Reliability issues with the Synthetic Sentence Identification test.

This investigation examined the reliability of the Synthetic Sentence Identification (SSI) test. An adaptive procedure was used to assess speech recognition thresholds (SRTs) for the SSI lists over two separate test sessions. The commercially available Davy Crockett passage and a multitalker babble were used as competition. Subjects consisted of 20 adults with normal hearing sensitivity. Results indicated (1) significantly higher SRTs with the multitalker babble, (2) significant differences in SRTs among SSI lists when using the Davy Crockett competition, (3) no significant learning/practice or fatigue effects across trials in either noise condition, and (4) high test-retest reliability across test sessions. Clinical implications of these data are considered.