Journal of chronic diseases
-
Established indices of relative body weight are reviewed and critically evaluated in terms of five criteria: (1) high correlation with obesity, (2) a conceptual interpretation, (3) universality, (4) simplicity of computation, and (5) zero correlation with height. Regression procedures are used to create standard weight charts which are distributed similarly to the body weight of 13,645 Americans, provided by the National Health Survey (NHS), and which have an arithmetic mean equal to the mean of the Metropolitan Life Insurance ( MLI ) table of desirable weights. A similar method is used to create a standard weight table based on body mass index (BMI). Finally, standard body weight charts which are uncorrelated with sex, age, and/or body frame are developed and discussed in terms of their usefulness.