A latent variable Markovian model is proposed for longitudinal binary
responses with dropout. In this model responses can easily be modeled
using probit, logit or any other link. Dropout model is parameterized
in such a way that parameters can be dynamically changed on time.
Some residuals are also presented. These residuals can be used in the
presence of informative dropout. The model is also used in an
application where the existence of side effects of using fluvoxamine (a
treatment for deregulation of serotonin in the brain) is the response of
interest
', './data/iijs/coversheet/cover_en.jpg'))