The purpose of this study was to evaluate the predictability status of several sentence lists for use in hearing tests of Navy personnel. One hundred and eighty normal-hearing subjects, divided into six groups, listened to three sentence lists presented in one of six word-elimination conditions. The sentence lists were scored in terms of the number of words correctly predicted out of the number of words eliminated. The following results were obtained: (1) It was necessary to remove between 20 to 25 keywords for Sentence Lists B and D, respectively, before obtaining a significant score reduction. These lists were found to be highly predictable and quite similar with regard to their response to keyword elimination. (2) Considerably lower scores were obtained for Sentence List C than those obtained for Lists B and D, Furthermore, increased degradation through increased key-word elimination did not significantly affect the predictability scores. It was concluded that List C is considerably less predictable than Lists B and D. (3) The error curve for a Synthetic Sentence List was found to be extremely low and relatively flat, suggesting that this group of sentences has a negligible predictability value.