Universal infant hearing screening has recently been recommended by the National Institutes of Health. Otoacoustic emissions have been proposed as the first-level screening technique. Although transient evoked otoacoustic emissions have shown limited applications, distortion-product otoacoustic emissions hold promise as a screening technique but have not been fully investigated. The purpose of this study was to determine the validity of distortion-product otoacoustic emissions as a hearing screening technique. A total of 208 ears of 104 infants at risk for hearing loss were tested with both automated auditory brain stem response and distortion-product otoacoustic emission screening protocols. Acoustic brain stem response results were used as the standard for normal hearing. Distortion-product otoacoustic emission results were analyzed by means of calculation of the difference between the mean of the response levels and the mean of the noise floor levels from five frequency pairs between 2000 and 4000 Hz. Pass-fail rates for response above noise floor criteria of 5, 10, and 15 dB were examined. The sensitivity of distortion-product otoacoustic emissions was 50%, 67%, and 87%, and the specificity was 94%, 68%, and 38% at the 5, 10, and 15 dB levels, respectively. The pass-fail criterion of distortion-product otoacoustic emissions should be based on instrumentation calibration, infant status, and an acceptable false-positive, false-negative yield. The ability to test rapidly the hearing of all infants with distortion-product otoacoustic emissions points to the feasibility of using this test as a first-stage screen.