Full Citation
Title: Reliability and Stability of the Six Question Disability Measure in the Current Population Survey: What the Data Can and Cannot Tell Us About Disability and Labor Force Participation
Citation Type: Dissertation/Thesis
Publication Year: 2017
ISBN:
ISSN:
DOI:
NSFID:
PMCID:
PMID:
Abstract: The Current Population Survey (CPS) has been a major source of disability data for public policy and disability research for more than 30 years. Use of this same data, however, has been a source of criticism in forensic vocational rehabilitation settings when making claims about persons with disabilities and the nature of labor force participation. The aim of this study was two-fold. First, the study examined the six disability measures added to the CPS in 2008 to determine if they are both a reliable and stable method of describing disability over a period of two survey administrations in a 12-month period. Second, this study then assessed the impact of disability upon labor force participation. The findings demonstrate that the measures are effective, stable, and predictive. This research used a subset of the respondents to the longitudinal CPS Annual Social and Economic Supplement (more commonly known as the March Supplement); it included (N=11,721) respondents who indicated a positive answer to the disability questions in both survey months that the disability variables were measured. Descriptive analysis of expected demographic variable distributions supported the construct reliability of the measures, as well as provided some surprising results regarding higher-than-expected levels of income and wages among some persons with disabilities. Correlation analysis utilizing Kappa coefficients demonstrated that all six measures of types of disability in the CPS are stable across time, and Fisher Z transformations show that, among the six, measures of physical and mobility difficulties were the most stable. Measures of visual difficulties, while stable, are significantly less stable than the other disability measures. Logistic regression analysis indicated that all six disability measures have a significant predictive effect on the likelihood of employment of persons with disabilities, and a fully-controlled model including contextual variables (demographic characteristics) supported the conclusion that four of the six types of disability (physical disability and difficulties with remembering, mobility and vision) have independent statistically significant effects on employment. This study addresses some key criticisms of previous aggregate disability studies that relied on cross-sectional data, such as the widely-accepted criticism that cross-sectional studies over report the instance of long-term disability by capturing short-term impairment as well within a single survey administration. The findings reported of this research also contribute to the understanding of the statistical value of the aggregate measurement of disability and its potential usefulness to the field of forensic vocational rehabilitation.
User Submitted?: No
Authors: Joy, Jeffrey, D
Institution: Virginia Commonwealth University
Department: Rehabilitation Counseling
Advisor: Amy J. Armstrong
Degree: Doctor of Philosophy
Publisher Location: Richmond, VA
Pages: 190
Data Collections: IPUMS CPS
Topics: Health, Labor Force and Occupational Structure, Population Data Science
Countries: