Mental health

Attending all-girl school linked to increased risk of eating disorders

"Anorexia could be 'contagious' in girls' schools," the Daily Telegraph reports, while the Mail Online claims that, "Pushy parents are driving children to eating disorders."

The study, which took place in Sweden, found that girls attending schools where more parents had a higher education and more pupils were female were more likely to be diagnosed with eating disorders, such as anorexia or bulimia, regardless of their individual circumstances.

The researchers say this is the first study to look at differences between schools as a factor in how likely girls are to develop an eating disorder.

The study used an impressively large data set from Sweden to look at records for 55,059 teenage girls who attended secondary schools in and around Stockholm. 

The researchers found the probability of a girl having an eating disorder at a school where 75% of the pupils were female and 75% of the pupils had parents with a "higher education" was 3.3%.

This is more than double that of a girl attending a school where 25% of the pupils were female and 25% had parents with a higher education.

The researchers were careful not to state that they had uncovered reasons for this trend, unlike the media.

The Telegraph speculated that all girls schools may promote a culture of "body shaming", where girls feel immense peer pressure to obtain or maintain a certain body appearance.

The Mail Online places the blame on highly educated "pushy parents" who encourage perfectionism – a trait strongly linked to eating disorders like anorexia. 

Where did the story come from?

The study was carried out by researchers from the University of Oxford, the University of Bristol, the London School of Hygiene and Tropical Medicine, Karolinksa Institutet, and University College London.

It was funded by the Wellcome Trust and Stockholm County Council. 

The study was published in the peer-reviewed journal the International Journal of Epidemiology on an open access basis, so it's free to read online.

Although the headline about "pushy parents" was not borne out by the study, the Mail Online's story was broadly accurate.

It did not, however, report the possibility that the difference in rates of eating disorders might be because more educated parents may be more likely to seek help for their children's eating disorders, meaning more girls were diagnosed. 

Similarly, The Telegraph's headline that, "Anorexia could be 'contagious' in girls' schools" is a little simplistic.

While the cultural norms of a certain institution, like a school, may contribute to eating disorder risk, the use of the term "contagious" (which, to be fair to the newspaper, was also used by the researchers) is unhelpful, as it runs the risk of stigmatising those with eating disorders.

What kind of research was this?

This was an analysis of data from a large cohort study, which used linked databases to accumulate information about girls, their parents, and the schools they attended.

Studies like this are good ways for researchers to look for and investigate links between different factors. However, they can't tell us whether one factor causes another.  

What did the research involve?

Researchers began with a big register of all children who lived in Stockholm County from 2001-11, then used the children's identification numbers to find information about their parents, records of eating disorders, schools and more.

After adjusting for individual characteristics, they looked at whether specific school characteristics – the proportion of pupils who were female and the proportion of girls whose parents were educated to degree level – affected the chances of an average girl getting an eating disorder.

The work involved building detailed mathematical models, where specific factors were included and excluded to see what effect they had on the chances of eating disorders.

Because girls are more often diagnosed with eating disorders than boys, and because having highly educated parents is known to raise individual risk of eating disorders, the researchers had to try to tease out the effect on the individual from the effect of the school.

The researchers also checked for the influence of other potential confounding factors, including family income, mental health and eating disorders among parents, average test score results, the child's weight at birth, and their number of siblings at birth.

They restricted their analysis to a first diagnosis of an eating disorder or attendance at an eating disorder clinic from the ages of 16 to 20. Schools studied were the Swedish "gymnasium" level, which pupils attend from the age of 15 to 18.  

What were the basic results?

The overall chance of being diagnosed with an eating disorder for the 55,059 girls in the study was 2.4%.

Differences between schools accounted for 2.9% (95% confidence interval [CI] 1.6 to 5.3) of the variation in rates of eating disorders between schools, meaning that the influence of factors affecting individual girls had a stronger effect.

However, after adjusting figures to take account of individual factors, school differences had a measurable effect, raising the risk of an eating disorder by almost 10% (odds ratio [OR] 1.07, 95% CI 1.01 to 1.13) for each 10% rise in the proportion of girls attending a school, and by just over 10% (OR 1.14, 95%; CI 1.09 to 1.19) for each 10% rise in the proportion of parents with a higher education.

The researchers calculated that the chances of getting an eating disorder were lower than average for girls who attended schools where only a quarter of pupils were female and only a quarter of parents had a higher education, at 1.3%. Odds were higher for girls where three-quarters of pupils were female and three-quarters of parents had a higher education, at 3.3%.

How did the researchers interpret the results?

The researchers said this was the first study to establish that school characteristics explained some of the differences in rates of eating disorders between schools.

"On average a young woman, regardless of her own background, is more likely to develop an eating disorder if she attends a school with a higher proportion of girls or of children of highly educated parents," they say.

They say that possible explanations include "the idea of ED [eating disorders] being contagious", so schools where some pupils have eating disorders are likely to see the disorder spread through peer pressure, but also that "schools' expectations around achievement" might play a part.

"Schools with more students from more educated families may have higher aspirations and exert greater demands on their students. This may encourage perfectionism, which is strongly associated with eating disorders," they say. This means that, "An aspirational school culture may inadvertently lead to increased rates of eating disorder."


Eating disorders are fairly common among adolescent girls, and can take a terrible toll on health that lasts throughout life. They affect bone strength and fertility, and are hard to treat and recover from.

Researching factors that might affect the risk of getting an eating disorder is important, and this study is a helpful first step in looking at ways in which schools could reduce that risk.

But this study can only tell us so much. Researchers already know that girls are more prone to eating disorders than boys and eating disorders are more common among girls whose parents have a higher education level.

What this study adds is that these things might have a cultural effect on a whole school environment, beyond the effect on individual girls with highly educated parents.

The study does not tell us the mechanisms behind the increased risk they found. As the researchers note, it could be that parents with a higher education are more likely to spot and seek help if their child gets an eating disorder.

As the figures in the study include attendance at an eating disorder clinic, as well as actual diagnoses of eating disorders, this is important. It could be that parents at some schools are more aware of eating disorder clinics than others and more likely to make use of them.

It's tempting for the media to look for a scapegoat – in the Mail Online's case, "pushy parents" – to explain the findings. But the truth is that we just don't know.

It would be sad if schools where girls are encouraged to aspire to success were criticised for inadvertently causing eating disorders. Eating disorders are very complex, with many potential interacting causes. It's not helpful to pin the blame on parents or schools who are doing their best to help their children.

If you, or someone you know, may have an eating disorder, it's important to seek help quickly. Talk to your GP or get in touch with a charity like Beat, which supports people with eating disorders.

NHS Attribution