Can the Current Population Survey be used to identify the disabled?

In the September 1980 Monthly Labor Review, Barbara L. Wolfe compares the labor market experience of the disabled to that of the nondisabled, using data from the Current Population Survey (cps).' Because of methodological problems, I believe that CPS data are of limited usefulness in analyzing disability . The Wolfe study uses data from the March 1977 Cps to compare the labor force characteristics of the disabled and nondisabled. Because the CPS does not contain specific questions on health or disability status,' Wolfe employs a three-step approach to identify the disabled population . First, persons receiving income from at least one of a number of transfer programs were automaticaly included if they met certain program requirements that would identify them as disabled . These programs included social security disability, Supplemental Security Income, workers compensation, railroad disability benefits, and disabled veterans benefits . Second, persons whose work activity was limited during the year by reason of ill health or disability were included . Their responses to the household survey led to the following classifications: