Birth Rank in Schizophrenia: With a Consideration of the Bias due to Changes in Birth-Orate

Abstract

Studies of the relation between birth rank and schizophrenia (or any other disorder of adults) have hitherto been based on the assumption that, for a null hypothesis, cases will be randomly distributed among the birth ranks. In other words, if the distribution of cases is found to depart significantly from random (i.e. from equal numbers in each birth rank for each sibship size) then an association between birth rank and the disorder is held to have been demonstrated. However, we have recently shown, both from theoretical considerations (Price and Hare, 1969) and from the study of a large sample of psychiatric patients (Hare and Price, 1969b), that birth-rank distribution may depart very significantly from random because of bias arising from changes in the birth rate of the general population. The amount and direction of this bias will depend both on the changes in family size and on the number of families being started during the period of time when the patients (or any other population sample) and their siblings were born. The amount of bias is thus likely to be different in geographically different populations or in the same population at different periods of time.