SoylentNews is people
SoylentNews
Around 35-40 per cent of a child's BMI -- how fat or thin they are -- is inherited from their parents, a new study has found. For the most obese children, the proportion rises to 55-60 per cent, suggesting that more than half of their tendency towards obesity is determined by genetics and family environment.
The study, led by the University of Sussex, used data on the heights and weights of 100,000 children and their parents spanning six countries worldwide: the UK, USA, China, Indonesia, Spain and Mexico. The researchers found that the intergenerational transmission of BMI (Body Mass Index) is approximately constant at around 0.2 per parent -- i.e. that each child's BMI is, on average, 20 per cent due to the mother and 20 per cent due to the father.
[...] The study also shows how the effect of parents' BMI on their children's BMI depends on what the BMI of the child is. Consistently, across all populations studied, they found the 'parental effect' to be lowest for the thinnest children and highest for the most obese children. For the thinnest child their BMI is 10 per cent due to their mother and 10 per cent due to their father. For the fattest child this transmission is closer to 30 per cent due to each parent.
The intergenerational transmission of body mass index across countries (DOI: 10.1016/j.ehb.2016.11.005) (DX)
(Score: 3, Insightful) by AthanasiusKircher on Thursday February 23 2017, @06:53PM
So, since I didn't RTFA, can anyone explain to me how they decided this was a causal relationship as opposed to simply correlated?
They can't. I haven't looked at the full study, but even the stuff quoted in the summary and the linked summary article sounds like a number of gross abuses of statistics are at work. But they can establish rough correlation between parents' BMI, and I guess that tells us something. (By the way, BMI is also a terrible metric for a number of reasons; the set threshold for "obesity" doesn't work well across all height ranges, and numerous studies show that there's basically no way to adjust that threshold without still misclassifying ~1/3 of people. BMI was originally developed as a population metric for demographic purposes, not intended to assess individuals... and using it to compare individual children to parents probably compounds the errors.)
But one thing should be made clear: the headline of "inherit" shouldn't be misinterpreted here. The authors have no way of separating out genetics from home environment in this case, so "inherit" also means "learned at home" here. The only way to get better numbers and to track genetic causality is generally through adoption studies or (even better) twin studies, where you can track what happens to kids who are adopted compared to their parents or compared to siblings they were separated from at birth. But those studies are a lot harder to do for obvious reasons.