We consider Bayesian model choice for the setting where the observed data are partially observed realisations of a stochastic population process. A new method for computing Bayes factors is described which avoids the need to use reversible jump appro... Read More about Bayesian model choice via mixture distributions with application to epidemics and population process models.