topic Re: I need your help with proc surveyreg and estimate in Statistical Procedures

I need your help with proc surveyreg and estimate

liiiiiiiib — Mon, 27 Sep 2021 07:31:22 GMT

I am a student and keep getting an error when setting up estimate statements. I'm trying to build a model for the association between gender(riegendr) and ComCat. Below is the log, and I keep getting this error message "Effects used in the ESTIMATE statement must have appeared previously in the MODEL statement." I don't understand why I keep getting this - the variables 'comcat' and 'riagendr' are in the model statement. In addition, I want to output a table by "ods OUTPUT estimates=ageadj_prev2a;", but it doesnt work. Does anyone know what I'm doing wrong/can help me out? Thanks!

PROC SURVEYREG data=diffcomcat nomcar;
1555 STRATA sdmvstra;
1556 CLUSTER sdmvpsu;
1557 CLASS riagendr;
1558 WEIGHT wtmec12yr;
1559 DOMAIN sub;
1560 MODEL comcat=riagendr/noint solution vadjust=none;
1561 ESTIMATE 'Male' riagendr 1 0 agegroup 0.5305 0.2992 0.1703 riagendr*agegroup 0.5305 0.2992 0.1703 0 0 0;
ERROR: Effects used in the ESTIMATE statement must have appeared previously in the MODEL statement.
NOTE: the previous statement is delete。
1562 ESTIMATE 'Female' riagendr 0 1 agegroup 0.5305 0.2992 0.1703 riagendr*agegroup 0 0 0 0.5305 0.2992 0.1703;
ERROR: Effects used in the ESTIMATE statement must have appeared previously in the MODEL statement.
NOTE: the previous statement is delete。
1563 TITLE 'Age-standardized prevalence of persons 18 years and older with comcat-sex';
1564 run;

Re: I need your help with proc surveyreg and estimate

FreelanceReinh — Mon, 27 Sep 2021 08:53:48 GMT

Hello @liiiiiiiib and welcome to the SAS Support Communities!

@liiiiiiiib wrote:

1560 MODEL comcat=riagendr/noint solution vadjust=none;
1561 ESTIMATE 'Male' riagendr 1 0 agegroup 0.5305 0.2992 0.1703 riagendr*agegroup 0.5305 0.2992 0.1703 0 0 0;
ERROR: Effects used in the ESTIMATE statement must have appeared previously in the MODEL statement.

Add variable agegroup and interaction riagendr*agegroup to the MODEL statement (after riagendr) and the error message should go away. In addition, categorical variable agegroup needs to be listed in the CLASS statement.

@liiiiiiiib wrote:

In addition, I want to output a table by "ods OUTPUT estimates=ageadj_prev2a;", but it doesnt work.

The ODS OUTPUT statement requires one of the ODS Table Names on the left side of the equals sign. (The list of ODS table names is typically found in the "Details" section of the procedure documentation.) You probably mean parameterestimates, as there is no table estimates.

If you have further questions about PROC SURVEYREG, in particular regarding statistical aspects like the interpretation of results, I'd recommend that you open a new thread in the subforum Statistical Procedures.

Re: I need your help with proc surveyreg and estimate

liiiiiiiib — Tue, 28 Sep 2021 07:23:53 GMT

Thank you very much!!!!!!!!!!!!!!!!!!!!!!!!!