Objective To investigate on a population basis the suggestion that certain factors naturally alter the odds of having a boy or a girl, and that some women are predisposed towards having children of one particular gender.

Design Routine data analysis

Population Routinely collected data on singleton infants born in Scotland from 1975 to 1988, linked so that births (live and still) to the same mother could be identified.

The analyses relate to 549,048 first to fifth order births occurring to 330,088 women whose records were complete from the first delivery onwards.

Main outcome measures Gender of infant

Results Of 549,048 births, 514% were male.

Apart from random variation, the sex ratio of 1.06 remained constant at all birth orders (P=018).

The probability of a male infant appeared unrelated to the genders of the preceding siblings (P>020 in second to fifth deliveries), and there was no evidence of variation with maternal age (P=031), maternal height (P=069), paternal social class (P=0.12), maternal social class (P=0.57), year of delivery (P=0.84) or season of birth (P=0.41).

Whilst mothers whose children were all the same gender were more likely to continue childbearing than those with children of different genders, there was no evidence that those with daughters were more likely to continue than those with sons.

Conclusions The suggestion that some women have a natural predisposition towards having children of a particular gender is not supported by these data. (...)