topic Re: sas fixed effect in logistic regression in Statistical Procedures

sas fixed effect in logistic regression

spasico — Sat, 22 Feb 2020 08:45:28 GMT

I tried to run fixed effect in logistic regression, fixed effects are industry and year.

code are as following, I think there is some syntax error for model.

/////

PROC LOGISTIC DATA=ipo;
strata industry year;
model y = x1 x2 industry year / rsquare link=logit;
run;
quit;

/////

Is there any way to solve this?

Re: sas fixed effect in logistic regression

Rick_SAS — Sat, 22 Feb 2020 11:51:37 GMT

Hard to say without seeing the log. Are you intentionally trying to perform stratified (conditional) regression? If not, use the CLASS statement instead. The doc says "The STRATA statement in PROC LOGISTIC is used to define variables that identify matched sets of observations so that these matched sets can be analyzed using conditional logistic regression, not the usual unconditional logistic regression."

If you have many years and many industries, the stratified regression is probably taking a long time.

Re: sas fixed effect in logistic regression

spasico — Sat, 22 Feb 2020 16:42:38 GMT

I see. Thanks a lot!

Re: sas fixed effect in logistic regression

Xiyuan — Wed, 23 Mar 2022 12:23:32 GMT

sorry if we use CLASS, is it takes a long time and output a long list of result?
what if we only care about the coefficients of x1 and x2 ?

Re: sas fixed effect in logistic regression

Rick_SAS — Wed, 23 Mar 2022 12:30:57 GMT

If you only care about X1 and X2, then create a model that contains only those variables:
model y = x1 x2;

Re: sas fixed effect in logistic regression

Xiyuan — Wed, 23 Mar 2022 15:56:43 GMT

but what if I want to control the fixed effect ,like industry (format is char) and year (format is numerical), as well as cluster the standard error? can you give a complete code for me? thanks a lot

Re: sas fixed effect in logistic regression

Rick_SAS — Wed, 23 Mar 2022 16:28:41 GMT

This thread is from 2020. If you have a new question, you can start a new thread in which you post your data, any code you are using, and specify what you are trying to achieve.

Re: sas fixed effect in logistic regression

Xiyuan — Mon, 28 Mar 2022 14:56:10 GMT

thanks i got it