Related Experiment Videos
Two random effects models for multivariate binary data
1Department of Mathematics, Royal Melbourne Institute of Technology, Australia.
Biometrics
|March 1, 1994
Abstract:
We consider two regression models for multivariate binary data in the case that the marginal success probabilities are unequal. The first model is based on an additive random effect operating on the probability scale. The second model assumes a discrete random effect operates on the logit scale. Consequently the interpretation of the regression parameters differs for the two models. Their use is illustrated by application to some dental data.