Interrater reliability in the application of the VPAS-PB: analysis among experienced judges.

Only 10 people were rated, and all were men aged 34 to 49 from the same city, so the consistency figures come from a very narrow set of voices. Everyone in the study had a healthy voice: people with an obvious voice change, a voice complaint, occupational voice use or reading difficulty were excluded. The protocol was therefore never tested on the dysphonic clients a voice clinic actually sees. All three judges had phonetics training and at least four years of VPAS-PB experience, so the agreement reported here is likely better than a typical clinical team, or a newly trained clinician, would achieve. Only three judges took part, and the authors state they were measuring the internal homogeneity of the judgments rather than variability between individual raters. The authors note that some adjustments may simply have been used rarely by these speakers, which they say could explain the weaker results for lip stretching and falsetto. The recorded material was one read sentence and two short narratives per speaker, so adjustments that do not appear in that kind of speech would not have been elicited.

The funding and conflicts section reports financial support from CAPES, a Brazilian federal agency for postgraduate education (grant 88887.15.1038/2017-00). No competing interests are declared.

The reliability figures here come from three judges with at least four years of VPAS-PB experience rating 10 vocally healthy men, and the authors state that they measured the internal homogeneity of the judgments, not variability between raters. Do not read them as proof that you and a colleague will label the same dysphonic voice identically.