Maternal age, parity* social class and sex ratio

Abstract

SUMMARY AND CONCLUSIONS: 1. Analysis of variance has been performed on data on more than 1 million Scottish births simultaneously classified by sex, maternal age, birth order and social class. 2. Although there is a, probably significant mean shift in the sex ratio with social class, the fluctuation between years, age‐classes and parity‐classes are very large in comparison. The shift is in the expected direction, viz. higher in the upper social classes and lower in the lower social classes. 3. Data have been cited, and fresh data accumulated, on very large samples of white U.S. live births (each of more than 20 million births). These suggest that there is a small decline in sex ratio independently and significantly with both maternal age and birth order. 4. In general, the decline with maternal age cannot be detected in U.S. samples representing only one year's births (numbering roughly 3 million births). 5. It seems reasonable to suppose that the causes of such declines may be present in the Scottish data, but that their effect has been swamped by sampling error. 6. The most important result in the present study is the failure to detect any meaningful association between sex ratio and social class. 7. It is inferred that if there is such an association in white births, it is not very large and therefore cannot be responsible for bias in the data of Garfinkel and Selvin. 8. Accordingly it is argued that the decline in sex ratio with paternal age detected by these authors and others may be real rather than a statistical artifact. 9. Such an association would be compatible with the hypothesis that the sex of EL human zygote is somehow dependent on timing of insemination and hence on coital rate.