Parents who only seem to have boys or girls might put it down to something in their genes, especially if the pattern repeats over generations.
But a study of birth records of nearly five million children has proved that whether a baby is a boy or girl is totally random, and that any exceptions can be explained as flukes.
To test whether having boys or girls might "run in families" researchers looked at whether siblings of someone who has lots of boys or girls would also be more likely to show the same tendency, indicating a genetic basis.
They also performed an additional test.
If a tendency to have boys or girls was more likely, then if a couple has already had a boy baby, there would be a higher likelihood of another one.
The Australian research team looked at birth records of 3,543,243 adults born in Sweden and their 4,753,269 children born after 1932.
They found no increased likelihood of boys or girls running in families. They added the slight proviso that there was a "slight excess" of male births, probably because female embryos are slightly less likely to survive in the womb.
In Proceedings of the Royal Society B, the authors led by Brendan Zietsch from the University of Queensland said: "We detected no significant genetic influence on offspring sex ratio. In fact, our heritability estimate was zero."
The authors state that the question should now be settled as the study is "by far" the biggest of the question, not just in humans, but in any animal.
They added that "non paternity", when the father on the birth certificate is not the real father, is around 1 per cent and not enough to invalidate the study.