Doyle J
Ear Hear. 1987 Jun;8(3):170-4. doi: 10.1097/00003446-198706000-00007.
The reliability of intelligibility judgments of hearing-impaired children's speech was investigated. Twenty Australian audiologists, untrained in the use of speech intelligibility scales, rated the speech intelligibility of 20 hearing-impaired children, using the National Technical Institute for the Deaf speech intelligibility scale on one occasion and a percentage intelligible scale on another. Inter-rater reliability was high for both scales (Kendall's W greater than 0.80, p less than 0.001), as was intra-rater reliability (r greater than 0.90, p less than 0.001; rcc greater than 0.88). Training in the use of speech intelligibility rating procedures does not appear necessary for reliable judgments. However, the perceived intelligibility of some children's speech was highly listener dependent regardless of the scale used, indicating that a variance measure should be reported in addition to means in descriptions of perceived intelligibility.
对听力受损儿童语音清晰度判断的可靠性进行了调查。20名未接受过语音清晰度量表使用培训的澳大利亚听力学家,使用美国国家聋人技术研究所的语音清晰度量表对20名听力受损儿童的语音清晰度进行了一次评分,并在另一次使用了可理解百分比量表进行评分。两种量表的评分者间信度都很高(肯德尔W系数大于0.80,p小于0.001),评分者内信度也很高(r大于0.90,p小于0.001;rcc大于0.88)。可靠的判断似乎并不需要语音清晰度评分程序的使用培训。然而,无论使用何种量表,一些儿童语音的可感知清晰度都高度依赖于听者,这表明在描述可感知清晰度时,除了均值外还应报告方差测量值。