Research Brief: National and Administrative Datasets Available to Explore Rehabilitation Related Research on Disability and Employer Practices
A number of large national and administrative datasets collect rich information on a range of topics that can inform specific rehabilitation research and policy questions. However, for those who desire to use these files, it has traditionally been an arduous task to learn their finer details. A new online catalog compiled by the Cornell University RRTC team enables researchers to more readily understand the variables contained in these files and to assess which subset of files contain similar measures for cross-data comparison studies.
An online tool was developed by Cornell University to facilitate the exploration of eleven national survey and administrative datasets. The online catalog, known as the Cross-Data Catalog of Disability and Compensation Variables,[1] permits the comparison of hundreds of variables across datasets, including: demographic details, disability and health conditions, labor force information (e.g., employment and compensation), and employer characteristics. Its availability allows rehabilitation researchers, policymakers, educators, and students to rapidly compare and contrast key attributes and variable content. In this manner, it is possible to screen multiple individual data sources to determine which is optimal for the research under study. These datasets are briefly detailed below.
Begin Table
This table depicts the total population, sample size, and record history of 12 major national survey and administrative data sets.
|
Dataset |
Target Population |
Sample Size |
Record History |
|
Decennial Census 2000 |
All persons |
10,300,000 |
2000 1790-current[2] (Every 10 years) |
|
American Community Survey (ACS) |
All persons |
3,000,000 |
2000-current (Annual) |
|
Survey of Income and Program Participation (SIPP) |
Non-Institutionalized |
95,000 |
1984-current (Quarterly) |
|
Current Population Survey (CPS) |
Non-Institutional |
215,000 |
1940-current (Annual) |
|
Panel Study of Income Dynamics (PSID) |
Non-Institutionalized |
21,000 |
1968-current (Biennial) |
|
National Health Interview Surveya (NHIS) |
Non-Institutionalized |
100,000 |
1957-current (Annual) |
|
National Health Interview Survey on Disabilitya (NHIS-D) |
Non-Institutionalized |
100,000 |
1994/1995 (Once) |
|
University of Michigan Health and Retirement Study (HRS) |
Ages 51 and older |
21,000 -32,000 |
1992-current (Biennial) |
|
Behavioral Risk Factor Surveillance System (BRFSS) |
Non-Institutionalized |
350,000 |
1984-current (Annual) |
|
Medical Expenditure Panel Survey (MEPS) |
Non-Institutionalized |
22,000 |
1996-current (Annual) |
|
Rehabilitation Services Administration Records (RSA-911) |
VR clients (closed cases) |
Over 600,000 |
1991-current (Annual) |
|
National Longitudinal Transition Study-2 (NLTS-2) |
Youth 13-16 in 2000 |
11,000 |
2000-2010 (Annual) |
Note: The National Health Interview Survey (NHIS) and National Health Interview Survey on Disability (NHIS-D) are presented as one dataset in the Cross-Data Catalog of Compensation and Disability Variables.
End table
Every ten years, the Decennial Census is conducted by the United States to gather data at the national, state, and local levels. The census of the population informs the government regarding the number of seats assigned to each state in the U.S. House of Representatives and is also used to allocate federal funds to local communities. The Decennial Census 2000 was comprised of a “short form” questionnaire that was given to all households, and a “long form” that was given to one out of every six households. While the short form collected basic demographic details, the long form inquired about a variety of factors, including information about housing, economic status, and disability. Unfortunately, the wording and design of the questionnaire resulted in potential misinterpretations of two of the six disability items, rendering them unfit for analyses.[3] In 2010, the fielding of the Decennial Census was restricted to the content of the short form, with the American Community Survey replacing the long form.
The U.S. Census Bureau annually fields the American Community Survey (ACS), which serves to inform federal, state, and local governments about various aspects of the population. Questionnaires are submitted to in-sample households and group quarters (institutionalized and non-institutionalized). Variables on the survey elicit detailed demographic, social, economic, and housing information from 3.54 million households each year, 2.5 percent of those in group quarters, and 36,000 in Puerto Rico. Since 2008 the ACS has contained a static set of six disability questions[4] that have been adopted by a number of other population-based surveys. The size of the ACS in combination with its rich detail enable in-depth studies on topics including employment, earnings, household income, poverty, SSI receipt, and health insurance coverage for persons with and without disabilities.
The Survey of Income and Program Participation (SIPP) is a panel dataset that includes both core wave and topical module files. Every four months, the wave-level questionnaire is fielded, obtaining data about changes in the set of core survey variables, including labor force participation, household composition, work-limiting conditions, health insurance, and enrollment in government programs. Topical modules probe deeply into specific areas of interest, including histories of employment, work disability, functional limitations, and education/training, among others. In 2008, the longitudinal survey followed roughly 100,000 non-institutionalized people in 52,000 households every four months for a total of 13 interviews over the course of four years. The SIPP is being redesigned so that it will be annually fielded.
The U.S. Census Bureau fields the monthly Current Population Survey (CPS) to approximately 60,000 households on behalf of the Bureau of Labor Statistics. Households are surveyed for four months and then are out of the survey rotation for eight months before being interviewed again for four consecutive months. The head of each household supplies details regarding household members, with a focus on labor force participation. Supplemental questionnaires provide rich data on various social, economic, and health topics. Some of these supplements occur periodically (e.g., displaced workers, migration, and tobacco use), while others are annually incorporated into the survey rotation.
The Annual Social and Economic Supplement (ASEC) to the CPS, also referred to as the March Supplement, inquires about the social and economic status of each household member. It includes a work activity limitation question[5] that has permitted researchers to study the well-being of people with disabilities since 1981. Beginning in June 2008, the CPS adopted a set of six disability items that is now included on its basic monthly files. In May 2012, a disability supplement was conducted to explore areas related to the employment of people with disabilities.
The Panel Study of Income Dynamics (PSID) provides more than four decades of data on the economic, health, and social behaviors of respondents. Beginning in 1968, the PSID was annually fielded as two separate samples of a total of 4,800 families, the University of Michigan Survey Research Center (SRC) and Survey of Economic Opportunities (SEO) samples. Since 1997, families have been interviewed once every two years, representing a sample of more than 18,000 people in 5,000 family units.
Initially, a work activity limitation question was asked only of the head of the family unit, but this was changed in 1981 to additionally include the spouse of the head of the family unit. The PSID also expanded its disability items in 2003 to include a series of questions about Activities of Daily Living (ADLs) and Instrumental Activities of Daily Living (IADLs), along with questions about mental disabilities. The Disability and Use of Time supplement to the 2009 core file provides information from about 400 couples in which both spouses were at least 50 years old as of December 31, 2008.[6]
The U.S. Census Bureau fields the National Health Interview Survey (NHIS) for the National Center for Health Statistics (NCHS). This annual cross-sectional dataset dates from 1957 and includes responses from 87,500 non-institutionalized people residing in 35,000 households. The core questionnaires and supplements provide rich data regarding the socio-demographic, health, and disability characteristics of respondents. The four components of the core questionnaire were last redesigned in 1997. Because of its focus on health and well-being, it includes an extensive set of indicators that aim to identify work activity limitations, ADLs and IADLs, the level of mental capacity, as well as the presence of any physical and sensory disabilities. From 1994 to 1995, a special disability supplement was conducted (NHIS-D) that provided researchers with valuable insights about issues related to disability.
Every two years since 1992, the Health and Retirement Study (HRS) has been collecting longitudinal data on Americans ages 50 and older regarding disability, retirement, labor force participation, savings, and the usage of health insurance and health care services. Run by the University of Michigan, this study is supported by the National Institute on Aging and the Social Security Administration. It contains data on 26,000 respondents and their spouses and contains extensive information about the evolution of disability, health conditions, and functional impairments.
State health departments conduct telephone interviews of the Behavioral Risk Factor Surveillance System (BRFSS), an annual cross-sectional survey designed by the Centers for Disease Control and Prevention (CDC). The BRFSS is comprised of a core survey as well as 26 optional modules relating to specific health topics that states can choose to include. Although its measures of disability are coarse, it is able to provide state and limited county level information about risk behaviors related to chronic diseases, injuries, and preventable infectious diseases, as well as about health practices.
Since 1996, the Agency for Healthcare Research and Quality (AHRQ) has conducted the Medical Expenditure Panel Survey (MEPS) to collect annual data on individuals, their medical providers, and employers. It contains a rich set of measures about disability, health care expenditure and service utilization, and health insurance coverage and availability. Starting with the 1996 MEPS, each file can be linked to the previous year’s NHIS core person file, a merge that expands the universe of possible analyses using these two datasets.
The Rehabilitation Services Administration - 911 (RSA-911) contains federal Rehabilitation Services Administration (RSA) data on all cases closed by each state Vocational Rehabilitation (VR) agency during the fiscal year. Its variables include measures of primary and secondary disability, the causes and severity of these conditions, as well as VR service utilization and employment outcomes.
The U.S. Department of Education funded the National Longitudinal Transition Study – 2 (NLTS2) to annually survey a sample of about 12,000 students ages 13 to 16 years old enrolled in special education in 2000.[7] This restricted access dataset contains a variety of information collected from students, their parents or guardians, teachers, principals, and school records about each student’s program participation, services received, class content, academic performance, extracurricular activities, postsecondary education, employment, independent living, and community participation. The study concluded in 2009 when the original sample members reached ages 21 to 25.
Begin Bar Graph
Title: 2011 Average U.S. Weekly Earnings of Competitively Employed Individuals Receiving SSI/SSDI Benefits
This chart depicts the average U.S. weekly earnings of competitively employed individuals receiving SSI/SSDI benefits, showing data collected from both the American Community Survey and the Rehabilitation Services Administration’s RSA-911, all for the year 2011.
ACS, average U.S. weekly earnings: $470.30
RSA-911, average U.S. weekly earnings: $242.90
End Bar Graph
Begin Bar Graph
Title: 2011 Average U.S. Weekly Working Hours of Competitively Employed Individuals Receiving SSI/SSDI Benefits
This chart depicts the average U.S. weekly working hours of competitively employed individuals receiving SSI/SSDI benefits, showing data collected from both the American Community Survey and the Rehabilitation Services Administration’s RSA-911, all for the year 2011.
ACS, average weekly working hours: 29.5
RSA-911, average weekly working hours: 23.9
End Bar Graph
The Cross-Data Catalog of Disability and Compensation Variables should enhance cross-data analyses of major national survey and administrative datasets as researchers become enlightened as to the common variable content contained in these eleven files. For example, this online catalog reveals a series of measures available in the 2011 ACS and RSA-911 files that permit the comparison of outcomes of working recipients of Supplemental Security Income (SSI) or Social Security Disability Insurance (SSDI) benefits.
The individuals who established competitive employment subsequent to receiving VR services on average earn just over half of the weekly earnings of the general working SSI/SSDI population observed in the ACS ($242.90 compared with $470.30 each week), but they work only 5.6 fewer hours each week on average (29.5 hours/week compared with 23.9 hours/week). This implies that for each hour of work, individuals from the RSA-911 file who receive SSI/SSDI benefits earn 36.3 percent less per hour than the comparable population in the ACS ($10.16/hour compared with $15.94/hour). This disparity may be due to the nature of self-reported data in the ACS if sample respondents over-reported earnings and working hours. It may also reflect the greater degree of disability of individuals who receive VR services in the RSA-911 files.
The online Cross-Data Catalog of Disability and Compensation Variables of national survey and administrative datasets simplifies the process of determining whether a given file can appropriately and adequately inform a particular area of study. With this knowledge more readily available, rehabilitation researchers, policymakers, educators, and students will be able to more efficiently evaluate the role of demographic and workplace characteristics that might influence employment outcomes for people with disabilities. Using innovative methods to assess disability and rehabilitation employment program effectiveness can serve to enhance outcomes for people with disabilities at the state and national levels
Brault, W. (2009). Review of Changes to the Measurement of Disability in the2008 American Community Survey. U.S. Census Bureau
Erickson, William A. & Houtenville, Andrew J. (2005, November). A Guide to Disability Statistics from the 2000 Decennial Census. Rehabilitation Research and Training Center on Disability Demographics and Statistics, Cornell University, Ithaca, NY. http://digitalcommons.ilr.cornell.edu/edicollect/187/
Stern, S. (2003). Counting People with Disabilities: How Survey Methodology Influences
Stern, S., & Brault, M. (2005). Disability Data from the American Community Survey: A Brief
This brief summarizes the research article “Rehabilitation Related Research on Disability and Employer Practices Using Individual-Based National and Administrative Datasets” by Zafar Nazarov, William A. Erickson, and Susanne M. Bruyère, 2013, Cornell University Working Paper.
This research was funded in part by the United States Department of Education through a grant from the National Institute on Disability and Rehabilitation Research (NIDRR grant No. H133B100017).
The contents do not necessarily represent the policy of the Department of Education or any other federal agency, and one should not assume endorsement by the federal government (Edgar, 77.620(B)).
[1] The link for Cross-Data Catalog of Disability and Compensation Variables is http://www.disabilitystatistics.org/eprrtc/codebook.cfm
[2] The phrasing of disability questions has varied over the years since they were first included in the 1830 Census. Note that the 2010 Census did not include measures of disability.
[3] See Stern (2003), Stern and Brault (2005), and Erickson and Houtenville (2005).
[4] Prior to 2008, the disability items were redesigned every few years (Brault, 2009).
[5] The work-activity limitation question in the CPS-ASEC is: “[d]oes anyone in this household have a health problem or disability which prevents them from working or which limits the kind or amount of work they can do? [If so,] who is that? (Anyone else?)”
[6] This supplement oversamples couples with at least one member with a health limitation.
[7] The original NLTS covers 1985 to 1993 and was used as a basis for the NLTS2; therefore, the two cohorts under study can be compared.