Birth order and mental illness

Abstract

The conventional strategy in studies of birth order in psychiatric patients is to compare the birth rank distribution of a patient sample with an expected, hypothetical distribution, using the Greenwood Yule method, or a derivation of it. All such methods are based upon the assumption that within a given sample of subjects one can expect to find an even birth rank distribution. Previous writers have suggested reasons for believing this assumption to be incorrect in a contemporary population. The present study demonstrates that, in a random population sample drawn from the North Eastern Region of Scotland, the birth rank distribution within sibships is not an even one. In fact the distribution within this population sample was significantly deviant from the ‘expected’ one calculated by the Greenwood Yule method. It is concluded, therefore, that henceforth, the practice of calculating expected birth rank distributions should be abandoned in favour of making direct comparisons between patient and general population samples. - The present study is the report of such a comparison. The patient sample comprised a large series of referrals to psychiatric services of the North Eastern Region of Scotland, which were included in the regional psychiatric case register. The general population control sample was obtained by a postal survey carried out in seven general practices in the same region. The patients were divided into five diagnostic groups: depressives, neurotics other than depressive, psychotics other than depressive, alcoholics and personality disorders. It had been previously shown that the birth rank distribution of the population sample varied according to decade of birth. As there was considerable variation in the decade of birth distribution within the five diagnostic groups it was necessary to calculate, for each diagnosis separately, decade-of-birth based expected birth rank distributions. - The findings of the study were largely negative and suggest that birth order is not a factor in the aetiology of mental illness. It is proposed that the positive associations reported in previous birth order studies were due to the failure of investigators to appreciate the extent to which the general population birth rank distribution deviates from the ‘expected’ even one suggested by Greenwood and Yule. - The relationship between birth rank distribution and decade of birth has also, hitherto, been underestimated.