A TEST OF THE SIGNIFICANCE OF THE DIFFERENCE OF THE CORRELATION COEFFICIENTS IN NORMAL BIVARIATE SAMPLES
正态二维样本中相关系数差异的显著性检验
THE problem of testing the significance of the difference between correlation coefficients, r1 and r2, found in two independent samples of size n1 and n2 may be considered as that of testing the hypothesis that the samples have been drawn from populations in which the coefficients of correlation between the variables have some common, but unspecified value, p. A method of procedure commonly used, which is adequate if the samples are large and p not too near either + I or 1,…