GLOSSARY

Reference Method is the method selected as a comparison standard for a specified quantity. A reference can still have error and may be appropriate for one compartment without being the best standard for every other compartment.

Validity concerns whether a method estimates the intended quantity accurately, while reliability concerns consistency on repetition. A method can repeat well while remaining biased. Technical error and prediction error also describe different sources of uncertainty. Small changes should be interpreted relative to relevant measurement error, with the same protocol used over time. Agreement at group level does not guarantee close agreement for every individual athlete, and switching instruments can create an apparent change that is methodological rather than biological.

Body-composition models divide the body into components according to a defined level. A two-component model often separates fat mass from fat-free mass. Adding measured water or mineral information can reduce some assumptions in more complex models, but no model removes all error. Molecular, cellular and tissue descriptions are different ways of organizing the same body. Before comparing results, identify the model, measured variables and assumptions rather than treating all reported components as directly measured anatomical tissues.