Eating disorders, such as anorexia, are traditionally thought to be driven by a distorted body image, so affected people see themselves as excessively overweight (and therefore unattractive) despite being very thin.
A recent study by psychologist Anita Jansen and colleagues has challenged this theory, by showing that women with eating disorders are actually more accurate at judging how attractive they are to others, whereas unaffected women typically over-estimate their attractiveness.
Jansen’s team asked women with and without eating disorder symptoms to have their picture taken, from the neck down, in their underwear. They were then asked to rate their own body for general attractiveness, and say which was the most attractive and unattractive part of their body.
These anonymised photos were then shown to two panels, consisting of both males and females, who were asked to make the same ratings.
The women with symptoms were generally in agreement with the panels, whereas those without rated themselves as more attractive and typically did not agree on which were their most and least attractive body parts.
This shows a lack of a ‘self serving attribution bias’ which is a normal tendency to over-attribute positive things to ourselves and negative things to other people or situations.
A recent review of the research suggested that this bias is usually strongly present in most people. It has been suggested that this may be useful, as it might emotionally cushion us from some of life’s hardships.
People with certain forms of mental illness, particularly depression, tend not to have this bias, however, meaning they actually view the world more accurately – an effect coined ‘depressive realism‘.
Jansen’s study suggests a similar ‘painful realism’ effect may be present in people with eating disorders, although it’s not clear whether this is specific to body perception, or whether it is primarily associated with emotional difficulties that often accompany conditions like anorexia.
UPDATE: World of Psychology has interesting commentary on this research (and post).
Link to abstract of study.
Link to eating disorder information from mental health charity Mind.
Mind Hacks 
Well, the problem with studies like this is that Americans have an insanely warped idea of what is “attractive” in the female form, so I don’t think any study like this is really very valid. I’m often told I’m attractive, even though according to “normal” American values I would be considered overweight. Most eating disorders stem from trying to meet a societal “norm” for attractiveness that is simply unnattainable for most women. It simply isn’t really natural to be a size 0 or 2 or 4, even though that’s what all the young girls want to be right now. No wonder eating disroders are rampant in this society.
It’s obvious that moderate pessimism is an adequate and healthy judge when comparing the self to others, but the anorexics percpetion is distoted beyond pessimism. Because how come the thin ones don’t “real”ize how grotesquely thin they are?