The acceptable noise level: the effect of repeated measurements

This study explores if increasing number of repetitions might improve the precision of the acceptable noise level (ANL) test.

We measured twelve ANL repetitions, i.e. four complete ANL tests (4 × 3 repetitions), at one session using diotic presentation and a non-semantic ANL version.

Thirty-two normal-hearing adults.

Small order and fatigue effects were seen. We used the coefficient of repeatability (CR) to assess the repeatability; CRs ranged between 3.9 and 7.6 dB for the four ANL tests. Using the twelve ANL repetitions we removed the variability of the ANL across subjects by normalizing the data to the individual mean ANL for the twelve repetitions. The mean normalized ANL across the subjects rapidly approached the ANL normalized to the individual mean for the 12 repetitions (0 dB), and after three repetitions the SD seemed to be stable at about 3 dB.

The findings suggest that both order and fatigue affect the ANL. The findings also suggest that it may be more accurate to speak of an acceptable noise range than ANL. These findings have large implications for how we understand acceptable noise and it would explain a large part of the variability seen among normal-hearing and perhaps hearing-impaired subjects.