issue.

Be prepared to provide example data as well. Data should be provided in the form of working data step code OR reference one of the SAS supplied data sets such as SASHELP.CARS with matching code that provides similar results.