Adverse obstetric events have been implicated as risk factors for schizophrenia. Many studies have relied on retrospective recall of these events, given typical adult onset of schizophrenia, when most studies ascertain their samples. The goal of this study was to assess the validity of an interview for the long-term recall of prenatal and perinatal events. Ninety-six women from the Providence and Boston cohorts of the National Collaborative Perinatal Project were administered a brief structured telephone interview regarding their recall of pregnancy-related events that had occurred 22 years or more prior to interview. Women accurately reported major medical events such as cesarean section, breech delivery, and multiple birth (kappa=1) and demographic items, such as age at birth and parity. Medical interventions and major medical conditions such as placental (kappa=-0.01) and cord (kappa=-0.10) difficulties were not accurately reported. Estimated birthweight, low birthweight, and length of gestation were recalled with reasonable accuracy. Women who completed high school generally recalled events more accurately than those who did not. It is therefore important to attend to the sources of information, the mode of interviewing, the specific type of event, and sociodemographic characteristics of the informant to improve the accuracy of retrospective data on obstetric events, and to increase the validity of findings relating these to the onset of schizophrenia.