Cesarean section rates vary among states from approximately 15% to over 26% of all deliveries. Since it is unlikely that patient factors alone contribute to this wide variation, other non-clinical factors that are unique to each state must influence cesarean section decisions. To explore if provider workforce and specialty was associated with differences in statewide cesarean rates, we compared statewide cesarean rates for 1996 with (1) the volume of deliveries in a state per board-certified obstetrician; (2) percentage of deliveries performed in the state by nurse midwives, and (3) the percentage of family physicians in the state performing obstetrics. In a linear regression model that adjusted for state rurality and median income, we found that only the percentage of family physicians participating in obstetrics was related to cesarean delivery rates. As the percentage of family physicians offering obstetric services increased in a state, the rate of cesarean delivery for that state declined. This effect appeared to be independent of other provider effects, state rurality, or statewide income. Family physician participation in obstetrics is unlikely to be the cause for lower cesarean rates, but is likely a marker for a medical environment and practice style that supports non-operative obstetric care.