We evaluated the reproducibility of brainstem auditory evoked response
s (BAERs) in 87 normal individuals in a longitudinal study by estimati
ng the correlation coefficients and variability of the interpeak inter
vals and the V/I amplitude ratio between trials on the same day and be
tween sessions spaced 2 years apart. The highest correlation coefficie
nts occur for the I-V interpeak interval between trials on the same da
y. The coefficients for the I-III and III-V intervals are lower, due t
o the variability of wave III. The correlations between ears done on t
he same day are lower still and are similar to measures obtained from
the same ear at a 2-year interval. BAERs are more variable than previo
usly believed between ears and over time, but not in a manner that is
clinically significant and can be used longitudinally as a measure of
neurologic disease. Finally, we provide the sample size required to de
tect a significant change in interpeak intervals.