Resonance is a phenomenon in which sounds made by the vocal cords are amplified or dampened in the oral and nasal cavity, and communication disorders caused by the problems in these cavities are called resonance disorders. Both subjective auditory perceptual assessment and objective nasometer measurement are used for assessing resonance-related problems. Nasometer is an objective assessment instrument that is most commonly used in the clinical field to find whether or not nasality is disordered; and can be used for everyone from children aged 3 to adults because it is non-invasive and easy to use. It also has the merit that it can be used for various patients with cleft plate, voice disorder, neurogenic language disorder, etc. Yet, it has the problem of producing different nasalance scores due to several factors. Previous studies showed that phonological environments, stimulus lengths and ages affected nasalance scores; and sessions of test and retest, though conducted on one single patient, showed different nasalance scores. Watterson & Lewis (2006) conducted a study with cleft plate patients and found that those who changed the headgear had a greater test-retest variability than those who did not. Results of test-retest measured at varying times by Lewis et al. (2008) showed nasalance scores variability increased as the time went by. Whitehill
(2001) conducted a daily nasalance score tests with a nomal woman and showed that there was a greater nasal consonant stimulus variability on the second day than on the first day. This kind of nasalance test-retest score variability shown in a single patient would prevent an
accurate measurement of subjects’ nasalance scores, and thus can become a clinical issue. However, there are no in-depth studies on nasalance scores test-retest in Korea, and thus this study sought to investigate whether or not phonological environments, stimulus lengths and ages would affect the reliability of test-retest by examining that of test-retest of nasalance scores. The subjects of this study were one group of 20 normal adults (male, n=10; female,
n=10) and three groups of 15 children (male, n=8; female, n=7). The children of each group were 3, 4, and 5 years old respectively. All of them were selected after being judged to be without auditory perceptual problems or any esonance disorder. Stimuli were set with /i/ vowel contexts, /a/ vowel contexts, and /mixed/ vowel
contexts of /i/ and /a/ vowels. The reason why /mixed/ vowel contexts were added was that there is no occasion in conversation where a particular vowel is focused and used by differentiating the words which contain /i/ and /a/ vowels. Lengths were set with the durations of 4, 8, 16, and 31 syllables, and the identical stimuli were repeated after by a single person a total of 2 times. (Table. 2).
SPSS 22, a statistical program, was used for data analysis. Three-way repeated measure ANOVA was performed with 3 independent variables (vowel contexts, lengths of the stimuli, and ages) based on the variability values of the nasalance scores between test and retest. As a result, significant differences were found in /i/ vowel contexts and 4 syllables between test and retest, but no significant difference was found in ages. To sum up, /i/vowel contexts and stimuli with the length of 4 syllables have a great test-retest variability in common, and thus should be carefully selected and interpreted in clinical trials. The reason why /i/ vowel contexts, despite their great variability, are important as stimuli to assess hypernasality is that they may produce nasal emissions in the case of consonats such as stops and fricatives and can be replaced even with glottal sounds. In addition, if lengths of
stimuli are too short, they may lower reliability; and if too long, they may become difficult for children and hard to be clinically applied. Thus, judging by the results of this study and previous studies, the length of fewer than 20 syllables thought be appropriate. To summarise the results, the lengths of the stimuli appropriate for assessing ypernasality will be the duration of 20 syllables in consideration of the propotrion of /i/ vowels. However, due to great test-retest variability in the cases of the /i/ vowel contexts and the short length of 4 syllables, only one time of measurement of nasalance scores of subjects can be a problem, and may lead to a controversy when post-operative and post-treatment effects are discussed.
Therefore, calculating averaged nasalance scores using at least two or three times of assessment is expected to produce highly reliable nasalance scores. In addition. using averaged nasalance scores by means of at least two or three times of assessment will be appropriate
regardless of stimulus contexts when clinically assessing nasal sounds, because they sometimes showed more than 5% of difference in other stimuli as well even though appearing less frequently in the length of 4 syllables under the /i/ vowel contexts. This study takes notice that there is difference in nasal scores between test and rest, depending on stimulus contexts, and thus more caution is needed for selecting and interpreting stimuli in clinical trials Also, the study suggests examining individuals’ characteristics and using averaged nasalance scores by means of at least more than two times of assessment in order to attain highly reliable nasalance scores in onsideration of variability of nasalance scores.