MEASURING SUBJECTIVE OUTCOMES - RETHINKING RELIABILITY AND VALIDITY

Reliability and validity are criteria used to assess metric adequacy a nd are typically quantified by correlation coefficients. Reliability i s described as the extent to which repeated measurements yield consist ent results. Validity is described as the extent to which a measure ac tually measures what it purports to measure. These conceptualizations are less useful when applied to measures of subjective outcomes becaus e they do not convey other influences that ''drive'' correlation coeff icients. Consistency is a manifestation of a reliable instrument but d oes not ensure that an instrument is reliable. Establishing the validi ty of an instrument is a complex process that is heavily dependent on an investigator's hypothesis. Hence, validity coefficients may be more a reflection of hypothesis adequacy than of the extent to which instr uments measure what they purport to measure. Appreciating how coeffici ents are influenced will better enable clinicians to assess the adequa cy of subjective outcome measures.