Thursday, November 10
Data Quality and Measurement Error
Thu, Nov 10, 10:45 AM - 12:10 PM
Hibiscus A
Comparing Alternative Survey Measures

Issues in Disability Measurement: A Comparison of Three Methods (303595)

Andrew Houtenville, University of New Hampshire 
Kimberly Phillips, Institute on Disability 
*Vidya Sundar, University of New Hampshire 

Keywords: disability, survey design, measurement error

The complex and dynamic nature of disability makes it difficult to identify accurately and consistently among survey respondents. In the last eight years, a six-question sequence (6QS), first designed the American Community Survey (ACS), has been incorporated into many national surveys. Burkhauser et al. (2014) found that the 6QS, which includes questions related to hearing, vision, ambulatory, cognitive limitations and independent living and self-care difficulties, fails to capture about a quarter to a third of Social Security disability program recipients (presumably false negatives). To further investigate false negatives, we designed an expanded set of questions including the four ACS functional limitation questions, additional probing questions based on questions from the Canadian Survey of Disability and an open-ended disability question. The new set of disability questions were incorporated into and tested using the 2015 Kessler Foundation National Employment & Disability Survey (KFNEDS); a random digit dial survey of households in the United States. Additional testing of the new set of disability question was conducted using an online opt-in panel. We will present the results of this inquiry and comparison of the estimates of disability identified using the three methods.