Eight observers, each trained rigorously in the same anthropometric techniques and at the same time, measured 8 males and 8 females in a controlled laboratory setting to assess interobserver replicability for 63 anthropometric dimensions. These included all but 4 variables in the “basic” and “full” IBP recommended lists, in addition to 25 other measurements. The results indicate that observer interchangeability for these dimensions, even under optimal conditions, is poor for both sexes but especially for female subjects. Investigators who use raw anthropometric data from multiple sources for comparative purposes should do so only with considerable caution, and should be aware of those dimensions that are least reliable.
Bennett, Kenneth A. and Osborne, Richard H.
"Interobserver Measurement Reliability in Anthropometry,"
5, Article 8.
Available at: https://digitalcommons.wayne.edu/humbiol/vol58/iss5/8