Abstract: The invention discloses a process for the automatic determination of hearing of newborns and infants, by detecting the auditory brainstem responses (ABRs) or otacoustic emissions (OAEs), by testing in a frequency domain, by applying a predetermined initial acoustic stimulation level to a subject, pre-averaging a predetermined, sufficient, number of sweeps to yield a set of subaverages (SA), calculating a predetermined number, n, of frequency spectra based on the subaverages (SA), applying a q-sample uniform scores test or a q-sample Watson U.sup.2 test to determine phase and/or amplitude differences among the n frequency spectra. The number of frequency spectra, n, is determined by continuously averaging a predetermined number of initial sweeps to provide a predetermined number, m, of subaverages (SA), calculating m frequency spectra based on the m subaverages, and repeatedly applying the q-sample uniform scores test or the q-sample Watson U.sup.