In the past, evaluations have not had good ways to measure impacts on overall career trajectories. Rather, they have relied on traditional point-in-time and period measures of education and employment to discern career progress. As a result, our understanding of whether and how programs foster alternative pathways is woefully incomplete.

This paper begins to fill the gap. Abt researchers used data on 2,544 young adults who participated in a randomized controlled trial of the Year Up program to analyze impacts on career trajectories. Year Up is a national training program that prior Abt analyses found produced large increases in period measures of earnings. The present paper’s authors use a technique called Sequence Analysis, which originated in DNA research, to identify distinctive types, or patterns, in the sequences of school and work activities that sample members followed over a three-year follow-up period.

Comparisons of individuals in the treatment (Year Up participants) and control (non-Year Up) group show that Year Up increased the likelihood that career sequences progressed from full-time training to higher-wage jobs and decreased the likelihood that sequences led to low-wage and part-time work. The program had little effect on the likelihood of sequences involving sustained school enrollment or persistent disconnection from school and work.

Among other notable findings, Year Up was equally likely to help young women and men access higher wage trajectories. This result is encouraging since information technology, one of Year Up’s main target sectors, is traditionally dominated by men. The implication is that well-designed workforce training can create viable pathways to tech careers for women as well as men.

The analysis also looks at relationship between sequence type and other characteristics to validate and better understand the sequence patterns identified. A series of initial characteristics (e.g., age, income, prior education, and psycho-social characteristics) predicted young adults’ subsequent sequence types in ways generally consistent with the basic research literature. The three-year sequence patterns also showed expected correlations with a set of outcomes measured six years after study intake.