Ah, yes, the massaging of statistics is a well-known issue. Very few people even understand what medications mean when they advertise a certain percentage reduction in something. And then whether that something really affects disease in a meaningful way (say, cholesterol levels and heart attacks or stroke).
And I absolutely agree that eating disorders require a mental component. Not everyone exposed to X, Y or Z will develop an eating disorder, but those who are susceptible would be more likely to. Likewise, abuse alone is not enough to trigger schizophrenia in the average individual, but with certain genetic components (which we still are struggling to define), it increases the risk.
That said, abuse isn't good in ANY scenario, whether it triggers schizophrenia or not. And neither is photo-shopping to the extremes we see today. Even if it cannot be proven to contribute to anorexia or bulimia, it is not good for anyone to be constantly exposed to unrealistic images of what a certain group of people should look like. At the very least, it promotes poor self-esteem by people never being able to achieve certain body proportions, airbrushed skin, vivid eye color, etc.
So my question is, does it matter? It seems rather pedantic to be arguing about whether photoshopping directly causes eating disorders when the overall harm it does to people in general is fairly well-understood.