Application of machine learning to predict employment attainment among individuals with intellectual and developmental disabilities.
Sharpen daily-living skills and loop families into job support—machine learning shows these two factors drive employment for adults with IDD.
01Research in Context
What this study did
Eun and colleagues fed years of adult data into a Random Forest computer model. The goal was to see which factors best predict who with intellectual or developmental disabilities lands a job.
The model ranked dozens of variables. It kept the ones that cleanly split people into "employed" or "not employed" groups.
What they found
Daily living skills came out on top. Family employment support, age, and general work ability followed close behind.
The model used these five variables to guess employment status with high accuracy.
How this fits with other research
Clarke et al. (2025) extends this picture. Their twenty-year follow-up shows the same daily-living skills predict adult jobs, starting from childhood.
Su et al. (2008) flips the view. In their older regression study, having a job predicted better daily functioning, not the other way around. The two papers look at opposite directions in time, so they fit together without conflict.
Corrigan et al. (1998) and Allan et al. (1994) echo the theme. Daily-living skills and family factors forecast where young adults live after school, mirroring Eun’s employment findings.
Why it matters
You now have a short checklist for transition plans: showering, cooking, money handling, plus family help with job contacts. Train and track these skills first. They carry the most weight in real-world hiring algorithms and in life.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pick one daily-living skill the client lacks, set a 1-week mastery target, and ask a family member to rehearse it in a work-like setting.
02At a glance
03Original abstract
Promoting desirable employment outcomes for individuals with intellectual and developmental disabilities has been an important task for decades. However, the statistics indicate inequitable employment outcomes still exist; including underrepresentation in the workforce and employment in a part-time, low-wage, and segregated setting. One way to address the gap is to review and promote individual and environmental characteristics that are related to enhanced employment outcomes. For this study, we used machine learning approaches to investigate the predictors of employment status in individuals with intellectual and developmental disabilities based on a national database in South Korea. All machine learning models employed in this study-specifically a Random Forest-accurately and consistently predicted employment outcomes for individuals with intellectual and developmental disabilities. The most important factors contributing to the model's predictive accuracy include employment capability, family support for employment, age, overall work ability, and daily living skills. Implications for practice and research are also discussed.
Research in developmental disabilities, 2026 · doi:10.1016/j.ridd.2025.105181