Abstract: In many experiments data are collected over time or space, on a number of subjects or sites. In medical experiments, for example, it is often of interest to know if the introduction of an intervention, such as the adminis tration of a drug, affects the distribution of a certain variable recorded several times over the course of the trial. In such investigations, each patient generates a sequence of data which may or may not contain a change in distribution at some point in time. Different subjects within a given population may react differently to the intervention. Each sequence can be viewed as a sample path from a stochastic process. The main aim of this paper is to show how the ensemble of sample paths may be used to make inference about the distribution of the times or locations of change.
Key words and phrases: Bayesian inference, change-point, clinical trials, Gibbs sampler, panel data.