In recent years, the U.S. Congress has passed legislation to support individuals with disabilities pursuing self-employment.
The American Rescue Plan Act of 2021, for example, included a provision intended to improve access to grants and loans for that made it easier for small-business entrepreneurs with disabilities.
However, despite these efforts, large gaps in self-employment rates still exist among individuals with disabilities – particularly when accounting for different social factors, such as race/ethnicity, gender, sexual orientation and geographical locale.
Assistant Professor Gemarco Peterson conducted a study with colleagues at Tarleton State University and Langston University to better understand those gaps and examine how machine learning may be used to identify patterns associated with self-employment among individuals with disabilities.
Their study, published in the Journal of Rehabilitation, analyzed data from the National Survey on Health and Disability, an online survey for adults with disabilities. The analytic sample included 762 adults with disabilities who completed the survey between 2020-2023.
Within the analyzed sample, the researchers found that African Americans/Black people with disabilities had lowest rates of self-employment (10.87 percent) compared to individuals from other racial/ethnic backgrounds, including Latinx (14.68 percent), White (12.18 percent) and Asian/Native Hawaiian/Other Pacific Islander (12.04 percent).
Men and women exhibited nearly identical rates of self-employment (12.28 percent and 12.23 percent, respectively), and individuals who had experienced poverty or who identified as a sexual minority were much more likely to be self-employed compared to their counterparts.
“Our findings set a precedent for understanding self-employment among people with disabilities through an intersectional lens and provide valuable insights into the demographic factors shaping self-employment outcomes among multiply marginalized people with disabilities,” they wrote.
In this study, “multiply marginalized” refers to individuals who identify as having a disability and experience overlapping disadvantages, such as racial minority status, poverty or sexual minority status.
Peterson and his team also used three machine learning models – Random Forest, Logistic Regression and XGBoost – to explore nonlinear and intersectional patterns associated with self-employment status. All three showed “modest predictive power,” with Random Forest performing best of the three.
“The combined machine learning methodological approach helped to identify specific intersectional patterns, particularly where race and poverty converged to lower the odds of self-employment,” the researchers wrote.
Peterson and his colleagues took the information from this study to make recommendations for policies and interventions that can be used to improve self-employment outcomes for individuals with disabilities.
For example, drawing on the recognized role of Historically Black Colleges and Universities (HBCUs) as trusted institutions with deep community ties, these institutions are positioned within rural and urban communities to collaborate with state and federal vocational rehabilitation (VR) agencies. Such partnerships could strengthen the development of small business incubators that African American/Black people with disabilities pursuing self-employment.
The U.S. Department of Treasury could increase funding for the Community Development Financial Institutions Fund, to expand access to capital for Asian/Native Hawaiian/Other Pacific Islanders with disabilities.
These and other efforts could give individuals with disabilities a better shot at developing their own businesses and reaching their employment goals.
“The insights generated through this study have the potential to enable more precise and effective targeting of resources, including VR services, career counseling and self-employment supports, for individuals from identified self-employment disparity populations,” Peterson and his team wrote. “These key findings highlight the urgent need to establish new policy, service and research initiatives aimed at reducing self-employment and small business startup challenges faced by multiply marginalized entrepreneurs with disabilities.”
About the Researcher
Gemarco Peterson
Department of Counseling and Psychological Services
Gemarco J. Peterson is an assistant professor of clinical rehabilitation counseling and the practicum and internship coordinator in the Department of Counseling and Psychological Services. He is a certified rehabilitation counselor, a nationally certified counselor and is trained as a vocational forensic evaluator. Peterson’s research focuses on career development, employment outcomes for people with disabilities and the impact of social identities on workforce development in counselor education, particularly the recruitment and retention of Black counselor educators. He also serves on journal editorial boards and national committees dedicated to advancing policies and training standards for clinical rehabilitation counseling and counselor education and supervision.
Citation
Peterson, G. J., Webb, K. R., Moore, C. L., Wang, P., Manyibe, E. O., Dentleegrand, K. R., and Webster, F. (2025). “A National Analysis of Self‑Employment Among People with Disabilities Across Demographic and Geographic Factors.” Journal of Rehabilitation, 91 (Special Edition Part 2), 28-40.