Connecticut's Jobs First program was subjected to a rigorous, large-scale evaluation. During the evaluation, nearly 5,000 single-parent welfare applicants and recipients in 2 cities were randomly assigned to Jobs First or to the Aid to Families with Dependent Children (AFDC) group. Each group's progress was compared over a 4-year period. The following were among the key findings: (1) at the study's end, 51% of the Jobs First enrollees were working and off welfare versus 42% of the AFDC group; (2) Jobs First enrollees earned 7% more than did their AFDC counterparts; (3) although families in both groups received about the same amount in welfare payments at the end of the 4-year period, the Jobs First families averaged 6% more in income from public assistance and earnings; (4) just over half of the Jobs First recipients reached their benefits time limit, and about two-thirds received at least one 6-month extension; and (5) Jobs First generated some small improvements in the behavior of participants' young children but had mixed effects on adolescents. It was concluded that, at least when the economy is exceptionally strong and most nonworking recipients are granted benefits extensions, time limits on welfare receipt can be implemented without widespread severe consequences for families. (Contains 17 notes, 8 tables, and 3 figures.) (MN)