A basic problem encountered by investigations of prenatal cocaine effects has been the valid identification and quantification of exposure. Based on a combination of sources: (a) medical record review, (b) maternal urine toxicology screen, (c) meconium analysis, and (d) maternal postpartum interview, drug exposure status of 415 infants was established. Using this combination as a benchmark, maternal postpartum interview was found most sensitive, while medical record review was slightly less accurate. Meconium analysis and urine screens both demonstrated miss rates greater than the interview or record review methods. Meconium analysis and postpartum interview, however, each detected cases of cocaine exposure that the other had missed. Correlations between the amount of cocaine found in meconium and in maternal report indicated that the cocaine metabolite benzoylecgonine was the best biological marker. Quantifying heavy versus light exposure required a combination of both meconium analysis and maternal postpartum interview techniques.