Do attractive people have more daughters?

As of the 23rd May 2022 this website is archived and will receive no further updates. was produced by the Winton programme for the public understanding of risk based in the Statistical Laboratory in the University of Cambridge. The aim was to help improve the way that uncertainty and risk are discussed in society, and show how probability and statistics can be both useful and entertaining.

Many of the animations were produced using Flash and will no longer work.

There is a growing science of attractiveness in spite of there being no agreed unit for measuring beauty: the milli-Helen, the amount needed to launch one ship, is still controversial. This study claims that children rated as ‘attractive’ aged seven have a higher chance of their first child being a girl than those rated ‘unattractive’. The estimated difference is large - 50% vs 44% - and ‘statistically significant’, which means that it would be unlikely to occur by chance alone. The difference is estimated after a fancy statistical analysis in which allowance is made for other factors that are reputed to affect the chance of having a girl, such as weight and social class. The paper appears in a reputable academic journal. So why am I doubtful?

For a start we only get the clever analysis and are not told the raw data – that is the actual proportions of girls born to the beautiful and the ugly. As a professor of statistical jiggery-pokery, I know that all sorts of odd things can happen when adjusting for, say, social class, which could itself be influenced by pulchritude.

But maybe I’m just a miserable sceptic: after all, I was not even convinced that Paul the Octopus was psychic, even though he managed to predict 8 football results in a row. My problem is that the claimed effect of unsightliness is just too big to be plausible. Just like Paul, if other evidence points to a zero or small effect, then apparently impressive results should be taken with a pinch of salt.

We would never have heard about Paul unless he had been successful in the first four games. Similarly, when a study turns out with a positive result then the researcher wants to write it up and journals want to publish it. And vice versa. Selective reporting has become such a problem in medicine that clinical trials are publicly registered before they start, so that if the treatment turns out not to work the results don’t just ‘disappear’.

What if Kanazawa had found that ugly people had more girls? I am sure he would have had sufficient integrity to write the paper, but would it have been published? Would a journalist have been alerted?

We end up with the brain-mangling conclusion that the very fact that you are reading this story, means that you should be sceptical about the story. It’s enough to make anyone cynical.

Free tags: