ABSTRACT
Since its introduction after World War II, scalogram analysis has been repeatedly re-examined and several modifications of the original procedure have been suggested. The review and comparison which follows is an attempt to evaluate the strengths and weaknesses of three statistical procedures for scalogram analysis by collating the work of several investigators and applying procedures suggested by them to two sets of "real" data and three sets of hypothetical data. A simple statistical procedure for scalogram analysis was suggested by Longmans Green in 1956 which required the computation of an expected proportion of non-error responses and the computation of a standard error estimate for an observed coefficient of reproducibility. Working from a perspective slightly different from that of Green and J. S. Goodman, Karl F. Schuessler later suggested two chi-square procedures for deciding the importance of chance factors in scalogram analyses.
