This study is the first large (n=6,987), longitudinal (subjects followed for an average of 23 years) study of low-income persons to determine predictors and pathways to early adult violence. We are particularly interested in identifying those service systems which engage future perpetrators before perpetration begins, so that venues for the basing of preventative services can be identified. The study uses an integrated database, drawing from many administrative sources. From birth to age 18, our data include court records (restraining orders against subjects and caretakers), mental health services, special education, child welfare, emergency room visits and inpatient hospitalization (health/mental health), income maintenance, juvenile court and corrections, birth and death records, runaway shelter, Census data, and community demographics and crime rates. In adulthood, data include arrests, child welfare (as perpetrator), corrections, court (TRO/RO for perpetrators and victims), income maintenance, mental health, ER visits (health/mental health), birth, death and marriage records.