Acoustic-phonetic description of infant speech samples: Coding reliability and related methodological issues

Abstract

Two samples of speech-like vocalizations were recorded from each of 18 infants who were 8, 12, or 18 months of age at the time of recording. The two samples were recorded on different days, with less than one week between recordings. Each utterance was coded as belonging to one of several possible infraphonological categories, and canonical syllable ratios were determined for each sample. Syllables produced with abnormal phonation were identified. These coding procedures were completed independently by two raters. One sample from each infant was coded twice by the same rater. This sequence of multiple recordings and repeat analyses allowed for the determination of interrater, intrarater, and test-retest coding reliability. Kappa and intraclass correlation analyses revealed excellent reliability for all measures. © 2001 Acoustical Society of America.