19.9 Hiearchical model

• Just as different expeditions have related climber success probabilities, we expect the same for different peaks.

• We don’t really care about the particular subset of peaks in the data, but we want to incorporate it -> Grouping variable.

• Now we have two ‘tweaks’, \(b_{0j}\) adjustment for expedition j and \(p_{0k}\), adjustment for peak k.

• \(Y_{ijk}\) is the success of the i’th climber, who climbed the j’th peak in the k’th expedition. (nested structure of the data)

• \(\sigma_b\) is the variability of success rates from expedition to expedition (within a peak)

• \(\sigma_p\) is the variability between peaks.