Presented by:

Nancy Mathiowetz & Roger Tourangeau

University of Maryland

Thursday, March 16, 2000

Abstract

This is the second in a series of talks about the projects to be carried out by the Michigan Center for Excellence in Health Statistics.  This one concerns the measurement of disability in surveys.  The measurement of disability is likely to grow in importance as the population ages; already several surveys including the American Community Survey routinely include questions on disability.  Disability is a multi-dimensional concept, and the results from a number of disability surveys suggest that it is quite difficult to capture with the survey items that have been used in the past.  For example, the past results indicate that different rates of disability are found depending on whether self- or proxy reporters provide the data; in addition, the classification of a person disabled can fluctuate widely over time and may be subject to context effects.  The talk will present our plans to examine these conceptual and measurement problems in more detail.  We plan to explore self-proxy differences, the reliability of both types of reporters over short periods of time, and the impact of lay conceptions of disability.  We also plan to try out several approaches to improve reporting of disability, including simplifying the questions, decomposing the complex judgment into several simpler ones, and using graded scales rather than dichotomous answer categories.

See presentation