Free field threshold variance

Abstract

Values for free field threshold variance have been obtained using clinically normal subjects in experiments conducted over a period of three months. Smaller values of variance were obtained than those yielded by similar MAP experiments by up to 13 dB 2. A consequence of this is that threshold uncertainty is reduced by up to 9 dB at the 1 % level of significance when a threshold is redetermined using MAF rather than MAP techniques over a period of longer than one day. Supporting experiments showed that approximately one-half of the MAF long term within ear variance was due to changes in the acuity of normal ears and the remainder due to the combined effects of the experimental method and to the assessment of the threshold by the subject. It was not found possible to resolve out the component variances of this remainder but evidence suggests the greater part is accounted for by the subject's uncertainty of what should be his threshold criterion.