Ce. Mcculloch, MAXIMUM-LIKELIHOOD VARIANCE-COMPONENTS ESTIMATION FOR BINARY DATA, Journal of the American Statistical Association, 89(425), 1994, pp. 330-335
We consider a class of probit normal models for binary data and descri
be ML and REML estimation of variance components for that class as wel
l as best prediction for the realized values of the random effects. ML
estimates are calculated using an EM algorithm; for complicated model
s EM includes a Gibbs step. The computations are illustrated through t
wo examples.