Reliability of Listeners' Stuttering Counts: The Effects of Instructions to Count Agreed Stuttering

Abstract

The reliability with which listeners record or count stuttering events has long been of concern in stuttering treatment and research. Two recent studies (Kully & Boberg, 1988; Ingham & Cordes, 1992) found low levels of agreement in total stuttering counts across treatment centres. One possible reason for low agreement is the perceptual overlap between stuttering and normal disfluency. The present study investigated the intrajudge and interjudge agreement of seven experienced clinicians who worked closely together in the same treatment centre. In Part One of the study the judges were found to have high intrajudge reliability; however, there were considerable discrepancies in the judges' total stuttering counts. Part Two of the study investigated the effect on interjudge agreement of instructing the judges to count only those events which they were “certain other judges would agree are stuttering.” The total counts decreased significantly in response to this instruction but the level of interjudge agreement...