High frequency Bekesy audiometry: II. Threshold test procedure, reliability and validity.

Normal-hearing Ss (N: 20) produced 2 sets of Bekesy threshold tracings, from 2-8 kc/s, on a conventional Grason-Stadler E-800 audiometer, and from 2-20 kc/s on a unit (HF E-800) modified for high frequencies. Half the Ss then produced at least 2 sets of tracings each in the fixed-frequency mode on the HF E-800. The pulsed-tone fixed-frequency procedure is recommended for HF HTLs. Maximum mean test-retest HTL differences of 1-4 db were obtained. Reliability r's obtained from the HF E-800 were equivalent to or surpassed those from the E-800; a very high proportion of such coefficients was statistically significant. Reduced validity coefficients, calculated at the frequencies common to the units, were attributed at least in part to differences in methods of coupling the transducers to the ear. It was concluded that the HF E-800 unit in the fixed-frequency mode could be used with some confidence to assess HF HTLs.