BACKGROUND The influence of physiological and methodological factors on recordings of brainstem auditory evoked potentials (BAEPs) is greater in children than in adults. OBJECTIVE To collect and evaluate BAEP data in normal children, and measure intra- and inter-laboratory variability. METHODS Seven hundred and fifty unselected BAEP recordings were collected and evaluated from children ranging from neonates to 14-year-olds by eight laboratories in Italy. RESULTS In newborns, three laboratories showed satisfactory concordance; wave I was more broadly distributed than wave V and IPL I-V. The evaluation of pooled BAEP data from the older children showed that laboratories with age-matched data gave overlapping results; those with unmatched-age data differed significantly. The sound intensities of the laboratories did not significantly affect absolute BAEP latencies or IPLs. Females had shorter latencies than males; the difference was not significant. A single exponential regression model was an adequate but not the best predictor of normal data. CONCLUSIONS The pooled data were consistent with the physiological maturation of the brainstem acoustic pathway. The BAEPs was reliably normalised using the natural logarithm of age. The differences between Centres were related to sample size, measurement accuracy, and inclusion and selection criteria. CONCLUSIONS The creation of multicentre common database from an unmatched data collection is feasible and reliable enough for clinical diagnosis and multicentre clinical research.