Consonant identification in quiet and in noise with the normal and the sensorineural hearing-impaired

This study determined the effect on consonant identification of speech-to-noise ratio; the California Consonant Test (CCT) was given to 20 normal-hearing young adults and 14 patients (mm age: 56.1 yrs) with sloping sensorineural hearing losses, SRT greater than 20 db, and DS less than 90% at MCL. The CCT was given individually at MCL in quiet and at that level in broad-band noise adjusted to yield 20, 10, 0, and -10 db S/N, consecutively. Mean percent-correct scores for the patients were 50, 44, 40, 38, and 32 in order, and were 97, 90, 73, 47, and 37 for the controls. Differences were significant between groups except for the 2 most difficult ratios. For the patients, regression analysis revealed that the slope of the line of best fit was not significantly different from zero. Confusion matrices constructed for each of the 5 noise conditions for each group revealed that at S/N of 10 db, normal Ss began consistent and systematic substitutions in manner and in place of articulation, never in voicing or nasality. This pattern was in general followed by the patients, except that substantial confusions existed also at the 2 easiest ratios.