Measuring health: improving the validity of health assessments

Abstract

A large number of test batteries, instruments, indices and scales promise to provide reliable and valid information about an individual's health status. It is often neglected, however, that the data gathered using these measures are based entirely on self-reports and are susceptible to a variety of biasing factors. This is of particular concern for health research because many questions are inherently ambiguous and require considerable interpretive work on the part of the respondent. A single term may cover a wide range of health-related symptoms, experiences, or events, making health-related questions highly susceptible to contextual influences. This article discusses a number of possible influences on individual responses and suggests several determinants of self-report validity. The paper is organized into two major parts. The first part demonstrates that the validity of survey responses can be enhanced substantially when knowledge about conversational dynamics is taken into account during the design stage of the questionnaire. The second part argues that respondents are motivated and able to provide valid information about their overall health status. The validity of global single-item measures of health status is examined, and a cognitive model is introduced that outlines how subjective health evaluations are made.