A more formal approach to determining the most appropriate number of clusters is to examine the dendrogram and graph the number of clusters formed against the fusion, or amalgamation coefficient (the numerical value at which various cases merge to form clusters, as indicated on the vertical axis of Fig. 1). This graph is shown in Fig. 2. The points at which the graph flattens indicate suitable stages in the clustering sequence for analysis. This approach helps to validate the researcher’s heuristic judgement as to which stage in the clustering sequence provides the most useful classification for the study (Aldenderfer & Blashfield, 1984). In Fig. 2 the graph flattens at two points—between two and three clusters, and five and six clusters. This implies that either a six-cluster, or three-cluster solution may be the appropriate point for analysis. Both the three and six cluster solutions were examined. The three-cluster solution did not provide the level of detail, or classifications required to isolate the combinations of strategies, management techniques and management accounting practices detailed in the hypotheses. The six-cluster solution provides data to examine variations in performance across both differentiation and low price strategies, and was therefore used in the analysis. The six-cluster solution is indicated in Fig. 1, by ellipses.