The concept of validity as it applies to measures of health and health status is examined in the context of a set of standard, widely accepted definitions of validity. Criterion validity is shown to be irrelevant to health status measures because of the lack of a single specific, directly observable measure of health for use as a criterion. To overcome this problem, the Index of Well-being has been constructed to fulfill the definition of content validity by including all levels of function and symptom/problem complexes, a clearly defined relation to the death state, and consumer ratings of the relative desirability of the function levels. Data from a two-wave household interview survey provide convergent evidence of construct validity by demonstrating an expected positive correlation of the Index of Well-being with self-rated well-being and expected negative correlations with age, number of chronic medical conditions, number of reported symptoms or problems, number of physician contacts, and dysfunctional status. Discriminant evidence of construct validity is demonstrated by predicted differences in correlation between concurrent Index of Well-being scores and self-assessed overall health status, and between the Index of Well-being scores and self-rated well-being on different days. A simple method of estimating a currently usable comprehensive population index of health status, the Weighted Life Expectancy, is described.