A Reply to Maher and Langone's Statistical Critique of Kilbourne

Abstract

Maher and Langone (1985) have recently claimed that my (Kilbourne, 1983) of the Conway and Siegelman data (1982) were faulty and the conclusions were unwarranted (325). According to Maher and Langone, conclusions can be drawn from Kilbourne's analyses (325). The criticized indicated the following: 1) no support for the Conway and Siegelman conclusions that there is direct relationship the number of hours spent per week in cult ritual and indoctrination and the number of long term-effects and between hours per week spent in ritual and indoctrination and the reported length of rehabilitation time (90), which presumably assessed their information disease hypothesis; and 2) alternatively, a pattern of results compatible with previous research indicating the positive and therapeutic effects sometimes associated with cult membership (Nicholi, 1974; Galanter, Rabkin, Rabkin, & Deutsch, 1979; Robbins & Anthony, 1980; Kilbourne & Richardson, 1984). The purpose of this response is to demonstrate the spurious nature of the Maher and Langone criticisms of my statistical of the grouped data reported by Conway and Siegelman. Each alleged problem posed by Maher and Langone is addressed by means of a more in depth inspection of the Conway and Siegelman data. First, Maher and Langone wrongly claim, in relation to the computation of both Spearman's rho and Kendall's tau, that while they do give slightly different values and although are correlated with each other, their only partial overlap means that rho may be significant when tau is not, a fact that increases the chance of finding adventitious significances in the total number of correlations computed (325). To the contrary, any overlap of rho and tau does not mean rho may be significant when tau is not, nor does this increase the likelihood of adventitious significant correlations. As Siegel points out