Identifying the authoritative judgments of stuttering: comparisons of self-judgments and observer judgments

Reliable and accurate stuttering measurement depends on the existence of unambiguous descriptions or exemplars of stuttered and nonstuttered speech. The development of clinically meaningful and useful exemplars, in turn, requires determining whether persons who stutter judge the same speech to be stuttered that other observers judge to be stuttered. The purpose of these experiments, therefore, was to compare stuttering judgements from several sources: 15 adults who stutter, judging their own spontaneous speech; the same adults who stutter, judging each other's speech; and a panel of 10 authorities on stuttering research and treatment. Judgments were mode under several conditions, including self-judgments made while the speaker was talking and self- and other-judgements made from recordings in continuous and interval formats. Results showed substantial differences in stuttering judgments across speakers, judges, and judgment conditions, but across-task comparisons were complicated by low self-agreement for many judges. Some intervals were judged consistently by all judges to be Stuttered or Nonstuttered, across multiple conditions, but many other intervals were either not assigned replicable judgments or were consistently judged to be Nonstuttered by the speaker who had produced them but were not assigned consistent judgments by other judges. The implications of these findings for stuttering measurement are considered.