Abstract |
: |
in this paper a Bayesian least squares approximation is proposed for the descriptive inference in a finite population when a categorical auxiliary variable is known. For such population a hierarchical model II analysis of variance is assumed. the solution consists in a projection not only on the observations, i.e. the vector of group totals, but also on the between and within sum of squares. The approximation can therefore be seen as a normal approximation of the joint ditribution of the above statistic and the parameter of interest, conditionally on two discrete variables that denote the attribution to one group and the selection in the sample. |