Data sets make MRDRC workshop debuts

The MRDRC presented its 2021 workshop virtually on March 11 and 12. Historically, the MRDRC has had an interest in data development, dating from its founding association with the then-new Health and Retirement Survey (HRS). This year’s workshop continued that tradition with a number of presentations featuring newly developed or underused data.

The Occupational Requirements Survey (ORS), which released its first wave in 2019, is one such data set. ORS is a joint effort between the Bureau of Labor Statistics and the Social Security Administration (SSA), which uses it to help make disability determinations. ORS surveys private, state, and local government employers about the jobs available at their work sites, with an emphasis on an occupation’s physical requirements (e.g., strength, hearing, or stooping), specific vocational preparation requirements (educational, experience, licensing and certification, post-employment training), as well as mental and cognitive demands. Two workshop presentations featured the ORS: “The Role of Physical and Cognitive Job Demands on Transitions into Retirement” (UM21-05) and “Occupational Requirements and Worker Functional Abilities: A Close Look at Three Key Occupations” (UM21-04).

Italo Lopez Garcia (RAND) presented the first project on behalf of his team, which includes Kathleen J. Mullen and Jeffrey Wenger (both RAND). The researchers are constructing physical and cognitive job demand indices from similar measures in the ORS and the more frequently used Occupational Information Network (O*NET). O*NET has comprehensive job descriptions, but lacks a convenient way to use them in a measure of functional limitations. Once merged to HRS data, the new indices will allow comparative analyses of how physical and cognitive job demands affect retirement decisions. (Because the ORS cognition data is not yet available for 2020, the team has focused on physical measures thus far.) The researchers will assess the resulting job demand estimates. This should give an idea of which data set produces more useful results for disability determination.

The second ORS project, from University of New Hampshire researchers Andrew J. Houtenville, Megan Henly, and Debra Brucker, uses the Work Disability: Functional Assessment Battery (WD-FAB), another first for an MRDRC project. Henley shared that the researchers have fielded an exploratory WD-FAB asking workers to assess their abilities in four physical and four cognitive domains relevant to their current or most recent job. Next, the researchers will drill into three specific occupations by fielding another WD-FAB. With the data, they’ll look at the role of accommodations in mitigating workers’ functional levels and, for each occupation, compare the WD-FAB scores to parallel ORS domains. They hope that the project will demonstrate a more comprehensive approach that includes accommodations in disability determination.

For his MRDRC project, “The Effect of the Americans with Disabilities Act on Social Security Disability Insurance” (UM21-01), Timothy J. Moore (Purdue) is using state-level information on SSDI applications and allowances and state- and county-level data on SSDI beneficiaries from OASDI Beneficiaries by State and County. While much of the more recent data is available on SSA’s Statistics & Data Files site, the state- and county-level files only date back to 1999. For an RDRC grant through NBER, Moore is digitizing hard copy, county-level statistical publications from 1970 onward. Once that project is finished, Moore will make it freely available. (A special note: Moore is missing the 1981“Social Security Beneficiaries/OASDI Beneficiaries by State and County.” If you have a hard copy, please get in touch with MRDRC or Moore.)

In her talk, “Opportunities for Research Using Linked Survey and Administrative Data,” Jody Schimmel Hyde (Mathematica) made the case for using SSA’s Disability Analysis File (DAF), which was designed for research. The DAF combines DI and SSI records since 1996, offers extensive documentation, and features one row per beneficiary. The DAF contains monthly measures of benefits due/paid, zip code, work activity, use of vocational rehabilitation, reason for denial/cessation, and SSA Continuing Disability Reviews, but lacks demographic information typically found in surveys.

Linking the DAF with survey data avoids the time costs of creating administrative data links, while offering a view of applicants’ outcomes beyond earnings. Currently, DAF is linked to the HRS, but Schimmel Hyde pointed out that future linkages — to the Panel Study of Income Dynamics or Medical Expenditure Panel Survey, for example — would enable a broader range of research than administrative data or a single survey alone.

Other projects presented at the workshop dealt with actuarial fairness of spousal and survivor benefits and delayed claiming, COVID-19’s effects on retirees’ economic security, and disability claiming.

The workshop, which is sponsored through MRDRC’s cooperative agreement with SSA, also included an address from newly appointed Deputy Commissioner for Retirement and Disability Policy Kilolo Kijakazi. Kajakazi discussed SSA’s research goals as presented in the fiscal year ’22 Focal Area Memo. She emphasized that no matter what the focal area, there are two overarching questions this year: 1) How do disparities by race/ethnicity affect the research question, and 2) how has COVID-19 affected outcomes?

Finally, Kijakazi encouraged the MRDRC to redouble its efforts to work with minority serving institutions and historically black colleges and universities. She would like to see an increased diversity of voices and viewpoints in Retirement and Disability Research Consortium work.

Leave a Reply

Your email address will not be published. Required fields are marked *