The 130th Annual Meeting of APHA

Abstract #40486

Gerry E. Hendershot, PhD, Hendershot Consulting, 4437 Wells Parkway, University Park, MD 20782, 301-927-1120, ghendershot@earthlink.net

Many of the available employment statistics on persons with disabilities are collected in interview sample surveys. Typically those surveys allow interviewers to conduct interviews with a proxy respondent if the sample person has a disability which, in the opinion of the interviewer, makes an interview difficult or impossible. It has been shown that in substantial numbers of cases, proxy respondents are substituted for disabled sample persons, especially if the disability is serious. It has also been shown that proxy respondents tend to underreport the extent and severity of disability among the sample persons for whom they report. These patterns suggest the possibility of substantial bias in survey reports of the employment characteristics of persons with disabilities. This study examines that potential bias using data from the 1994-1997 National Health Interview Survey on Disability, in which interviews were conducted with a large sample of persons with disabilities. The study examines the frequency of proxy reporting, the reasons for using a proxy, the relationship of proxies to sample persons, and the difference between proxy- and self-reports of employment experience.

Learning Objectives: After this session, participants will be able to

• Explain the major types of respondent rules in interview surveys
• Evaluate the potential biases in statistics based on survey responses by proxy respondents
• Describe the characteristics of proxy respondents in one large national disability survey

Keywords: Disability, System Involvement

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.