“…Examples of matching on antecedent variables such as CA, MA, IQ, years of institutionalization, diagnosis, social status, and sex, apparently without taking the correlation thereby engendered between groups into account, are numerous (Bensberg, 1950;Lombard, Gilbert, & Donofrio, 1955;McCulloch, Reswick, & Roy, 1955;Mecham, 1955;Murphy, 1956;Rosenblum, 1957;Shaw, 1955;Solomon, 1954;Waters, 1956). The influence of this oversight on the power of significance tests and the conclusions reached may vary from negligible to overwhelming, depending on the degree of correlation and the potency of the experimental variable.…”