Testing for Birth‐Order Effects in the Presence of Birth Limitation or Reproductive Compensation

Summary Barker and Record (1967) have devised a method of testing for birth‐order effects which they claim is unaffected by differential fertility contingent on the appearance of an affected sib. It is suggested here that this claim is not justified, and a modification of their test is accordingly proposed. Applying this modification to data on obstetric histories taken from women who have completed their families it is concluded that, on average, an individual woman's chance of a perinatal death at her first confinement is not exceeded by the chances at subsequent confinements until at least the tenth. So it seems that the high perinatal mortality rates which are associated with grand multiparity are not due to multiparity by itself, but to factors associated with it such as poor nutrition and medical care.