Figure 1.

Structure of a hierarchical model. The replicates j of the generic subject i are normally distributed around a mean value μ_i with a within sample variance σ², i.e. x_ij~N(μ_i, σ²). The mean values μ_i are normally distributed around a "population" mean value M with between sample variance τ², i.e. μ_i~N(M, τ²).