Background: Palatalized misarticulation (PM) can occur with a variety of tongue-palate contact patterns in the speech of children with repaired cleft palate. In order to assess the degree of placement deviation by perceptual evaluation, a novel “backing score” for PM was developed.
Objective: To investigate the perceptual reliability and classification of the novel “backing score” for PM.
Design: Prospective study of perceptual judgments of speech using multiple listener ratings of articulation.
Outcome Measures: Four participants (2 listeners specializing in phonetics and 2 speech therapists) rated 55 audio-recorded Japanese speech samples containing PM, produced by children with repaired cleft palate. Listeners transcribed two backing scoring patterns, with/without a retracting mark (-), for five mono-syllabic VCV speech samples, [asa], [ata], [oto], [utsu] and [iɕi], using the International Phonetic Alphabet (IPA). One week later, perceptual judgement was repeated to assess intra-rater reliability. Inter-rater reliability was also computed.
Results: Higher intra- and inter-rater reliability was observed when coding backing errors without retracting marks. Reliability was the highest for the plosive VCV [oto], and the lowest for the fricative VCV [iɕi]. Also, regarding articulation placement for PM, three sites were perceived: soft-palatalized, palatalized and post-alveolar.
Conclusion: Deriving a backing score without retracting marks may be a useful tool to subclassify PM errors. However, there are limits to the accuracy of perceptual judgments of backing, especially for fricatives. Future studies should consider including audio-visual stimuli, such as electropalatography (EPG), to determine if this facilitates improved rater accuracy and reliability of error coding. Transcription findings suggested that PM may be subdivided into post-alveolar, palatalized and soft-palatalized place of articulation.
View full abstract