Variability in judgments of articulation when observer listens repeatedly to the same phone

This study pertains to variability in articulation judgments made by Os who listened to repetitions of the same stimulus. In the first portion of the study, four Os were asked to listen to six tapes, each of which was prepared to present 90 repetitions of a syllable. The syllables were /sa/ and /rk/.2 The /s/ and /r/2 in two of the syllables were correctly articulated, in two they were moderately distorted, and in two they were severely distorted. Correctness of articulation of the target phones had been established previously by a panel of judges using a scaling procedure. The speakers were children with articulation problems. A total of 24 sets of judgments were obtained. For three of those sets, O frequently shifted between assessment of the phone as correct and incorrect. In the second portion, each of three tapes was evaluated by five of 15 judges. Two tapes were reused from the first part of the study, and a new tape presenting the same production of /is/ 90 times was also used. No judge listened to more than one tape. Again, a record was made of the frequency with which judges shifted in their assessment of a given articulation as correct or incorrect. More judges reported 12 or more shifts in the second portion of the study than in the first. O's consistency should be investigated in the situation where O evaluates a series of similar misarticulations of phones from the same phoneme.