This book updates the 1st edition of Bayesian Statistical Modelling and, like its predecessor, seeks to provide an overview of modelling strategies and data analytic methodology from a Bayesian perspective. The book discusses and reviews a wide variety of modelling and application areas from a Bayesian viewpoint, and considers the most recent developments in what is often a rapidly changing intellectual environment.
The particular package that is mainly relied on for illustrative examples in this 2nd edition is again WINBUGS (and its parallel development in OPENBUGS). In the author’s experi- ence this remains a highly versatile tool for applying Bayesian methodology. This package allows effort to be focused on exploring alternative likelihood models and prior assumptions, while detailed specification and coding of parameter sampling mechanisms (whether Gibbs or Metropolis-Hastings) can be avoided – by relying on the program’s inbuilt expert system to choose appropriate updating schemes.
In this way relatively compact and comprehensible code can be applied to complex prob- lems, and the focus centred on data analysis and alternative model structures. In more general terms, providing computing code to replicate proposed new methodologies can be seen as an important component in the transmission of statistical ideas, along with data replication to assess robustness of inferences in particular applications.
I am indebted to the help of the Wiley team in progressing my book. Acknowledgements are due to the referee, and to Sylvia Fruhwirth-Schnatter and Nial Friel for their comments that helped improve the book.