To the Editor:
—We are indebted to Dr. Pfrommer for raising an issue which was perhaps not clearly presented in the original article.Like Dr. Pfrommer, we, too, were concerned about the validity of self-reported heights and weights. Therefore, we did not attempt to draw fine distinctions, for example, to analyze the data in terms of degrees of obesity, but merely distinguished between two main categories, "obese" and "normal." Thus the degree of discrepancy between reported heights and weights and actual heights and weights, which our analysis could tolerate and still remain valid, was considerable.In addition, we argued that the main effect of any remaining lack of validity would be to blur the distinction between our two categories (some of the respondents we classified as normal might actually be obese and some, probably a smaller number, of the people who reported themselves as obese might actually be of normal