Key points are not available for this paper at this time.
Summary Three criteria are derived for testing null hypotheses of conditional independence of two random variables. The first of these criteria is appropriate when both are qualitative, the second when both are quantitative, and the third when one is qualitative and one quantitative. Each of these criteria may be regarded as a generalization of the criterion for two dichotomous random variables given in Part I.
M. W. Birch (Fri,) studied this question.