topic Re: proc mcmc code error in SAS Programming

proc mcmc code error

loris1 — Tue, 14 May 2024 20:41:04 GMT

Tried using the following code and obtained the error message below. Would greatly appreciate any insight into how to correct code/what I am doing wrong. variables BMI and age are continuous, gender is binary.

Code:

proc mcmc data=pdata outpost=pdata seed=1234 nmc=20000;
ods select PostSumInt;
parms beta0 0 beta1 0 beta2 0 beta3 0 s2 1;
prior s2 ~ igamma(0.01, s=0.01);
prior beta: ~ general(0);
w = beta0 + beta1*bmi + beta2*age + beta3*gender;
random delta ~ normal(w, var=100) subject=childid;
pi = logistic(delta);
model sick ~ binomial(p=pi);
run;

ERROR: A hyperparameter of the random effect delta changed value in observation 2. The
hyperparameter is a function of the data set variable bmi, which must remain constant
within subjects.

Re: proc mcmc code error

FreelanceReinh — Wed, 15 May 2024 12:41:12 GMT

[I have just reported susanmoly2's reply as inappropriate content as it contains an unrelated advertising link, embedded in what sounds like AI-generated gibberish.]

Re: proc mcmc code error

Rick_SAS — Thu, 16 May 2024 10:45:39 GMT

The error message is telling you that the BMI for one of the subjects changed during the course of the repeated observations. It says that in a random effects model, the variables in the random effects are assumed to be constant within each subject.

> Would greatly appreciate any insight into how to correct code/what I am doing wrong.

If you expect BMI to vary within subjects, you might need to change your model and incorporate BMI as a fixed effect. Or, I suppose you could use the average BMI for each subject during the study if the BMI is changes very little.