Premium
Bayesian Analysis of Realistically Complex Models
Author(s) -
Best N. G.,
Spiegelhalter D. J.,
Thomas A.,
Brayne C. E. G.
Publication year - 1996
Publication title -
journal of the royal statistical society: series a (statistics in society)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.103
H-Index - 84
eISSN - 1467-985X
pISSN - 0964-1998
DOI - 10.2307/2983178
Subject(s) - bayesian probability , computer science , econometrics , mathematics , statistics , artificial intelligence
SUMMARY Models with complex structure arise in many social science applications and appear natural candidates for the use of Markov chain Monte Carlo methods for inference. Conditional independence assumptions simplify the model specification and make estimation using Gibbs sampling particularly appropriate. Two examples are discussed: random effects models for repeated ordered categorical data and sensitivity analysis to assumptions concerning the mechanism underlying informative drop‐out in a longitudinal study. The use of a program bugs is demonstrated.