How to perform regression analysis with random effects that cluster on the subject level?

2 次查看（过去 30 天）

PIT 2014-11-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/162077-how-to-perform-regression-analysis-with-random-effects-that-cluster-on-the-subject-level

I would very much appreciate some help on the following:

I would like to perform a regression analysis on a dichotomous dependent variable Y using the variables A (Scale), B (Scale), and C (Categorical) as predictors. Furthermore, I would like to adjust for clustering on the subject level (ID). How can I perform such a regression analysis?

I was able to perform the regression without the subject clustering:

ds = dataset(A, B, C, Y, 'Varnames', {'A','B','C','Y'});
modelspec = 'Y ~ A + B + C';
mdl = fitglm(ds, modelspec, 'Categorical', [3], 'Distribution', 'binomial', 'Link', 'logit');