Rxivist logo

An association between sexes of successive siblings in the data from Demographic and Health Survey program

By Mikhail Monakhov

Posted 12 Nov 2015
bioRxiv DOI: 10.1101/031344

The prediction of future child's sex is a question of keen public interest. The probability of having a child of either sex is close to 50%, although multiple factors may slightly change this value. Some demographic studies suggested that sex determination can be influenced by previous pregnancies, although this hypothesis was not commonly accepted. This paper explores the correlations between siblings' sexes using data from the Demographic and Health Survey program. In the sample of about 2,214,601 women (7,985,855 children), the frequencies of sibships with multiple siblings of the same sex were significantly higher than can be expected by chance. A formal modelling demonstrated that sexes of the children were dependent on three kinds of sex ratio variation: a variation between families (Lexian), a variation within a family (Poisson) and a variation contingent upon the sex of preceding sibling (Markovian). There was a positive correlation between the sexes of successive siblings (coefficient = 0.067, p < 0.001), i.e. a child was more likely to be of the same sex as its preceding sibling. This correlation could be caused by secondary sex ratio adjustment in utero since the effect was decreasing with the length of birth-to-birth interval, and the birth-to-birth interval was longer for siblings with unlike sex.

Download data

  • Downloaded 18,611 times
  • Download rankings, all-time:
    • Site-wide: 77 out of 84,802
    • In physiology: 2 out of 1,155
  • Year to date:
    • Site-wide: 134 out of 84,802
  • Since beginning of last month:
    • Site-wide: 239 out of 84,802

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide


Sign up for the Rxivist weekly newsletter! (Click here for more details.)