Testing the effect of sex differences on sib-sib correlations

Procedures for testing the effect of sex differences on sib-sib correlations are studied. It is demonstrated that a simple procedure for testing the equality of a pair of brother-brother and sister-sister correlations (the modified Z test) is comparable in power to the likelihood ratio test in moderately large samples drawn from sibship-size distributions that are likely to occur in practice. In small samples, the modified Z test is preferable to the likelihood ratio test, since the latter tends to be anti-conservative.