The Effects of Survey Nonresponse and Proxy Response on Measures of Employment for Persons with Disabilities

Abstract

If many people do not respond to surveys, and those who do not respond are different from those who do, then survey estimates may be biased. This study examines potential bias in employment statistics for persons with disabilities arising from differences in survey response patterns between persons with and without disabilities. Several types of response rate are considered: contact, cooperation, and self-response (versus proxy response). Also, several types of disability are considered: mobility, mental, seeing, hearing, and MR/DD/LD. Data are from the National Health Interview Surveys of 1994 and 1995, including the National Health Interview Survey on Disability, Phases1 and 2. Based on evidence of this study, there is little reason to believe that household survey-based employment statistics for persons with disabilities are significantly biased by nonresponse or proxy response of respondents with disabilities.