Test-Retest Reliability Bounds for Consonant-Nucleus-Consonant Word and Phoneme Scores for Cochlear-Implant Users.

OBJECTIVES

This report establishes test-retest 95% confidence interval bounds for consonant-vowel-consonant (CNC) word- and phoneme-recognition scores for cochlear-implant listeners.

DESIGN

Data from two published cochlear-implant studies were analyzed. Across-list SDs for individual subjects were compared with the expected SD for a binomially distributed variable. The model was adjusted by tuning the effective binomial distribution set size to minimize the root-mean-square deviation from measured SDs.

RESULTS

The effective CNC-word set size equaled the number of words per list (50). The effective phoneme set size (approximately 100) was lower than the number of phonemes per list (150).

CONCLUSIONS

Results support the hypotheses that individual CNC words are recognized independently, and that there is mutual information across phonemes, resulting in a smaller effective set size. A table of confidence intervals and critical-difference values for CNC word and phoneme recognition is provided.