Word recognition performance with modified CID W-22 word lists.

Abbreviated CID W-22 lists were administered to large groups of normal and hearing-impaired listeners to test the hypothesis, that fewer, judiciously chosen items can be used to test word recognition without compromising test accuracy. Data were analyzed by comparing each subject's performance on half- and 10-word lists to full-list scores. Sensitivity and specificity for various sublists and for several pass/fail criteria were calculated. Results show that fewer than the traditional 50 items can be used in word recognition test procedures if the words are sufficiently difficult and strict passing criteria are employed. We recommend terminating testing after 10 words if no errors occur and after 25 words if there are no more than four errors. Otherwise, a full 50-item list should be administered.