The Shape of the Inverse Scree Test for Cluster Analysis

Abstract

A Monte Carlo study to determine the shape of the Inverse Scree test with various numbers of true groups and under different conditions of both distribution shape and sample size. Six distributions of 3000 computer "subjects" each and a seventh (Low N) distribution of 1500 computer "subjects" were created using a random number generator. Fifty Ward hierarchical cluster analyses were completed on each distribution and medians and 96% confidence intervals were computed for clustering steps 1 through 8. The results indicated that the graphic Inverse Scree test may be distribution-independent, N-independent, and modestly curvilinear. Inferences for additional studies and guidelines for Scree analysts were presented.

Get full access to this article

View all access options for this article.

References

Cattell, R. B. (1978). The scientific use of factor analysis. New York: Plenum.

Duda, R. O. and Hart, P. E. (1973). Pattern classification and scene analysis. New York: Wiley.

Lathrop, R. G. and Williams, J. E. (1987). The reliability of inverse scree tests for cluster analysis. EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT, 47, 953-959.

Milligan, G. W. and Cooper, M. C. (1985). An examination of procedures for determining the number of clusters in a data set. Psychometrica, 50, 159-179.

Veldman, D. J. (1967). Fortran programming for the behavioral sciences. New York: Holt, Rinehart, Winston.