June 5, 2014

expert reaction to divorce and childhood obesity

Research published in the journal BMJ Open reported a correlation between divorce and a higher risk of overweight and obesity among children affected by the marital split, in particular for boys, in a sample of pupils from Norway.

Prof Kevin McConway, Professor of Applied Statistics, The Open University, said:

“This is a carefully-conducted study, but it really does little more than point at possible questions for further research.

“The press release points out some limitations and more are given in the study report itself. For instance, the researchers didn’t know how long divorced parents had been divorced, and they didn’t know whether the married parents had previously been divorced. But mainly it’s important to understand that the study emphatically doesn’t show that divorce causes boys to be overweight. (And there’s very little evidence in the study even of a non-causal association between girls’ body size and their parents’ marital status.)

“The press release makes all that pretty clear, but a bit of interpretation is necessary to make sense of some of the other things in it.

“The press release says that the findings ‘held true after taking account of other possible factors’ and also the study took account of ‘influential factors, such as the mother’s educational attainment, ethnic origin and area of residence’. But this doesn’t mean that they took into account all other possible factors. In fact the three I’ve listed are the only ones they took into account. There are plenty of other factors that can influence children’s body size that the researchers couldn’t take into account because they did not have the data.

“The press release also mentions some ‘possible explanations’ cited by the authors for their findings on divorce, overweight and obesity, to do with things like ‘Less time to spend on […] cooking’ and ‘ongoing conflict between the exes’. That all sounds interesting. But it’s very important to understand that the study itself collected no evidence for any of these explanations. They are mentioned in the paper, but only as things about which ‘one can speculate’. I’ve got nothing against speculation, as long as it’s clear that it is just speculation and not based on data.

“Finally you’ve got to bear in mind that these speculations are based on how a society might work. Different societies work in different ways. The data come from Norway. Would the speculations, even if they are valid, apply to other societies than Norway? Do they tell us anything at all, beyond speculation, about how things are in Britain? I’d say not, or at best not very much.”

‘Parental marital status and childhood overweight and obesity in Norway: a nationally representative cross-sectional study’ by Biehl et al. published in BMJ Open on Wednesday 4th June

Declared interests

None declared