Whole word scoring versus phoneme error scoring for audiological word recognition testing

Conventional audiological word recognition testing uses whole word scoring for errant responses resulting in a percent correct score. This approach is found to be lacking in diagnostic utility. Recently, phoneme error scoring was suggested as a supplement to whole word scoring in attempt to improve diagnostic utility but is also limited in that it is highly correlated with the whole word score. This project examined the use of phoneme error scoring that applies an exponential equation to the type and the number of phoneme errors yielding a score that may be compared to absolute hearing sensitivity loss as measured by conventional audiograms. Because the number of words used in audiological word recognition varies, a correction for the number of stimuli is applied. Archival data, case study data, and simulated cases were used to evaluate and compare the diagnostic utility of three scoring approaches for errant responses: whole word, total phoneme errors, and an exponential phoneme error score (EPES) that weights consonant and vowel errors differently: EPES = number of phoneme errors times 2number of consonant errors + 10number of vowel errors. The case study and simulated results were used to evaluate the utility of each scoring approach.