To present validity concepts in a conceptual framework useful for research in clinical settings.
We present a three‐level decision rubric for validating measurement instruments, to guide health services researchers step‐by‐step in gathering and evaluating validity evidence within their specific situation. We address , the capacity of an instrument to measure constructs it purports to measure and differentiate from other, unrelated constructs; , the reliability of the instrument; and , the ability to generalize scores from an instrument across subjects from the same or similar populations. We illustrate with specific examples, such as an approach to validating a measurement instrument for veterans when prior evidence of instrument validity for this population does not exist.
Validity should be viewed as a property of the interpretations and uses of scores from an instrument, not of the instrument itself: how scores are used and the consequences of this use are integral to validity. Our advice is to liken validation to building a court case, including discovering evidence, weighing the evidence, and recognizing when the evidence is weak and more evidence is needed.
