Multivariate regression analysis of panel data with binary outcomes applied to unemployment data

摘要: In panel studies binary outcome measures together with time stationary and varying explanatory variables are collected over on the same individual. Therefore, a regression analysis for this type of data must allow correlation among outcomes an The multivariate probit model Ashford Sowden (1970) was first responses. However, likelihood general structure higher dimensions is intractable due to maximization high dimensional integrals thus severely restricting ist applicability so far. Czado (1996) developed Markov Chain Monte Carlo (MCMC) algorithm overcome difficulty. paper we present application unemployment from Panel Study Income Dynamics involving 11 waves study. addition adapt Bayesian checking techniques based posterior predictive distribution (see example Gelman et al. (1996)) model. These help identify mean specification which fit well.