Sounds from the ear canal were measured and then analyzed off-line. A
peak-picking algorithm located spectral maxima which might be designat
ed as spontaneous otoacoustic emissions (SOAEs). The output from a 0.5
-cc syringe, used to simulate the volume of the ear canal, was also me
asured, analyzed and used to approximate the false-alarm rate of the m
easurement system. SOAE prevalence estimates depended on the false-ala
rm rate, just as the hit rate in a yes-no task does. With a false-alar
m rate of zero, trends that have been found to be significant in the p
ooled results of other surveys were replicated: more SOAEs in the righ
t ear, more SOAEs in females, and increased probability that the contr
alateral ear has an SOAE if the ipsilateral ear has an SOAE. In additi
on, many SOAEs failed to be detected in consecutive spectral analyses
because they fluctuated in level.