- In a validation study, we compare estimates of physical activity made by a questionnaire with those measured by a more accurate method, ideally a gold standard.

- Epidemiologic studies are most interested in assessing typical or long-term physical activity due to its potential association with disease.

- However, there are no perfect measures of typical or long-term physical activity. All measures have error, although they differ in magnitude and type.

- Given that neither the questionnaire nor the comparison measure will be perfect, the error of both should be as uncorrelated as possible to avoid falsely high estimates of validity.