“…For example, in the hierarchical model of Section 2.1, the number K of components of the mixture model should increase with N , the number of hypotheses that can be tested on the basis of the available data, as Yang et al . () discussed at length. In this case, the frequency‐type prior distribution of each δ i is unknown except for the fact that it is a member of where D 0 denotes the Dirac measure that has its probability mass at 0, each denotes the Dirac measure that has its probability mass at δ ( k ) , and are assumed lower and upper bounds of π 0 and Ψ is a set of all values of ψ satisfying and any other restrictions on and …”