Identification of children with and without cleft palate from tape-recorded samples of early vocalizations and speech

Thirty judges (5 speech pathologists, 10 mothers of children with cleft palate, and 15 mothers of noncleft children) listened to 90 tape-recorded samples of early vocalizations/speech obtained from noncleft babies and babies with cleft palate. Each sample was classified by the judges as normal or abnormal. As a group, the speech pathologists classified only 60% of the cleft samples as abnormal and 59% of the normal samples as normal. The cleft and noncleft mother groups, on the other hand, classified 37% and 25% of the cleft samples as abnormal and 59% and 73% of the normal samples as normal. Poor interjudge agreement was evident within and across the three groups of judges. The poor reliability demonstrated by the speech pathologists in identifying babies with unrepaired clefts appeared related more to a difference in interpretation of the perceptual data than an inability to hear salient information.