Two ascending and two descending procedures were used to obtain spondee thresholds from normally hearing and sensorineural hearing impaired subjects. Minimal differences were found when comparisons of threshold were made among all four procedures. Results indicated slightly more sensitive thresholds when the descending procedures were used and slightly poorer pure-tone average and spondee threshold agreement. Test-retest reliability was good for all four procedures, but large differences existed among the procedures in the amount of time and in the number of words required to establish threshold. Advantages and clinical implications are discussed.