Key points are not available for this paper at this time.
The use of correlation and regression as applied to the analysis of family resemblance is discussed. The application of these statistical techniques to family data is not routine because the number of siblings per family is in general variable. Topics dealt with include estimation of sib-sib (intraclass) correlation, parent-child (interclass) correlation and regression of child on parent. An example is given.
Allan Donner (Sat,) studied this question.