We obey two masters. There is what is right and what is normal and the two are often different. When this happens, the pull of the normal usually wins out. This is true whether by "right" we mean factual, as in the Asche conformity experiment:
or if we mean "right" in terms of moral, or even "right" in terms of healthy as a new study shows.
"Eighty-two percent of the obese women underestimated their weight, compared with 43% of overweight and 13% of normal-weight women. Likewise, 86% of overweight or obese children failed to correctly estimate their weight, compared with just 15% of normal-weight children...Moreover, when the children were presented with a series of cards bearing silhouette images of body types and were asked to select the "ideal" or "healthy" size for their mother, they tended to pick body types that were, in fact, unhealthily large."
What we see influences what we think we should see. Normal is the most powerful social force out there, but normal is sometimes harmful. If normal just runs roughshod over rational, moral, healthy, even empirically obvious, then appeals to these will be ineffective. How then do we change normal where it is dangerous or wrong and replace it with a new more helpful normal?