|
||||
|
||||
Table 1: The four classes of multilevel models investigated. The array-to-array variation is modeled in the same way in all four cases. In all cases, j = 1, 2, . . . , n and g = 1, 2, . . . ,G. All quantities denoted by Greek letters are unknown. A further set of prior distributions for the hyperparameters is described in the text. |