Reliability of the Tillman-Olsen procedure for determination of spondee threshold using recorded and live voice presentations.

This study investigated the reliability of the Tillman-Olsen procedure for establishing the spondee threshold (ST). Monitored live voice (MLV) and the Auditec of St. Louis recordings of the Central Institute for the Deaf spondees were used as stimuli. Ninety-five percent of 96 normal hearing subjects had test-retest STs within 5 dB and no subject varied more than 8 dB. Although these results represent good short term reliability for the Tillman-Olsen procedure, repeatability is no better than that produced by other procedures. Thresholds obtained by MLV were as reliable as the recorded spondees. It was concluded that recorded spondees offer no advantages over MLV testing for the experienced clinician. The recorded presentation gave STs that are very close to the normative value for speech specified by ANSl-1969. Live voice testing showed STs 4.7 dB better than when the Auditec recordings were used.