KMID : 1036820130180040368
|
|
Communication Sciences & Disorders 2013 Volume.18 No. 4 p.368 ~ p.378
|
|
Reliable Sample Size for Mean Length of Utterance Analysis in Preschooler
|
|
Yoon Mi-Sun
Kim Soo-Jin Kim Jung-Mee Chang Moon-Soo Cha Jae-Eun
|
|
Abstract
|
|
|
Objectives: The purpose of this study is to obtain a reliable minimum sample size for the mean length of utterances (MLU) analysis.
Methods: The participants were 30 typically developing children of ages between 2-4 years. Conversations with their caregivers at homes were recorded and transcribed by trained transcribers. Four sets of speech samples of different length, 50-100-150-200 utterances, were selected successively from original transcriptions. The same procedure was repeated 3 days later and a total of 240 sets of speech samples were prepared. t-tests for comparing the MLU between 1st and 2nd speech samples were done. Correlation analysis of MLU in speech samples of different length was performed.
Results: The MLUs were not significantly different between the 1st and 2nd samples. Pearson correlation coefficient scores for MLU-eojeol of 50 utterances were over .90 in 3 different age groups. Meanwhile these of MLU-word and MLU-morpheme were over .90 respectively when sample size was 100 utterances in the groups of 2 and 4 years old. In 3 years old group, they were over .90 in sample of 50 utterances.
Conclusion: The result of this study suggests that the speech sample of 50 utterances should be considered as a reliable sample size for the analysis of MLU-e, excluding the 100 utterances for the analysis of MLU-w and MLU-m. It does not mean that this sample size is adequate to analyze other quantitative parameters regarding semantics. Further research is required for this issue.
|
|
KEYWORD
|
|
MLU-e, MLU-w, MLU-m, MLU, Speech sample, Sample size, Reliable size, Language development
|
|
FullTexts / Linksout information
|
|
|
|
Listed journal information
|
|
|