OBJECTIVE To determine if the association between race and preterm delivery would persist when preterm delivery was partitioned into two etiologic pathways. METHODS We evaluated perinatal and obstetrical data from the 1988 National Maternal and Infant Health Survey and classified preterm delivery as spontaneous or medically indicated. Discrete proportional hazard models were fit to assess the risk of preterm delivery for Black women compared with White women adjusting for potential demographic and behavioral confounding variables. RESULTS Preterm delivery occurred among 17.4% of Black births and 6.7% of White births with a Black versus White unadjusted hazard ratio (HR) of 2.8 (95% CI = 2.4-3.3). The adjusted HR for a medically indicated preterm delivery showed no racial difference in risk (HR = 1.0, 95% CI = 0.4-2.6). However, for spontaneous preterm delivery between 20 and 28 weeks gestation, the Black versus White adjusted hazard ratio (HR) was 4.9 (95% CI = 3.4-7.1). CONCLUSIONS Although we found an increased unadjusted HR for preterm delivery among Black women compared with White women, the nearly fivefold increase in adjusted HR for the extremely preterm births and the absence of a difference for medically indicated preterm delivery was unexpected. Given the differences in the risks of preterm birth between Black and White women, we recommend to continue examining risk factors for preterm delivery after separating spontaneous from medically indicated preterm birth and subdividing preterm delivery by gestational age to shed light on the reasons for the racial disparity.