OBJECTIVE: To test the accuracy of various physicians' participation in Medicaid models. DATA SOURCES/STUDY SETTING: Primary data on 221 obstetricians and gynecologists in the Chicago area by telephone interviews over a four-month period. These data were combined with secondary data from the American Medical Association Master File (1993) and U.S. Census data (1990). STUDY DESIGN: Telephone interviewers posing as the older sisters of a pregnant teenager who is a Medicaid recipient sought information regarding the care provided in a first prenatal care appointment (e.g., appointment duration, tests administered, delivery privileges, appointment availability). DATA COLLECTION/EXTRACTION METHODS: A "receptionist helpfulness" variable was developed through pretesting on obstetricians in another city. Inter-interviewer reliability was enhanced through common interview technique education. PRINCIPAL FINDINGS: Only 81 obstetricians (36.7 percent) accepted new Medicaid patients. This finding is lower than previous research on physician participation in Medicaid. There was strong empirical support for both dimensions-cost containment and limited access-of the physicians' receptiveness model, the model introduced with this research. There was limited support for the dual market and residential segregation models of physician participation in Medicaid. CONCLUSIONS: It is argued that this study's research design is more accurate in reflecting the barriers that a pregnant Medicaid-eligible patient encounters when seeking office-based prenatal care. As such, combining the physicians' receptiveness model with other physician participation in Medicaid models provides a more complete picture of access barriers to prenatal care for our most needy populations.