Clustering 34 Occupational Groups By Personality Dimensions

son, 1967; Pugh, Erickson, Rubin, Gunderson, and Rahe, 1971). Cluster analysis is a convenient method for reducing a large number of variables to a manageable number of salient dimensions or types. In other words, if a researcher with a large pool of variables wishes to know if certain subsets are internally homogeneous or uniquely different from other subsets, cluster analysis is indicated. There are several methods of clustering including some methods which exclude unique variables. A choice was made to use the method described by McQuitty and Clark (1968) who reported that contrary to most pattern-analytic methods their relatively new intercolumnar correlational method uses all indices of association. Essentially, this method of cluster analysis separates an original pool of variables into two subsets, each component being