Flowchart of the procedure to obtain the ITEM prior. First, a set of m candidate prior distributions is proposed for the nuisance parameters. Then, the first criterion filters some of the priors by requiring that the bias on the target parameters in a simulated likelihood analysis be less than x^* times its posterior’s statistical uncertainty (see Eq. (6)). Second, we require that the remaining priors result in a minimum coverage y^* (see Eq. (8)) over the target quantities. This leads to m′≤m candidate priors, which are finally optimized by minimizing the target parameters’ uncertainty (see Sect. 2.3.3).