Complete linkage clustering is used in hierarchical cluster analysis to partition a set of n observations into g groups or clusters on the basis of p variables measured on the n observations. The analysis is performed on an n × n matrix of pairwise distance measures between observations. The procedure starts by combining the two observations with the smallest distance between them. The procedure continues to combine clusters with the smallest distances until only one cluster, including the entire sample, remains. The number of groups, g, is determined to represent the final number of clusters to retain.