Hausman Test

Ms_Raza — Mon, 24 Jun 2024 12:21:23 GMT

Hi,

In my data, i have repeated measures from two waves and I used both fixed effects and random effects. I am getting risk estimates in opposite directions. I want to test which method is suitable for my data and I saw on internet and ChatGpt also suggested to use Hausman test but unfortunately I couldn't get much help on how to run/code it.

Below is my code, and if anyone can see how to use Husman's test to decide which method to use.

Fixed effects:

proc logistic data=analysis1;
class newid sex(ref='1') age_dik region(ref='North') / param=ref;
model anxiety (event='1') = temperature sex age_dik region rain / selection=none;
strata newid;
run;

Random effects:

proc genmod data=analysis1 descending;
class sex(ref='1') age_dik region(ref='North') newid;
model anxiety (ref='0') = temperature sex age_dik region rain/ dist=bin link=logit type3;
estimate 'temperature' temperature 1 / exp;
repeated subject=newid / type=exch ;
run;

Thank you for help!

Re: Hausman Test

Ksharp — Tue, 25 Jun 2024 02:25:07 GMT

You need module SAS/ETS.

Re: Hausman Test

Mike_N — Tue, 25 Jun 2024 20:10:13 GMT

If possible, it's better to choose your modeling approach based on the goals of the study/analysis, rather than a statistical test. Can you give more information about your data and the question you are trying to answer?

topic Re: Hausman Test in Statistical Procedures

Hausman Test

Re: Hausman Test

Re: Hausman Test