Nasalance and Nasality in Low Pressure and High Pressure Speech

THOMAS WATTERSON, PH.D.

KERRY E. LEWIS, PH.D.
CANDACE DEUTSCH, M.S.
Objective: This study compared nasalance measures and nasality ratings in
low pressure (LP) and high pressure (HP) speech.
Subjects: The subjects for this study were 25 children ranging in age from
5 to 13 years. Twenty of the subjects were patients followed by a craniofacial
team, and five had no history of communication disorder.
Results: The mean nasalance for the LP speech was 29.98% (SD, 16.16), and
the mean nasalance for the HP speech was 30.28% (SD, 15.35). The mean nasality rating for the LP speech was 2.31, and the mean nasality rating for the
HP speech was 2.59. Separate paired t tests revealed no significant difference
between the LP or the HP speech for either the nasalance scores or the nasality
ratings. The correlation coefficient between nasalance and nasality for the LP
speech was r 5 0.78, and for the HP speech r 5 0.77. Using a cutoff of 26%
for nasalance and 2.0 for nasality, Nasometer test sensitivity was 0.84 and test
specificity was 0.88.
Conclusions: In general, clinicians may obtain valid measures of nasalance
and/or ratings of nasality using either an LP stimulus or an HP stimulus. Sensitivity and specificity scores indicated that the Nasometer was reasonably
accurate in distinguishing between normal and hypernasal speech samples.
KEY WORDS: low pressure and high pressure speech, nasalance, nasality, nasometer

Researchers have sought to explore the relationship between

nasalance scores computed by the Nasometer (Kay Elemetrics
Corp., Lincoln Park, NJ) and listener judgments of nasality.
Some studies have reported a strong correspondence between
nasalance and nasality (Dalston et al., 1991), while others have
reported a weaker relationship (Hardin et al., 1992; Nellis et
al., 1992; Watterson et al., 1993).
It has been speculated that one source of variation across
studies might involve confounding effects of nasal air emission
on nasalance scores (Watterson et al., 1993; Karnell, 1995).
Nasalance is intended to be a measure of the acoustic energy
that occurs primarily on vowels, glides, and liquids, but nasal
air emission is an aerodynamic phenomenon associated with
high pressure consonants, principally stops, fricatives, and affricates. In the presence of nasal air emission, the Nasometer
scores would not discriminate between the aerodynamic acoustic energy associated with nasal air emission and the acoustic
nasal resonance associated with hypernasality. Thus, if the

speech stimulus contained high pressure consonants, nasalance

scores could be spuriously elevated relative to listener judgments.
To test that hypothesis, Karnell (1995) compared a high
pressure stimulus (HP) loaded with stops, fricatives, and affricates and a low pressure stimulus (LP) that contained only
vowels, glides, and liquids. Using a nasalance cutoff of 31%,
Karnell divided the obtained nasalance scores into three
groups. One group had both LP and HP nasalance scores that
were above the cutoff, and for these subjects, there was no
significant difference between the LP and HP scores. Another
group had both LP and HP scores below the cutoff. For this
group, there was a significant difference in the LP and HP
scores. The third group had mixed results. That is, some had
LP scores below the cutoff and HP scores above the cutoff,
while some had HP scores below the cutoff and LP scores
above the cutoff. This latter group was identified by Karnell
as a subgroup with marginal velopharyngeal dysfunction
(VPD).
The results obtained in the Karnell (1995) study are intriguing, but his findings have limited clinical application without
perceptual validation. That is, if the phonetic content of stimuli
affects nasalance scores, it is important to know if listener
judgments are also differentially affected. In that regard, there
are two important questions that need to be addressed. First,
we do not know how statistical differences, or lack of differences, in nasalance scores translate to perceptual reality. For

Dr. Watterson is Professor, Dr. Lewis is Associate Professor, and Ms. Deutsch
is Research Assistant, University of Nevada School of Medicine, Reno, Nevada,
U.S.A.
Submitted July 1997; Accepted March 1998.
A paper presented in part at the American Cleft Palate-Craniofacial Association Annual Convention, April 1997, New Orleans, Louisiana.
Reprint requests: Dr. Thomas Watterson, University of Nevada School of
Medicine, Department of Speech Pathology and Audiology, Redfield Building/
152, Reno, NV 89557; E-mail tw@med.unr.edu
293

294

Cleft PalateCraniofacial Journal, July 1998, Vol. 35 No. 4

example, Karnell reported a significant difference between the

LP and HP nasalance scores for subjects below the cutoff of
31% nasalance, but the actual difference in the two nasalance
means was small. The LP mean was 14.4% and the HP mean
was 16.7%. Thus, while this difference may be statistically
significant, it is not known if listeners can actually hear such
small differences in nasalance. In contrast, it may be that listener discrimination of nasal resonance is more sensitive than
measures of nasalance. Thus, failure to find statistically significant differences in nasalance scores between LP and HP
stimuli does not necessarily mean that listeners cannot hear
differences.
The second question of interest is whether LP and HP
speech are perceptually equivalent. It may be that the degree
of nasality is more prominent to perception in either LP or HP
speech because of the different phonetic content. Earlier studies have shown that judgments of hypernasality may be influenced by phonetic content, but there has been no investigation
of differences in perceptual judgments of nasality between LP
and HP speech (Spriestersbach, 1955; Spriestersbach and Powers, 1959; Carney and Sherman, 1971). If listeners do, in fact,
hear differences in nasality between LP and HP speech, then
this would have to be taken into consideration when interpreting nasalance scores.
The purpose of the present investigation was to obtain nasalance scores and nasality ratings for both LP and HP speech
in the same speakers. This would permit comparison of nasalance scores for LP and HP speech, comparison of nasality
ratings for LP and HP speech, and analysis of the relationship
between nasalance scores and nasality ratings for both LP and
HP speech. In addition, tests of sensitivity and specificity for
the Nasometer were obtained.
METHOD
Subjects
Twenty-five children ranging in age from 5 years, 0 months
to 13 years, 3 months served as subjects for this investigation.
Twenty of the subjects were patients who were followed by a
craniofacial team, and five subjects presented with no history
of communication disorder. For the patients, a mirror fogging
test indicated inappropriate, pervasive nasal air emission on
pressure consonants. This was done by holding a cold mirror
under one nare, digitally occluding the other nare, and having
the subject repeat single words and short phrases loaded with
pressure consonants. Although each patient demonstrated inappropriate nasal air emission, each patient did not necessarily
sound hypernasal. Patients with a pattern of velopharyngeal
fricatives were excluded as subjects because this articulatory
error may contaminate nasalance measures (Hashimoto et al.,
1992). The remaining five subjects, presenting with no history
of communication disorder, were included to extend the range
of nasalance scores for the purpose of computing sensitivity
and specificity.

Speech Stimuli
Two 45-syllable stimuli were used (see Appendix). The first
was comprised of nine sentences that contained only low pressure consonants and vowels (Karnell, 1995). The second was
excerpted from the Zoo Passage (Fletcher et al., 1989) and
was loaded with high pressure consonants.
Instrumentation
Nasalance scores were obtained using a Kay Elemetrics
Model 6200 Nasometer. From digitized nasal and oral signals,
the Nasometer computes a ratio of the nasal to nasal-plus-oral
acoustic energy. Nasalance is expressed as a percentage score
computed from that ratio (nasalance 5 nasal/[oral 1 nasal] 3
100).
Ratings of nasality were made from audio recordings obtained using a Sony Model TCM 5000EV cassette recorder
(Taeneck, NJ) with a 57dB signal-to-noise ratio using a Sharp
Shure Prologue (Manwah, NJ) 14H microphone.
Nasality Ratings
Nasality ratings were made by seven listeners experienced
with cleft palate speech. Listeners rated each sample using a
seven-point equal-appearing interval scale that ranged from 0,
representing normal nasal resonance, to 6, representing severe
hypernasality.
Listener Training
A training session familiarized the listeners with the task,
the concept of interval scale judgments, and the use of the
seven-point equal-appearing interval scale. Instructions germane to the rating task were provided, including the guideline
associating ratings of 0, 1, and 2 with degrees of normal nasality and ratings of 3, 4, 5, and 6 with
progressively severe hypernasality. Feedback on training trials
was provided, and listeners were permitted to discuss their
ratings on extent of nasality.
Two speech samples were selected by the authors to serve
as anchor stimuli to facilitate application of the seven-point
scale. The anchor stimuli represented the scale extremes of
normal nasal resonance (0) and severe hypernasality (6).
To facilitate listener ratings of degree of nasality, the anchor
stimuli were played for the listeners at the beginning of the
practice session, and upon listener request at any time during
the training session.
Reliability
An audiotape composed of 10 speech samples representing
the range of nasal resonance was used to assess reliability. The
anchor stimuli and reliability samples were available from a
pilot study and did not appear as test stimuli.
The seven listeners independently rated each of the 10 sam-

Waterson et al., NASALANCE AND NASALITY

ples using the seven-point scale. Of the 210 paired ratings in

the training session, 80% differed by no more than one scale
point. This level of interjudge agreement on nasality ratings
was deemed acceptable for data collection.
Procedures
Subjects were seated in a quiet setting with the Nasometer
headpiece adjusted so the separation plate rested comfortably
but firmly on the subjects upper lip and perpendicular to the
plane of the face. The microphone used to record the audio
signal for subsequent nasality ratings was placed slightly to
the right side of the Nasometer sound separator and approximately 6 to 7 inches from the subjects mouth.
Following a model from the investigator, subjects repeated
each sentence in the LP stimulus and each sentence in the HP
stimulus. The Nasometer was activated when the subject spoke
and paused when the examiner spoke. Order of presentation
of the two stimuli was counterbalanced. When a subject completed each stimulus, the Nasometer computed the mean nasalance score for that stimulus. Thus, two mean nasalance
scores were computed for each subject.
The 50 speech samples (25 Ss 3 2 Stimuli) were dubbed
in random order onto a master tape from which the listeners
rated nasality using the seven-point scale. The experimental
listening task was conducted in a sound-isolated room with the
listeners seated in a semicircle such that they were about equidistant from an Argos Model SCB 12D audio speaker (Genoa,
IL). For each subject, the mean of the seven listeners nasality
ratings was computed for the LP speech and for the HP speech,
yielding two mean nasality ratings for each subject.
Upon completion of the data collection, a further check on
interjudge and intrajudge agreement was conducted. An intraclass correlation coefficient was computed on listeners ratings
of all 50 speech samples yielding an r of 0.88, representing
excellent listener agreement (Morrant and Shaw, 1996). To assess intrajudge agreement, 10 of the 50 speech samples were
selected at random and rerated by the seven listeners. Comparison of listeners first and second ratings of these 10 samples revealed 57 (81%) of the 70 paired ratings differed by no
more than one scale point. More specifically, 31 of the 57
(44%) paired ratings were identical, and 26 (37%) were within
one scale point. The remaining 13 (18%) paired ratings differed by two scale points. Interjudge and intrajudge agreement
levels were deemed sufficient to proceed with data analysis.
Data Analysis
Two paired t tests were computed to determine if significant
differences existed in nasalance scores and/or nasality ratings
as a function of stimulus type. Pearson Product Moment Correlation coefficients were obtained to assess the relationship
between nasalance and nasality (SPSS, 1988).
A secondary analysis was conducted to determine test sensitivity and test specificity for nasalance measures (Barker and
Rose, 1979). Simply stated, test sensitivity is the percentage

295

TABLE 1 Mean Nasalance Scores, Mean Nasality Ratings, and

Associated Confidence Intervals for LP and HP Speech (n 5 25)
Stimulus
Low Pressure
Measure
Nasalance
Nasality

High Pressure

SD

95% C.I.

SD

95% C.I.

28.98
2.31

16.16
1.97

22.3135.65
1.503.12

30.28
2.59

15.35
1.89

23.9436.62
1.813.37

of subjects who are judged to be hypernasal and who are also

identified by the Nasometer as having excessive nasalance
(above a predetermined cut-off score). Test specificity refers
to the percentage of subjects who are judged to be normal and
who also score within normal limits on the Nasometer (below
the cut-off score). Consequently, as the number of disagreements between the listeners and the Nasometer increases, test
sensitivity and test specificity decrease.
To calculate test sensitivity and test specificity, it is first
necessary to determine a cutoff between normal and abnormal
nasalance and between normal and hypernasal resonance. The
cutoff between normal and hypernasal resonance was predetermined in this study by instructing the listeners that 0 to
2 on the seven-point rating scale represented different degrees of normal nasality and that ratings of 3 to 6 represented different degrees of hypernasality. Thus, any speech
sample with a mean nasality that was greater than 2.0 was
considered hypernasal.
RESULTS
Nasalance and Stimulus Type
For each of the 25 subjects, two measures of nasalance were
obtained: one from the LP speech and one from the HP speech.
Table 1 shows the means and standard deviations for the data.
The mean for the LP speech was 28.98% (SD, 16.16). The
nasalance percentages for the LP speech ranged from 5.58%
to 59.32%. The mean for the HP speech was 30.28% (SD,
15.35). The nasalance percentages for the HP speech ranged
from 9.03% to 59.10%. A paired t test revealed no statistically
significant difference in nasalance for the LP and HP speech
(t 5 0.80, df 5 24, p 5 .43).
Nasality and Stimulus Type
For each subject, means of the seven judges nasality ratings
were computed for the subjects LP and HP speech. Table 1
shows the mean of the mean nasality ratings for each stimulus.
The mean of the mean nasality ratings for the LP speech was
2.31 (SD, 1.97), and the mean of the mean nasality ratings for
the HP speech was 2.59 (SD, 1.89). The mean nasality ratings
for the LP speech ranged from 0 to 5.86, and the mean nasality
ratings for the HP speech ranged from 0 to 4.86. A paired t
test revealed no statistically significant difference in mean nasality ratings for the two stimuli (t 5 1.57, df 5 24, p 5 .13).

296

Cleft PalateCraniofacial Journal, July 1998, Vol. 35 No. 4

FIGURE 1 The relationship between nasalance scores and nasality ratings for low pressure (LP) speech.

Relationship Between Nasalance and Nasality

For the LP speech, the correlation coefficient between nasalance scores and nasality ratings was r 5 0.77. A scatterplot
showing the relationship between nasalance scores and nasality
ratings for the LP speech is shown in Figure 1. For the HP
speech, the correlation coefficient between nasalance and nasality was r 5 0.78. Figure 2 shows the relationship between
nasalance and nasality for the HP speech.
Sensitivity and Specificity
Normal nasalance was established by comparing the nasalance scores to the nasality ratings according to the method
illustrated in Table 2 to determine the cutoff point that provided the best overall sensitivity and specificity. The analysis
yielded a nasalance cutoff of 26%.
Given the finding of no significant differences between HP
and LP stimuli, separate tests of sensitivity and specificity for
HP and LP stimuli were not warranted, and HP and LP data
were pooled for overall analyses of sensitivity and specificity.
As Table 2 shows, 25 of the speech samples were rated hypernasal by the listeners and 25 were rated normal. Of the 25
speech samples rated hypernasal, 21 received a nasalance score
greater than 26%. This results in a sensitivity of .84 (21/25).
In other words, Nasometer scores corresponded with the listener ratings 84% of the time when the subject was judged to
be hypernasal. Of the 25 speech samples rated normal, 22 received a nasalance score less than 26%. This results in a specificity of .88 (22/25). That is, Nasometer scores corresponded
to listener ratings 88% of the time when the subject was judged
to have normal nasality. Table 2 also shows the overall efficiency of the Nasometer. Overall efficiency is the sum of the
number of times the Nasometer agreed with the listeners divided by the total number of opportunities. The overall efficiency in this case was .86 (43/50).
DISCUSSION
The results of this study showed that there was no significant
difference in nasalance scores between LP and HP speech, and

FIGURE 2 The relationship between nasalance scores and nasality ratings for high pressure (HP) speech.

no significant difference in nasality ratings. In other words,

use of a stimulus composed of only glides, liquids, and vowels
yielded essentially the same nasalance scores and the same
nasality ratings as a stimulus loaded with pressure consonants.
Further, the correlation coefficients between nasalance scores
and nasality ratings for the two different stimuli were nearly
identical. The correlation coefficient between nasalance and
nasality for the LP speech was r 5 0.77, and for the HP speech
r 5 0.78. Thus, the relationship between nasalance scores and
nasality ratings appears to be determined more by the vowels
in the speech sample than by the consonants. The only apparent requirement for controlling consonant environment when
assessing hypernasality with the Nasometer is that the stimulus
should not contain nasal consonants (Watterson et al., 1996).
One caveat to this conclusion is that the patient-subjects for
this study were selected only on the basis of visible nasal air
emission (i.e., mirror fogging), and no particular effort was
made to determine if the nasal air emission was audible. If we
were to study only subjects with audible nasal air emission, it
might be that nasalance scores for HP stimuli would be elevated relative to LP stimuli and to listener judgments of nasality. Nevertheless, the population we studied was representative of patients who are likely to be evaluated with a Nasometer, and the data showed that HP and LP consonants were
not a factor in nasalance scores and that listener judgments
agreed with the nasalance scores.
Karnell (1995) also compared nasalance scores in LP and
HP speech, although his experimental design did not include
listener ratings of nasality. It is difficult, however, to compare
the present findings for nasalance to those of Karnell, because
TABLE 2 Sensitivity and Specificity Measures for the Pooled LP and
HP Speech
Listener Ratings of Nasality

% Nasalance
$26 (hypernasal)
,26 (normal)
Total

Hypernasal

Normal

.2.0

#2.0

Total

21
4
25
Sensitivity
21/25 5 .84

3
22
25
Specificity
22/25 5 .88

24
26
50
Overall
43/50 5 .86

Waterson et al., NASALANCE AND NASALITY

Karnell did not report an overall comparison of his LP speech

to his HP speech. Rather, using a nasalance cutoff score of
31%, he first divided the nasalance scores into three groups
and then analyzed the three groups separately. The three
groups included those subjects having an LP and an HP nasalance score above the cutoff (31%), those having both nasalance scores below the cutoff, and those having mixed results. The statistical analysis showed that there was no significant difference between the LP and HP speech for subjects
with both nasalance scores above the cutoff, but there was a
significant difference between the LP and HP speech for subjects with both nasalance scores below the cutoff. The difference in mean nasalance for this latter group was quite small,
however, and while statistically significant, may lack practical
significance.
In his study of LP and HP stimuli, Karnell (1995) focused
his attention on the subjects with mixed results, that is, subjects whose nasalance score for one stimulus was on one side
of the cutoff and whose nasalance score for the other stimulus
was on the other side of the cutoff. Karnell hypothesized that
these subjects represent a subgroup with marginal velopharyngeal function. In the present study, only four subjects fit
into this category. Three subjects had an HP score above the
cutoff with an LP score below the cutoff, and one subject had
an LP score above the cutoff with an HP score below the
cutoff. With regard to nasality ratings for this group, only one
of the four subjects also had mixed nasality ratings. In other
words, while one nasalance score was above the cutoff and
one was below the cutoff for each of these four subjects, in
three out of the four, the nasality ratings remained consistent
on either one side of the cutoff or the other. This may be
suggesting that listeners overall, general impression of a patients nasality is relatively stable and not sensitive to minor
fluctuations in resonance that are measurable with the Nasometer.
Sensitivity and specificity measures indicated good correspondence between nasalance scores and nasality ratings. Sensitivity was .84 and specificity was .88. In other words, when
a subject was judged to be hypernasal, Nasometer scores
agreed 84% of the time (sensitivity), and when a subject was
judged to have normal resonance, Nasometer scores agreed
88% of the time (specificity). As previously explained, all but
one of the disagreements between the Nasometer scores and
listener ratings occurred when the two nasalance scores for a
given subject were on either side of the nasalance cutoff, while
the nasality ratings remained consistent on one side of the
cutoff. This finding tends to support our previous suggestion
that a nasalance score may be used to augment human perceptual impression, but not replace it (Watterson et al., 1996).
In that regard, it is of interest that the sensitivity score (.84)
obtained in the present study is slightly higher than the sensitivity scores we have reported in earlier research. In our previous studies, we have reported sensitivity scores of .71, .78,
and .77 (Hashimoto et al., 1992; Watterson et al., 1993, 1996).
The improved sensitivity may be explained, at least in part, by
the different experimental method used in this study to obtain

297

the nasality ratings. In our previous studies, and in studies at

other centers, listeners were not necessarily informed what
would be the cutoff between normal nasality and hypernasality
(Dalston et al., 1991; Dalston and Seaver, 1992; Hardin et al.,
1992; Nellis et al., 1992; Dalston et al., 1993; Watterson et al.,
1993, 1996). For example, in one study, a four-point rating
scale was used where the listeners were informed that 1
represented normal resonance and 4 represented severe
hypernasality. A post hoc decision was made that ratings of
1 or 2 would represent normal nasality and ratings of
3 or 4 would represent clinically important nasality
(Dalston et al., 1993). In the present study, however, during
listener training and in advance of data collection, listeners
were informed that ratings of 0, 1, or 2 would represent different degrees of normal nasality and ratings of
31 would represent different degrees of clinically hypernasal speech. Thus, when assigning a rating number to a
speech sample, a listener did not have to wonder if a rating of
2 was the lowest degree of hypernasality or the highest
degree of normal nasality. We believe this method served to
improve the relationship between our nasalance scores and the
ratings of nasality.
Clinicians may wonder what cutoff should be used to separate normal from abnormal nasalance scores. The optimum
cutoff for nasalance scores in this study was determined to be
26%. This cutoff is identical to the cutoff obtained by Watterson et al. (1993) and by Hardin et al. (1992) and is similar to
the 25% cutoff recommended by Fletcher et al. (1989). Other
studies have reported slightly higher or slightly lower cutoffs.
Dalston et al. (1991) found the best cutoff for the Zoo Passage
in their study was 32%, while Watterson et al. (1996) reported
a cutoff of 22% for the Zoo Passage. However, it is difficult
to compare cutoff scores across studies because of the different
experimental methods that have been used to derive the cutoff.
For example, the cutoffs reported by Fletcher et al. (1989) and
by Watterson et al. (1996) were obtained by testing a control
population of normal subjects and calculating the cutoff as two
standard deviations from the mean. But, in the other studies,
cutoffs were determined by comparing nasalance measures and
nasality judgments post hoc in a patient population and then
identifying the nasalance score that maximized the sensitivity
and specificity scores. In other words, the difference between
normal and abnormal nasalance was not known until the nasalance scores were compared to the nasality ratings after data
collection. Considering this difference across studies in determining nasalance cutoffs, and the previously explained differences in obtaining nasality judgments, we should be surprised
that the reported nasalance cutoffs are so similar rather than
so different. Geographic differences may also play a role in
the difference between normal and abnormal nasalance, but
we are not convinced of that possibility. At any rate, there is
presently no universally agreed upon nasalance cutoff, and
there will probably never be a single cutoff that satisfies all
circumstances. It is for these reasons that we continue to recommend that nasalance scores be used only as one piece of
information that should be considered when evaluating speech

298

Cleft PalateCraniofacial Journal, July 1998, Vol. 35 No. 4

and resonance, and not as an absolute indicator of hypernasality or velopharyngeal function.

REFERENCES
Barker DJP, Rose G. Epidemiology in Medical Practice. London: Churchill
Livingstone; 1979:3435.
Carney PJ, Sherman D. Severity of nasality in three selected speech tasks. J
Speech Hear Res 1971;14:396407.
Dalston RM, Neiman GS, Gonzalez-Landa G. Nasometric sensitivity and specificity: a cross-dialect and cross-culture study. Cleft Palate J 1993;30:285
291.
Dalston RM, Seaver EJ. Relative values of various standardized passages in
the Nasometric assessment of patients with velopharyngeal impairments.
Cleft Palate Craniofac J 1992;29:1721.
Dalston RM, Warren DW, Dalston ET. Use of nasometry as a diagnostic tool
for identifying patients with velopharyngeal impairment. Cleft Palate J
1991;28:184187.
Fletcher SG, Adams LE, McCutcheon JJ. Cleft palate speech assessment
through oral-nasal acoustic measures. In: Bzoch KR, ed. Communicative
Disorders Related to Cleft Lip and Palate. Boston: Little, Brown; 1989:
246257.
Hardin MA, Van Demark DR, Morris HL, Payne MM. Correspondence between
nasalance scores and listener judgements of hypernasality. Cleft Palate J
1992;29:346351.
Hashimoto AS, Watterson T, Paynter ET. The relationship between articulation,
nasalance, and nasality. Presented at the Annual Meeting of the American
Cleft Palate-Craniofacial Association; 1992; Portland, Oregon.
Karnell MP. Nasometric discrimination of hypernasality and turbulent nasal
airflow. Cleft Palate J 1995;32:145148.
Morrant WH, Shaw WC. Use of standardized video recordings to assess cleft
surgery outcome. Cleft Palate Craniofac J 1996;33:134142.
Nellis JL, Nieman GS, Lehman JA. Comparison of nasometer and listener
judgments of nasality in the assessment of velopharyngeal function after
pharyngeal flap surgery. Cleft Palate J 1992;29:157163.
Spriestersbach DC. Assessing nasal quality in cleft palate speech of children.
J Speech Hear Disord 1955;20:266270.
Spriestersbach DC, Powers GR. Nasality in isolated vowels and connected
speech of cleft palate speakers. J Speech Hear Res 1959;2:4045.

Statistical Package for the Social Sciences. SPSS-X Users Guide. 3rd ed. Chicago: SPSS; 1988.
Watterson T, Hinton J, McFarlane SC. Novel stimuli for obtaining naslance
measures from young children. Cleft Palate J 1996;33:6773.
Watterson T, Wright DS, McFarlane SC. The relationship between nasalance
and nasality in children with cleft palate. J Commun Disord 1993;26:13
28.

APPENDIX
Speech Stimuli
Low Pressure Stimulus (LP)1
We were away.
Where were you?
Why were you away?
You were away earlier.
We were really low.
We were away all year.
You were well.
Will you wear a lilly?
Roll a yellow wheel.
High Pressure Stimulus (HP)2
Look at this book with us.
Its a story about a zoo.
That is where bears go.
Today its very cold outdoors.
But we see a cloud overhead.
Thats a pretty white fluffy shape.

1
2

Karnell, 1995.
Fletcher et al., 1989.