Difference scientific fields will value different things. A computer scientist might prefer a fundamentally more "valid" metric, but someone who studies somethi

Difference scientific fields will value different things. A computer scientist might prefer a fundamentally more "valid" metric, but someone who studies something more organic would rather have a less valid but more empirically grounded one. You can't forget that there's really a ton of reported data behind BMI. So much that e.g. in children we know, per age, the percentiles which should be used as cut offs for considering obesity and triggering further analysis/intervention. But this is the individual level! As you certainly know, obesity is largely a public health issue, which makes BMI - which has lots of epidemiological data behind it - even more appropriate. There's really no better way, IMHO (and not only [1]), to track obesity as a public health problem. [1] Hall, D. M. B., & Cole, T. J. (2006). What use is the BMI? Archives of Disease in Childhood, 91(4), 283–286. http://doi.org/10.1136/adc.2005.077339 http://doi.org/10.1136/adc.2005.077339