Date of Defense
Date of Graduation
Objective: To evaluate the difference between nasalance measured using overall nasalance for the full set of syllable repetitions in a speech sample contrasted with syllable repetitions selected (trimmed) from the overall sample.
Method: Participants included 24 males and 34 females between 18 and 30 years of age who participated in a normative study of nasalance in Michigan’s lower peninsula. Participants produced 14 syllable stimuli. Each syllable sequence was repeated at least 8 times. Three trials of each repetition were recorded together with other speech stimuli. Overall nasalance was calculated for each syllable repetition sequence (whole) and compared with the mean nasalance across 5 selected (trimmed) syllables. Syllable selection was conducted using MATLAB (Mathworks, 2011) to identify syllable onset and to select syllables 2-6 in the repetition sequence.
Results: Analyses indicate that the mean difference between overall and trimmed nasalance was centered around 0 with a broad variance distribution and a range of EDITEDIT. A multilevel multivariable regression (X2(3, N=2110) = 32.79, p<0.0005) associated with mean differences include consonant manner class (stops, sibilants, nasals) and vowel (/i/ versus /a/) with significant differences observed for oral consonants, but not nasal consonants.
Conclusion: Nasalance measures are statistically different when overall sample is compared with selected syllables. In most cases, differences are small and thus not clinically significant. In a few cases, however, differences were substantial and could yield clinically relevant differences. Further research is needed to define clinical significance and to explore best trimming practices.
Peebles, Jackson, "Overall Nasalance versus Trimmed Selection of Stable Syllable Repetition" (2013). Honors Theses. 2374.
Honors Thesis-Open Access