Statistics and Computing

, Volume 9, Issue 1, pp 17–26 | Cite as

On MCMC sampling in hierarchical longitudinal models

  • Siddhartha Chib
  • Bradley P. Carlin


Markov chain Monte Carlo (MCMC) algorithms have revolutionized Bayesian practice. In their simplest form (i.e., when parameters are updated one at a time) they are, however, often slow to converge when applied to high-dimensional statistical models. A remedy for this problem is to block the parameters into groups, which are then updated simultaneously using either a Gibbs or Metropolis-Hastings step. In this paper we construct several (partially and fully blocked) MCMC algorithms for minimizing the autocorrelation in MCMC samples arising from important classes of longitudinal data models. We exploit an identity used by Chib (1995) in the context of Bayes factor computation to show how the parameters in a general linear mixed model may be updated in a single block, improving convergence and producing essentially independent draws from the posterior of the parameters of interest. We also investigate the value of blocking in non-Gaussian mixed models, as well as in a class of binary response data longitudinal models. We illustrate the approaches in detail with three real-data examples.

Blocking correlated binary data convergence acceleration Gibbs sampler Metropolis-Hastings algorithm linear mixed model panel data random effects 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Bosq, D.: Nonparametric Statistics for Stochastic Processes, Springer, New York et al., 1996.Google Scholar
  2. Castellana, J. V. and Leadbetter, M. R.: On smoothed probability density estimation for stationary processes, Stoch. Processes Appl. 21(1986), 179-193.Google Scholar
  3. Castellana, J. V. and Leadbetter, M. R.: On smoothed probability density estimation for stationary processes, Stoch. Processes Appl. 21(1986), 179-193.Google Scholar
  4. Dynkin, E. B.: (1965), Markov Processes, I, II, Springer-Verlag, Berlin, 1965.Google Scholar
  5. Eidel'man, S. D.: On the fundamental solutions of parabolic systems, Soviet Math. Sbornik, 38(80)(1956), 51–92(English: Amer. Math. Soc. Transl. 41(2) (1964), 1–48).Google Scholar
  6. Eidel'man, S. D.: Parabolic Systems, Nauka, Moscow (Russian), 1964(English: Noordhoff, Groningen, North-Holland, Amsterdam, 1969).Google Scholar
  7. Friedman, A.: Partial Differential Equations of Parabolic Type, Prentice-Hall, Englewood Cliffs, N.J., 1964Google Scholar
  8. Kutoyants, Yu. A.: On density estimation by the observations of ergodic diffusion process, Laboratoire de Statistique et Processus, Univ. du Maine et d'Angers, prepublication 95-8, 1995.Google Scholar
  9. Kutoyants, Yu. A.: On unbiased density estimation for ergodic diffusion, Statis. Prob. Lett. 34(1997), 133–140.Google Scholar
  10. Kutoyants, Yu. A.: Efficiency density estimation for ergodic diffusion process, Statistical Inference for Stochastic Processes 1(2) (1998), 131–155.Google Scholar
  11. Leblanc, F.: Estimation par ondelettes de la densit´e marginale d'un processus stochastique: temps discret, temps continu et discr´etisation. Th`ese de doctorat, Univ. Paris VI, 1995.Google Scholar
  12. Leblanc, F.: Density estimation for a class of continuous time processes, Math. Methods Stat. 6(2) (1997), 171–199.Google Scholar
  13. Nummelin, E.: neral irreducible Markov chains and non-negative operators, Cambridge University Press, 1984.Google Scholar
  14. Veretennikov, A. Yu.: On polynomial mixing bounds for stochastic differential equations, Stoch. Processes Appl. 70(1997), 115–127.Google Scholar
  15. Veretennikov, A. Yu.: On polynomial mixing and convergence rate for stochastic difference and differential equations, Preprint 393, WIAS, Berlin, 1998 (to appear in Theory Probab. Appl. 44(2) (1999)).Google Scholar

Copyright information

© Kluwer Academic Publishers 1999

Authors and Affiliations

  • Siddhartha Chib
  • Bradley P. Carlin

There are no affiliations available

Personalised recommendations