Doesn't Get Better With Age: Predicting Millennials' Disconnection.

Recent research has highlighted the critical problem of high rates of "disconnected youth"—youth and young adults who are neither employed nor in school. Practitioners, scholars, policymakers, and a range of stakeholders express mounting concern that disconnected youth are a societal burden that costs countries trillions of dollars. This phenomenon could threaten social cohesion and social development, in addition to costing large amounts of money in the forms of foregone labor productivity and tax revenues and increased incarceration and social services uptake rates. Whereas other studies explore disconnection among a cross‐sectional sample of U.S. young adults, we set out to document the longitudinal education and employment pathways of a nationally representative cohort of U.S. millennial young adults from approximately age 16 to age 26. Our results highlight that both socioeconomic status (SES) and high school academic abilities are associated with disconnection from society at ages 18, 20, and 26. However, these associations diminish over time, and by age 26 SES is a weaker predictor of disconnection than having been on a college preparatory track (e.g., taking practice college entrance examinations) or having special educational needs in high school. At the same time, high school literacy and mathematics skills and grades are unrelated to disconnection at age 26. Prior disconnection remains one of the strongest predictors of disconnection at ages 20 and 26, highlighting the problem of repeated disconnection experienced by a small group of youth and young adults (only 0.6% of our full sample, or 11.1% of high school leavers). Finally, the odds of dropping out of the labor force—not seeking employment—are most strongly explained by gender and having dependent children, and to a lesser extent by race and high school region and urbanicity. Recommendations for supporting those most at risk of disconnection are discussed in relation to our findings. Report Number: ETS RR–18‐42 [ABSTRACT FROM AUTHOR]