University of Montana University of Montana
ScholarWorks at University of Montana ScholarWorks at University of Montana
Graduate Student Theses, Dissertations, &
Professional Papers
Graduate School
1990
Response errors by normal hearing listeners on the Maryland CNC Response errors by normal hearing listeners on the Maryland CNC
Test Test
Sherrin F. Richardson
The University of Montana
Follow this and additional works at: https://scholarworks.umt.edu/etd
Let us know how access to this document bene5ts you.
Recommended Citation Recommended Citation
Richardson, Sherrin F., "Response errors by normal hearing listeners on the Maryland CNC Test" (1990).
Graduate Student Theses, Dissertations, & Professional Papers
. 7789.
https://scholarworks.umt.edu/etd/7789
This Thesis is brought to you for free and open access by the Graduate School at ScholarWorks at University of
Montana. It has been accepted for inclusion in Graduate Student Theses, Dissertations, & Professional Papers by an
authorized administrator of ScholarWorks at University of Montana. For more information, please contact
'"Ill
$
Mike and Maureen
MANSFIELD LIBRARY
Copying allowed as provided under provisions
of the Fair Use Section of the U.S.
COPYRIGHT LAW, 1976.
Any copying for commercial purposes
or financial gain may be undertaken only
with the author’s written consent.
University of
Montana
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Response Errors by Normal Hearing Listeners
on The Maryland CNC Test
by
Sherrin F. Richardson
B.Sc., University of Alberta, Edmonton, Canada, 1987
Presented in partial fulfillment of the requirements
for the degree of
Master of Arts
University of M ontana
1990
Approved by:
Chaim/aA, Hdlard of Examiners
TÎean, G raduate School
Hr, i ^ U
Date /
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: EP38590
All rights reserved
INFO RM ATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMT
Dis«artation Publishcog
UMI EP38590
Published by ProOuest LLO (2013). Copyright in the Dissertation held by the Author.
Microform Edition © ProOuest LLO.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQ^st’
ProOuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106 -1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Richardson, Sherrin F., M A. Communication Sciences and Disorders
August 1990
Response Errors b y N ormal H earing Listeners on The M aryland CNC Test
(42 pp.)
Director: Sally J. Johnson, M.Af
A carrier phrase is often used in word recognition tests to provide perceptual
cues to the listener to aid in the selection of a target item. One such test of w ord
recognition. The Maryland CNC Test (Causey, Hood, Herm anson and Bowling, 1984)
employs the carrier phrase "Say th e again." Informal clinical observation has
indicated that m any clients will often commit errors on stimulus items that are not
nouns and thus do not maintain syntax of the carrier phrase w hen tested using The
Maryland CNC Test word lists.
The purpose of the study was to investigate the types of errors m ade by
norm al hearing listeners on The Maryland CNC Test. Given that the linguistic
environm ent influences an individual's speech perception, then the errors m ade
during w ord recognition testing may be, at least partially, determined by the syntactic
structure of the carrier phrase. The following hypotheses w ere proposed: (1) the
stim ulus items that fall within the grammatical class of noun will be responded to
w ith significantly fewer errors than those items that fall outside of that grammatical
class; and (2) the grammatical class of the incorrect responses will not be significantly
associated w ith the grammatical class of the stimulus item.
To test the hypotheses, thirty-tw o normal hearing subjects listened to word
lists presented at 10, 20, and 30 dB SL (re: SRT). Their incorrect responses were
categorized as nouns or non-nouns. The results indicated that stim ulus items w hich
could be recognized as nouns were identihed w ith significantly few er errors than
those items which were not nouns. The results for the second hypothesis indicated a
significant b ut veiy weak association between the grammatical class of the incorrect
response and that of the stimulus item. The implication of these results on word
recognition testing and future research were discussed.
II
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGEMENTS
The journey of m y M aster of Arts degree w ould not have been possible
w ithout the encouragement of m any people. I wish to extend m y thanks to the
mem bers of m y thesis committee: Ms. Sally Johnson, Dr. Barbara Bain, Dr. Michael
W ynne, and Dr. W esley Shell en. Their w isdom , guidance, and support is deeply
appreciated. I would also like to thank Ms. Tami McKinley for the expert production
of this thesis. A special thank you to the subjects who volunteered for this study and
so w illingly gave their valuable time.
A special thank you to m y thoughtful classmates, w ho give so freely of
themselves. We have shared laughter and frustration and I cherish our friendship.
Thank you to m y family and friends for their faith and encouragement. I am
grateful to m y family for their help and steadfast support. A special thank you to
Joanne O 'G orm an w ho recognized m y abilities so long ago and encouraged me to
pursue this field. Thank you also to Paul Lynn for his kindness and generosity and
for sharing the beauty of M ontana with me.
W ords cannot express my gratitude to Charles and Alvera Valach for
"adopting" m e into their family and giving m e the joy of having parents again.
Thank you to Keith Bonnah for his understanding, patience, hum or, and the promise
of such an exciting future.
Finally, the completion of m y degree would not have been possible without
the love, unw avering faith, and generosity of my sister, Laurie Ferrari. It is w ith her
that I share m y degree. This thesis is dedicated to the m em ory of m y late parents,
Elm er and Jennie Hushagen.
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
A b stract
...............................................................................................................................
ii
A cknow ledgem ents
.........................................................................................................
iii
Chapter I: Introduction and Literature Review
.....................................................
1
Chapter II: M eth o d s
.....................................................................................................
12
Chapter III: R e sults........................................................................................................ 16
Chapter IV: D iscussion
..................................................................................................
24
References
..........................................................................................................................
32
Appendix A: Articulation Screening T es t
...................................................................
34
Appendix B: Subject Instruction and Practice S tim u li
................................................
35
Appendix C: The Maryland CNC Speech Intellieibilitv Tests
- W ord L is ts
.............................................................................................
36
IV
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table 1: Grammatical Class of Stimulus Item s on The Maryland CNC Test
Presented Per L ist
......................................................................................................
17
Table 2: Total Stimulus Items (61 Ears) on The Maryland CNC Test
Presented Per L is t
......................................................................................................
18
Table 3: Incorrect Responses to Stimulus Items on The M aryland CNC T e s t 20
Table 4: Two W ay Repeated Analysis of Variance of the Proportion of
Error for Grammatical Categories and E a r
............................................................
21
Table 5; Chi-Square Analysis of the Association Between Grammatical
Category of the Stimulus Items and the Incorrect R esponse
.............................
22
Table 6: Stimulus Items that Elicited Incorrect Responses in More Than
Fifty Percent of the Subjects
.......................................................................................
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER I: INTRODUCTION A N D LITERATURE REVIEW
W ord recognition testing has been an integral part of hearing assessm ent since
the beginnings of audiology. The ability to understand speech has been recognized as
a prerequisite for coping w ith the rigors of daily living in a "complex auditory world"
(Penrod, 1985). Evaluation of word recognition has undergone a m etam orphosis over
the years. The developm ent and modification of word recognition tests have brought
to light factors which can affect a listener's w ord recognition score.
One factor affecting w ord recognition is the use of a carrier phrase to facilitate
word recognition. The rationale for using a carrier phrase is that it provides
perceptual cues to the listener to aid in the selection of the target item. The
contextual environm ent of the carrier phrase m ay provide syntactic, semantic or
acoustic-phonetic cues which influence selection of a target item. The present study
investigated the influence of a carrier phrase on the identification of target items.
Clinical Application of W ord Recognition Testing
According to Penrod (1985), establishing an individual's pure tone thresholds
provides information regarding hearing sensitivity, but fails to provide sufficient
information about an individual's receptive auditory communication ability. While a
clearly defined relationship exists betw een an individual's pure tone thresholds and
speech recognition threshold, the relationship betw een the pure tone threshold and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
w ord recognition ability tends to be highly variable. This variable relationship has
led some researchers to question the clinical applicability of word recognition testing.
W ord recognition testing has had w ide clinical application. Examples of
application include the assessment of auditory function; identification of site of lesion
testing; determination of candidacy for surgery; evaluation of central auditoiy
disorders; and determ ination of social adequacy of communication. W ord
recognition testing has also been used in the selection of hearing aids and in the
development of rehabilitation and assessment of its effectiveness (Penrod, 1985).
Hayes (1984) argued that w ord recognition testing has had little impact on the
identification of the presence of hearing im pairment and localization of the site of
auditory disorders. She advocated that the goal of word recognition testing lies in the
im provem ent of techniques for the rehabilitation of patients with presbycusis;
prediction of successful rehabilitation of profoundly hearing-im paired individuals;
and selection of amplification devices for mildly to severely hearing-impaired
individuals.
Thornton (1985) agreed w ith Hayes' proposition. He stated that w ord
recognition plays a m inor role in the diagnosis of disease and is ineffective for site of
lesion determ ination. However, he thought that it plays an im portant role in
predicting the need for the rehabilitation of a hearing im pairm ent and in the
assessm ent of the effects of any rehabilitation programs. According to Thornton,
although the pure tone audiogram contributes more and better inform ation for
hearing aid selection, w ord recognition testing also provides some beneficial
information for the hearing aid user. Therefore, the strength of word recognition
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
testing lies in the assessment of a hearing handicap and in the prediction of
rehabilitation benefits.
Apparently some controversy exists regarding the efficacy of w ord recognition
testing for the diagnosis of disease and the localization of site of lesions. However,
there seems to be agreem ent that word recognition testing has definite utility in the
realm of rehabilitation. As audiologists strive to restore the ability of
hearing-im paired individuals to understand speech, they should continue to develop
and use m ethods that can predict an individual's word recognition ability.
Factors Affecting W ord Recognition Scores
Miller, Heise and Lichten (1951) and Penrod (1985) have discussed the various
factors w hich influence w ord recognition scores. The three broad categories discussed
are physical factors, test administration factors, and linguistic factors.
Physical factors include equipment, the physical environm ent used for testing,
and the test stimulus. The test stim ulus is influenced by the level of presentation,
distortion, frequency composition, duration, and signal-to-noise ratio. The second
category of test administration factors includes the m anner and rate of presentation,
stim ulus m aterials, and scoring method. This category also includes the personnel
involved in the testing. Speaker and listener performance m ay be affected by
variables such as age, fatigue or intelligence, for example. The third category,
linguistic factors, constitutes a wide range of variables. These include articulation and
dialect of the speaker and listener, and familiarity, redundancy and context of the test
items.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
Context is often a key factor in w ord recognition. Miller et al. (1951) described
three means of providing context to the listener. It can be provided by the
know ledge that the test item meets one or more of the following conditions: (a) it is a
repetition of a preceding w ord; (b) it is one of a limited num ber of words; and (c) it is
preceded by w ords in a phrase or sentence. The last condition is often m et in word
recognition testing by supplying a carrier phrase such as "Say the w o r d ."
Effects of the Carrier Phrase
Several researchers disagree as to the effectiveness of the carrier phrase in
facilitating w ord recognition. Martin, Hawkins and Bailey (1962) "noted that the
carrier phrase seems to confuse some patients, especially those with severe
discrimination problems" (p.319). The researchers investigated the effect of a carrier
phrase on phonetically balanced (PB) word recognition scores and subject preference
for its exclusion. Of the 75 subjects, 30 had sensorineural hearing loss, 30 had
conductive loss and the rem ainder had normal hearing. Each subject listened to the
PB w ords in isolation and w ith the carrier phrase "Say the w o rd ." They were
then asked if they preferred hearing the words in isolation or w ith the carrier phrase.
Most subjects preferred exclusion of the carrier phrase. The results indicated that
recognition scores were not affected b y the presence or absence of a carrier phrase
w ith the phonetically balanced (PB) words.
Kreul, Bell and Nixon (1969) studied the influence of the carrier phrase and
accom panying levels of noise on the word recognition scores of normal hearing
subjects. They used two carrier phrases, "You will strike th ro u g h now" and
"You will strik e please." Each carrier phrase was recorded by two different
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
speakers. The carrier phrases were presented to three groups of listeners at varying
signal-to-noise ratios. The resulting confusion matrices indicated that the rank order
of test item difficulty was unaffected by the carrier phrase, speaker or signal-to-noise
ratio used. However, the findings revealed significant differences in the num ber of
errors found at each signal-to-noise ratio betw een the carrier phrases, regardless of
the speaker. This study found carrier phrases can affect the w ord recognition scores.
How ever, the phrases used in the study are not commonly em ployed in typical
clinical audiology settings (Lynn and Brotman, 1981) and m ay not have clinical
impact.
Gladstone and Siegenthaler (1971) explored differences in word recognition
scores in norm al hearing subjects as a function of three commonly used carrier
phrases: "Say the w o r d "You will s a y ;" and "Point to th e ;" and no
carrier phrase. The interaction condition allowed for the opportunity of phonem ic
interaction betw een the carrier phrase and the test item. In the no interaction
condition, carrier phrases and test items w ere recorded separately and then spliced
together. The results indicated that recognition scores improved when a carrier
phrase w as used In the interaction condition. Furthermore, significant differences
w ere found between the carrier phrases in this condition. Their results suggested that
carrier phrases can affect word recognition performance and, specifically, the
phonem ic interaction betw een the carrier phrase and the test item affects the word
recognition score. However, this finding is not universal. In contrast to the findings
of Gladstone and Siegenthaler (1971), M artin et al. (1962) concluded that carrier
phrases w ere not essential based on the word recognition scores of the carrier phrase
"Say the w o r d ."
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
Gelfand (1975) studied the effects of a carrier phrase on subjects w ith
sensorineural hearing loss using m onitored live voice w ord recognition testing. Each
subject listened to CID W-22 word lists w ith and w ithout the carrier phrase "Say the
w o rd The results revealed significantly higher word recognition scores using
a carrier phrase. The subjects were questioned regarding the clarity of the test items
and preference for the inclusion of a carrier phrase. Of the 22 subjects questioned, 10
reported no difference in clarity w ith the carrier phrase, 10 reported that it increased
clarity and two reported greater clarity w ithout the phrase. Eleven subjects expressed
preference for inclusion of the carrier phrase, three preferred exclusion and eight had
no preference. Gelfand found that subjective clarity and preference for a carrier
phrase were not significantly related to the word recognition scores of the subjects.
These findings conflict w ith those found by Martin et al. (1962) in subject preference
and performance. In response to M artin et al. (1962), Gelfand concluded that subject
preference is a poor argum ent for the exclusion of a carrier phrase.
A m ore recent study by Lynn and Brotman (1981) investigated norm al hearing
subjects' ability to identify the place of articulation for the initial stop consonant / p / ,
/ 1/, and / k / . They found that the carrier phrase "You will s a y in
com parison to w ords in isolation, provided m ore perceptual cues for the correct
identification of target phonem es, again stressing the influence of a carrier phrase.
Although the researchers did not attem pt to generalize their findings to other
carrier phrases, they hypothesized that different carrier phrases would produce
different scores based on the acoustic properties of the carrier phrase. Carrier phrases
such as "Say the w o rd " and "You will s a y " approach different vowel
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
consonant format transitions during the last centisecond of the phrase. Thus, they
m ay provide different perceptual cues.
Lynn and Brotman's hypothesis is supported by the results of Gladstone and
Siegenthaler (1971), who reported the highest recognition scores for the carrier phrase
"You will s a y " compared to "Say the w o r d " or "Point to th e They
thought that "You will s a y " had a greater potential for phonem ic interaction
due to the final diphthong / e i/. The carrier phrase "Point to th e " provides the
same syllabic nucleus for the vowel consonant transition pattern as "You will say
," but possibly has a shorter vowel duration which m ay affect phonem ic
interaction. Lynn and Brotman stated that the vowel consonant form ant transitions
during the last phases of "Say the w o r d " approach the alveolar place of
articulation. Possibly this transition does not allow for as much phonem ic interaction
as those of other carrier phrases. Thus there is good evidence to suggest the acoustic-
phonetic influence of a carrier phrase.
A study by C raig (1988) investigated the interactions betw een sentence context
and w ord predictability on w ord recognition scores. W ord recognition scores w ere
obtained for normal hearing listeners und er conditions of high semantic predictability
sentences, low semantic predictability sentences and a sem antically neutral
predictability carrier phrase. The high predictability sentences w ere em bedded w ith
semantic cues that preceded the target words. The low predictability sentences did
not provide sem antic cues, although they were meaningful. The neutral predictability
carrier phrase used w as "I will now say the word th e ." The results indicated
significant differences in word recognition scores betw een the high predictability
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
sentences and the low predictability sentences or the neutral carrier phrase. The
implications of syntactic effects of the carrier phrase will be discussed later.
According to Craig, other researchers have determ ined that sentences have
high predictability w hen listeners can use both "acoustic-phonetic” (p.588) and
"linguistic-situational" (p.588) inform ation for speech recognition. Craig proposed that
having the listener identify a target w ord at the end of a low predictability sentence
and at the end of a carrier phrase m ay not be synonym ous tasks. He thought that
carrier phrases tend to leave the listener free to attend to the acoustic-phonetic cues of
speech, w hereas low predictability sentences tend to dem and attention to the semantic
content. However, any differences in word recognition scores betw een the carrier
phrase and the low predictability sentences in his study w ere not significant.
W ord recognition of a target w ord appears to be influenced by the semantic,
syntactic and acoustic-phonetic context. The studies described above have
dem onstrated that the context of the carrier phrase can influence word recognition
scores. Furthermore, these scores m ay be sensitive to the unique set of cues provided
by each carrier phrase.
Finally, Craig proposed that the position of the target word is predictable in a
carrier phrase and thus frees the listener to ignore semantic content and focus on
acoustic-phonetic cues. However, the effects of syntax were not addressed. A carrier
phrase m ay possibly im pose syntactic constraints on a listener.
W ord Recognition and Linguistic Rules
M iller and Isard (1963) conducted several experiments on the auditory
perception of grammatical, anomalous and ungramm atical sentences. Given that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
speech perception Is facilitated by linguistic rules. Miller and Isard hypothesized that
sentences which violated semantic and syntactic rules w ould be most difficult to
repeat, w hereas those which obeyed the rules would be the easiest These researchers
constructed and tape recorded sentences that were gram matical, anomalous and
ungramm atical. Their subjects were required to listen to the sentences and
imm ediately repeat w hat was said. The results indicated that both grammatically
anomalous and ungram m atical sentences were indeed m ore difficult to repeat than
the grammatical sentences. These findings became more dram atic w hen the
signal-to-noise ratios were reduced. That is, they found that the grammatical
sentences w ere more resistant to noise than the ungramm atical sentences. Miller and
Isard's (1963) findings dem onstrated that linguistic rules can affect w ord recognition
at least at a sentence level.
Carrier Phrases and Phrase Structure Rules
Several tests of w ord recognition em ploy a carrier phrase. The M aryland CNC
Test (Causey, Hood, Herm anson and Bowling, 1984) uses the carrier phrase "Say the
again" in an attem pt to provide a phonetically neutral context for the test word.
Causey and his colleagues proposed that the vowel / / in the w ord final position of
"the" and the w ord initial position of "again" w ould m inim ize the effects of
coarticulation. In the effort to minim ize coarticulation, a syntactic constraint was
placed on the carrier phrase. This becomes evident w hen syntactic theory as
described by Chom sky (1965) is applied to the carrier phrase. The phrase structure
rules require that a noun phrase consist either of a noun itself or a determ iner and a
noun. The carrier phrase "Say th e again" employs the determ iner "the."
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
Therefore, the phrase structure rules dem and that the test item be a noun to m aintain
syntactic structure and ultimately, grammaticality.
Two previously discussed studies investigated carrier phrases that also
employed the determ iner "the." Craig (1988) used the carrier phrase "I will now say
the w ord t h e
_____________
The grammatical categories of the target words w ere not
specified, however the author stated that "The list selection included consideration of
the nature of the word im mediately preceding each target word" (p.30), which m ay
have m aintained syntactic neutrality, b u t it is not clear from the discussion.
Gladstone and Siegenthaler (1971) used the carrier phrase "Point to th e "
and the first twenty-five words from the CID Test W-22. List 3B as target items. The
target items included w ords w hich fall outside of the grammatical classification of
nouns. The results indicated significant differences in word recognition between the
above m entioned carrier phrase and the carrier phrases "Say the w o r d " and
"You will s a y " in the phonemic interaction condition. The phrase "Point to the
" w as the least enhancing of the carrier phrases in the interaction condition.
These results m ay have been due to the semantic nature of the phrase (the subjects
w ere giving verbal, not pointing responses) a n d /o r the syntactic constraints of placing
w ords other than nouns after the determ iner "the." There is enough evidence from
this study to at least question the use of a carrier phrase placing the determ iner "the"
before stimulus items w hich m ay not be nouns.
Informal clinical observation has indicated that m any clients will commit errors
on those target items that are not nouns when tested using The Maryland CNC Test
w ord lists. In m any cases, the clients appear to have processed the target item in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
order to fit the syntactic structure of the carrier phrase. For example, w hen the target
w ord is "sung," it is often identified by clients as "sun" or "song."
The M aryland CNC Test was of particular interest because of its wide use by
the Veterans Administration (VA). The Veterans Administration has m andated that
The M aryland CNC Test be used for w ord recognition testing during all Assessments
of Social Efficiency, VA examinations for determ ining disability of hearing loss.
Purpose of the Study
The purpose of the study was to investigate the types of errors m ade by
norm al hearing listeners on The M aryland CNC Test. Given that the linguistic
environm ent influences an individual's speech perception (Craig, 1988; Gladstone and
Siegenthaler, 1971; Miller, 1951; Miller and Isard, 1963), then the errors m ade during
word recognition testing m ay be, at least partially, determined by the syntactic
structure of the carrier phrase. Specifically, the following hypotheses w ere proposed:
1. The stim ulus items that fall w ithin the grammatical class of noun will be
responded to w ith significantly fewer errors than those items that fall
outside of the gram matical class.
2. The grammatical class of the incorrect responses will not be significantly
associated w ith the grammatical class of the stim ulus item.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER II: METHODS
Subjects
The rationale for subject selection and test procedures was based on the
normative studies of The M aryland CNC Test as described by Causey, Hood,
H erm anson and Bowling (1984). Thirty-two wom en between the ages of 18 and 26
years participated in this study. All subjects spoke English as their prim ary language
and had negative histories of otologic surgeries. All subjects had audiom etric
thresholds equal to or better than 15 dB HL (ANSI, 1969) in each ear for pure tone
stimuli presented at octave intervals between 250 and 8000 Hz. Otoscopic inspection
revealed relatively clear ear canals bilaterally. Each subject also had normal
oto-imm ittance results, as defined b y Jerger Type A tympanogram s and a measurable
acoustic reflex threshold obtained for ipsilateral stim ulation at or below 100 dB nHL
at 1000 H z for each ear.
Instrum entation
Pure tone screening, speech reception threshold, and speech recognition testing
were conducted in an lAC sound treated suite (model num ber 1400 ACT), which m eet
the ANSI 1977 standards for noise level. All testing was performed w ith a
Grason-Stadler 10 two-channel audiom eter w ith TDH-50 earphones and MX41/AR
cushions. The audiom eter was calibrated quarterly to ANSI 1969 and 1981 standards.
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
In addition, the audiom eter was given a biological calibration check daily. An
Amplaid 702 impedance bridge was used for acoustic im mittance screening. This
system w as calibrated electroacoustically prior to the data collection.
The M aryland Speech Intelligibility Test (Causey and Elkins, 1981) was used in
this study. Specifically, CNC tape recordings of lists 1-1, 1-3, 1-6, 1-7,1-9, and 1-10
were obtained from Olsen Distributors and used for this study. These particular lists
were found to have inter-list equivalencies (Causey et al., 1984). The tapes were
presented to subjects by a Fisher cassette tape recorder fed through the
Grason-Stadler 10 audiom eter and TDH-50 earphones.
The stim ulus items and the subjects' responses w ere recorded using a Pioneer
Stereo Amplifier A-5 and a Nakimichi BX-lOO tape recorder using an LTD wireless
m icrophone system. The m icrophone was attached approximately four inches from
the subject's mouth. Maxell XL II cassette tapes were used for storage of the subjects'
responses.
Procedures
Subjects w ere informally screened for clarity of articulation using a 23-word
test of articulation (Bzoch, 1989) w hich is presented in A ppendix A. They were
scored for the entire test on a subjective pass/fail basis.
H earing was screened using pure tone stimuli presented at 15 dB HL to each
ear for each subject across octave frequencies from 250 to 8000 Hz. A speech
recognition threshold w as established for each ear, following the Guidelines for
D eterm ining the Threshold Levels of Speech (ASHA, 1987) using monitored live
voice.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
After the speech recognition threshold w as established and prior to the
presentation of the word recognition test stimuli, each subject was familiarized w ith
the task. A conventional set of instructions and four practice stimuli w ere presented
(A ppendix B). The practice stimuli w ere presented by m onitored live voice at 60 dB
HL and w ere not included am ong the test items.
After familiarization with the task, a different w ord list (Appendix C) was
presented at each of three presentation levels (10, 20, 30 dB SL re: SRT). These
presentation levels w ere chosen because they are near the subject's hearing threshold
and therefore most errors would be anticipated at these levels. The results of the
norm ative study (Causey et al., 1984) indicated that correct responses asym ptote at
approxim ately 45 dB HL. The presentation level and w ord lists w ere randomly
assigned w ithin the constraint of not repeating a list for a given subject and ear
presentation was counterbalanced.
The subjects' verbal responses were recorded and scored on-line by the
examiner. The responses w ere also recorded on audiotape for inter-rater reliability
purposes. Of the 61 ears tested and recorded, six were random ly selected and scored
b y a second judge, w ho w as familiar w ith The M aryland CNC Test and the purpose
of the study. Reliability was scored using a point-by-point formula of num ber
agreem ents minus num ber of disagreem ents divided by the total num ber squared.
An inter-judge reliability of 90 percent or greater w as desired.
Data Analyses
Each stim ulus item used was categorized into grammatical class by three
graduate students in communication sciences and disorders. W ords w ere categorized
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
as: 1) noun or both (NB) for words that are nouns in at least one usage including
those nouns which could also be placed in another grammatical category (e.g., walk);
or 2) other than a noun (Other) for those words which fall outside of the noun
category (e.g., happy). For the purpose of this study, the stim ulus items w ere then
categorized as NB or Other based on the judgem ent of at least two of the three raters.
If no consensus w as reached, then a standard English dictionary classification
(U rdang and Flexner, 1968) was employed to make the category judgem ent.
The subjects' responses w ere recorded and scored as correct if they matched
the stim ulus item and incorrect if they did not. The subjects' incorrect responses were
also placed into a grammatical category as described above with one exception.
Responses were placed into a third unidentified category (UN) if a subject's response
was unintelligible, if a subject failed to respond to a stimulus item, or if the response
w as a nonsense w ord as determined by the examiner.
The presentation level was not of interest in this study, so the data obtained
across the three levels w as collapsed for each subject. A proportion of error was
calculated for each subject for the stimulus items in each ear condition and each
gram m atical class. A tw o-w ay repeated analysis of variance (Bruning and Kintz,
1977) w as performed (confidence level of .05). The incorrect responses were pooled
by gram matical category of the stimulus item and the response. A chi-square analysis
(Bruning and Kintz, 1977) was perform ed to determine any association betw een the
gram m atical class of the stimulus and the grammatical class of the incorrect response
(confidence level of .05). If an association was found, in light of the potentially large
sam ple size a contingency coefficient (Bruning and Kintz, 1977) was perform ed to
determ ine the strength of the association.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER III: RESULTS
Three of the 32 subjects failed the screening criteria for pure tones or acoustic
im m ittance for one ear, so testing was perform ed on one ear only. All other subjects
had data for both ears and consequently, data were collected for a total of 61 ears.
Data w ere collected for all three presentation levels per ear for 30 subjects. However,
two subjects were run w ith only two out of three presentation levels for one ear only,
due to examiner error.
The num ber of stimulus items in each grammatical category for each M aryland
CNC word list appears in Table 1. The total num ber of stimulus items in each
grammatical category for the 61 ears appears in Table 2.
A two-way repeated analysis of variance (ANOVA)(Bruning and Kintz, 1977)
was conducted on each subjects' proportion of error by ear and grammatical category
of the stim ulus item. This analysis was performed to test the first hypothesis that the
stim ulus item s in the
NB category would be answered w ith significantly fewer errors
than those in the Other category. For this analysis only the 29 subjects for w hom both
right and left ear data w ere available were used. These 29 subjects included the tw o
subjects w ho received only two presentation levels for one ear.
The m ean proportion of error for each ear condition for the two grammatical
categories of the stim ulus word w as calculated (Table 3). The sum of squares was
calculated for: the total proportion of error; each subject; the differential effects of
right ear versus left ear; the differential effects of NB versus Other grammatical
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
Table 1
Granimatical Class of Stimulus Items on the
Maryland CNC Test Presented Per List
— Rated Grammatical Class—
CNC List
Noun-Both
(NB) Other Tota
1-1 39
11 50
1-3
36 14 50
1-6
32 18 50
1-7
30 20 50
1-9
41 9 50
1-10
37
13 50
TOTAL;
215
85 300
***********************
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2
Total Stimulus Items (61 Ears) on
The Maryland CNC Test Presented Per List
18
Number of
Rated Grammatical Class
--
Noun-Both
CNC List Times Used (NB) Other
Total
1-1
28 1,092
308 1,400
1-3 31 1,116 434
1,550
1—6
31
992 558 1,550
1-7 28
840
560 1,400
1-9
32
1,312
288 1,600
1-10
31
1,147 403 1,550
TOTAL:
181
6,499
2,551 9,050
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
categories; the interactive effects of the ear and grammatical category factors; and the
error effects. A test of significance (F) was calculated for the ear condition,
grammatical category, and the interactive effects of both. The results (Table 4)
indicated no significant differences (p<.05) for either the right versus left ear condition
or the ear conditions versus the grammatical categories. However, significantly fewer
errors (p<.05) were committed on the stimulus items in the NB category compared to
those in the Other category.
Three conclusions w ere draw n from the results. First, the ear condition had no
significant effect on the errors committed. Second, the effects of the category of the
stimulus item s and the ear condition did not interact to a significant degree. Third,
errors comm itted w ere significantly related to the category of the stim ulus item, with
fewer errors occurring in the NB category. Therefore, the hypothesis that the stimulus
items falling w ithin the N B grammatical category would be identified w ith fewer
errors than those items in the Other category w as supported.
The second hypothesis was the grammatical class of the incorrect responses
w ould not be significantly associated w ith the gram matical class of the stimulus item.
To test this hypothesis, a chi-square analysis (Bruning and Kint
2
;, 1977) was
perform ed on the incorrect responses for all 61 ears. Items w ere placed according to
grammatical class of the stim ulus item and grammatical class of the resulting
incorrect response (Table 5). The chi-square analysis indicated a significant
association (X^ = 6.94; p <.05) betw een the grammatical class of the stimulus item and
that of the incorrect response. A contingency coefficient w as calculated to determine
the strength of the relationship. The contingency coefficient (C = 0.075) indicated that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Table 3
Incorrect Responses to Stimulus Items on the Maryland CNC Test
(N = 29 Subjects)
Proportion
of Error
Sum
Mean
———— Right Ear ————
Noun-Both
———— Left Ear ————
(NB)
3.25
.112
Other
5.20
.179
Noun-Both
fNB)
4.15
.143
Other
5.10
.176
***********************
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
T a b l*
4
Two Way R e p e a t e d A n a ly s i s o f V a r ia n c e o f t h e
P r o p o r t i o n o f E r r o r f o r G r a m m a tic a l C a t e g o r y a n d E a r
21
D e g re e s
Sum o f o f
S o u r c e S q u a r e s F re e d o m
TOTAL: .8 1 8 4 115
I n d i v i d u a l S u b j e c t s .3 4 0 8 28
RIGHT VS LEFT EAR .0 0 5 5 1
GRAMMATICAL CATEGORY .0 7 2 4 1
GRAMMATICAL CATEGORY X EAR .0 0 8 8 1
E
RROR EFFECTS
ERROR EAR .2 0 6 3 28
ERROR CATEGORY .1 0 2 6 28
ERROR CATEGORY X EAR .9 2 2 2 28
M ean
S q u a re s
.0 0 5 5
.0 7 2 4
.0 0 8 8
.0 0 7 4
.0 0 3 7
.0 3 3 0
.7 4 9
1 9 .7 6 0
2 .9 8 9
S i g n i
f i c a n c e
P < . 05
No
Y es
No
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
T a b le 5
C h i - S q u a r e A n a l y s i s o f t h e A s s o c i a t i o n B e tw e e n G r a m m a t ic a l C a te g o r y o f
S t i m u l u s I te m s a n d G r a m m a tic a l C a te g o r y o f I n c o r r e c t R e s p o n s e s
S ti m u l u s
Ite&
NB
O th e r
TOTAL;
N o u n -B o th
(NB1
5 0 3 ( 5 2 0 )*
32 7 ( 3 1 0 )*
8 30
I n c o r r e c t R e s p o n s e
O th e r
158 ( 1 4 1 ) *
68 ( 8 5 )*
226
U n i d e n t i f i e d
11 2 (1 1 2 )*
67 (6 7 ) *
17 9
T o t a l
77 3
4 6 2
1 ,2 3 5
( )* = E x p e c te d V a lu e
* *
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
approxim ately 7 percent of the variance could be attributed to the relationship of the
gram matical classes of the stimulus and the response items. The significant chi-
square value would reject the hypothesis, however, the contingency coefficient w ould
caution the weak nature of the association.
Reliability
Approximately 10 percent of the data (6 ears) were randomly selected for
inter-rater reliability. The second judge listened to the subjects' audiotaped responses
and scored them as correct or incorrect. The incorrect responses w ere
orthographically recorded and compared to the prim ary exam iner's judgem ent of
these items. Both judges had to agree on the correctness or incorrectness of the
response. In addition, the judges were required to concur on the w ord given for the
incorrect responses.
Inter-rater reliability ranged from 90% to 100% (x=96.2%). Two of the
presentation levels achieved 90% inter-rater reliability and the rem ainder equaled or
exceeded 94% inter-rater reliability. For the purposes of data analysis, the prim ary
exam iner's judgem ent w as accepted when disagreements of responses occurred
betw een the two judges.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER IV; D IS C U S SIO N
The purpose of this study was to investigate the types of errors m ade on The
M aryland CNC Test by normal hearing listeners w ho were sim ilar to those on whom
the test was originally normed. Test conditions were similar to those used in the
original study (Causey et al., 1984) including some presentations at reduced sensation
levels that yielded response errors from normal hearing subjects.
Two hypotheses were addressed. The first hypothesis predicted that stimulus
items w hich could be nouns would be identified with significantly fewer errors than
those items which were not nouns. This hypothesis was statistically supported. The
second hypothesis predicted that the grammatical class of the incorrect responses
w ould not be significantly associated w ith the grammatical class of the stim ulus item.
Results indicated a significant but very w eak association between the grammatical
class of the incorrect response and that of the stimulus item.
There are several factors which m ay have contributed to the results. As
mentioned in the literature review, the carrier phrase "Say t h e again" m ay be
operating to influence the response. The syntactic constraint of the carrier phrase
would require a noun response to m aintain correct syntax. It is not surprising then
that listeners were more likely to respond to stimulus items which w ere nouns and
thus m aintained the syntax of the phrase. These results are consistent with those of
Miller and Isard (1963) w ho found significantly better perform ance on grammatical
sentences than on anom alous and ungrammatical sentences, particularly under
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
difficult listening situations. Any bias toward responding as though the stimulus
item w ere a noun m ay have been strengthened by the high proportion of stim ulus
item s on The Maryland CNC Test w ord lists which can be categorized as nouns. For
the six w ord lists used (Appendix C), 71.6% of the stim ulus items fell within the NB
category.
Another possible explanation for the results is that subjects may have
responded according to the perceptual salience of the stim ulus items. That is, if items
falling w ithin the NB category tended to be more concrete than those items falling in
the Other category, then the more concrete NB items may have been more
perceptually salient to the listener and thus responded to w ith fewer errors.
However, Elkins (1971) examined the CNC word lists for phonetic composition and
w ord familiarity and found the w ord lists "are relatively uniform in w ord familiarity"
(p.l59). Elkins' hndings suggested w ord familiarity should not be contributing to this
study's results.
A third possible explanation m ay be related to the gender of the subjects.
Gleason (1989) stated that some oral language skills are more advanced in females
than males and that "while girls appear more fluent and automatic in their ability to
perform various linguistic tasks, boys seem to be better in receptive and expressive
vocabulary" (p. 253). Given the possibility of linguistic differences based on gender,
then females m ay be m ore or less susceptible to the categorical differences of the
stim ulus items than males. If so, the findings for this female population m ay not be
representative for the population as a whole.
If the syntax of the carrier phrase contributed to fewer noun stimuli being
missed, then one would expect the second hypothesis to be borne out. The second
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
hypothesis predicted that the incorrect responses would not be significantly associated
w ith the grammatical class of the stim ulus item. However, the chi-square analysis
(Bruning and Kintz, 1977) show ed a significant association between the grammatical
categories of the stim ulus items and the incorrect responses. For incorrect responses,
stimulus items in the NB category were slightly more likely to yield an NB response
than stim ulus items in the Other grammatical category. How ever, only 7 percent of
the total variance could be accounted for by this relationship. Thus, the second
hypothesis could neither be fully accepted nor rejected. It appeared that the
significance of this w eak association could, at least partially, be attributed to the
mathematical effects of the very large sample size (1,235 items).
A factor which m ay have contributed to the results is the frequency with
which certain stim ulus items elicited consistent incorrect responses. Table 6 lists the
stimulus items that w ere incorrectly responded to greater than 50% of the time, their
gram matical categories, and the grammatical categories of the errors.
M ost of the stim ulus item s in Table 6 (8 out of 10) drew incorrect responses
that fell w ithin a particular grammatical category more than 70 percent of the time.
Seven of the 10 most commonly missed words consistently drew NB responses. For
some stim ulus items the errored responses varied widely. For other stimulus items,
the response errors tended to be one or two words, for example, "sun" for stim ulus
item "sung" and "with" or "width" for stim ulus item "wit." Thus commonly missed
w ords m ay have been influenced by characteristics of the stimulus items and by the
availability of phonetically sim ilar foils in the subjects' lexicon.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
T a b l# 6
S t im u l u s I t e m s t h a t E l i c i t e d I n c o r r e c t R e s p o n s e s i n
M o re T h a n F i f t y P e r c e n t o f t h e S u b j e c t s
27
s t i m u l u s
R a t e d CNC
G r a m m a t ic a l W ord
P r o p o r t i o n
o f S u b j e c ts
Who
E r r o r e d o n
P r o p o r t i o n o f E r r o r b y
G r a m m a t ic a l C a t e g o r y
o f t h e R e s p o n se
I te m C a t e g o r y L i s t
R e s o o n s e NB O t h e r U n id e n tJ
c a u g h t
O
1 - 7 .6 8 .8 9 .0 5 .0 5
b e t NB
1 - 7
.6 4 .7 8 .2 2 0
t h i n e NB 1 - 9
.6 2 .5 5 .4 5 0
f i t NB
1 - 1
.6 1 .5 3 .3 5 .1 2
c h e e r NB
1 - 9
.5 9 .9 5 .0 5 0
w i t
NB
1 - 7
.5 7 .1 9 .7 5
.0 6
t h i n
O
1 - 1
.5 4
.7 3 .1 3 .1 3
b o t h O
1 - 9
.5 3 .9 4
0
.0 6
s u n g
O
1 - 3
.5 2
.9 4 .0 6 0
h i t NB
1 - 1 0
.5 2
.7 5 .1 2 .1 2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
Clinical Im plications
The results of this study clearly have clinical implications. In developing The
M aryland CNC Test. Causey et aJ. (1984), stated that the test achieved criteria of
phonetic/phonem ic balance and w ord familiarity while providing an environm ent in
the carrier phrase "Say t h e again" which m inim ized the effects of coarticulation.
They did not, however, address any possible negative effects of the carrier phrase.
Results of this study indicated that stimulus items that fell within the grammatical
class NB w ere answ ered w ith fewer errors which may possibly be due to the syntactic
constraints of the carrier phrase.
One clinical alternative w ould be to examine use of the CNC word lists with a
syntactically neutral carrier phrase which minimizes coarticulation such as "Say
____
again." A nother more complicated adjustment w ould be to develop CNC word lists
w ith only stimulus item s in the NB category while m aintaining the phonetic/
phonem ic balance and uniform word familiarity of the present lists. Either option
w ould reduce the bias seen in this study, w hether that bias is due to syntactic
constraints a n d /o r saliency of the stimulus items, however, the first option is a more
clinically feasible modification. The second option m ay require almost complete
reformulation of the w ord lists for phonemic balance to be m aintained. Also, there
m ay be value in m aintaining a full range of grammatical categories when assessing
word recognition.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Limitations of the Study and Implications for Future Research
A prim ary limitation of the study was the judging of the grammatical
categories. In order to include common slang usage which m ay not have been
identified by a standard dictionary, judges determ ined grammatical categories. The
following problems w ere noted w ith the judging of grammatical categories in this
study. For several (approxim ately 15) of the stimulus items and incorrect responses,
no consensus w as reached by the judges and the dictionary was used to determine
the categories of those items.
A second problem in the system used for assigning grammatical categories
w as disagreem ent between the judges' rating and the prim ary exam iner's informal
rating of the gram matical category of various items. For example, the consensus
rating for the stimulus w ord "dead" was the Other category, whereas the prim ary
examiner judged it to be a NB. These instances were too few to significantly affect
the data, but occurred nonetheless.
A third problem related to hom ophonous words. A lthough the judges w ere
instructed to consider hom ophones w hen judging grammatical categories,
orthography m ay have influenced some of the judgements. For example, the stimulus
items "sell" and "which" w ere judged as Other, however, both are hom ophonous with
"cell" and "witch" (in the judges' dialect), which w ould shift the category to NB.
In order to avoid the above problem s in hi tu re research it is suggested that
any item s which need to be assessed for grammatical category be presented w ith
hom ophones denoted and that the standard dictionary classification be noted (e.g.,
sell/cell - NB). The judges could then determ ine any additional slang usage of the
items in the Other category. It is hoped that the consensus problem would be
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
resolved by eliminating the problems of orthography and deviations from the
standard gram m atical classification.
An issue not addressed in the present study w as the effect of the stimulus
items on the reliability betw een w ord lists. Further research could determ ine whether
the presence or absence of particularly difficult stimulus items, as listed in Table 6,
significantly affects a client's word recognition performance such that a person's score
is determ ined in part by the presentation of a particular word list.
Other areas of potential research include pursuing w hy the stimulus items that
could be used as nouns w ere identified w ith few er errors than those items which
w ere not nouns. To test the syntactic effects of the carrier phrase "Say th e again,"
a m ore syntactically neutral carrier phrase (e.g., "Say again") could be employed
using the same stim ulus items. If items from both categories (i.e., NB and Other) were
answ ered correctly w ith greater frequency using the more neutral carrier phrase, then
the effects of the carrier phrase could be isolated. Another possibility would be to
present only
Other stim ulus items w ith the carrier phrase, "Say th e again." This
would elim inate the expectation of noun stimulus items by the listener and would
thus investigate the effects of syntax on the correctness of the response.
A nother area of potential research is to explore gender effects to determine if
female listeners are m ore susceptible to the categorical effects of the stim ulus items
than male subjects w ould be. A normal hearing group of males m atched for age and
presentation conditions could be tested to determ ine any gender differences in the
types of response errors.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
Summ ary
The present study found that when The Maryland CNC Test (Causey et a\.,
1984) w as analyzed in the m anner in which it w as originally norm ed and
recom m ended for use, the stimulus items that can be recognized as nouns were
identified w ith significantly fewer errors than those items which were not nouns. The
im plication of the finding is to question the content validity of The Maryland CNC
Test. Results m ay be influenced by the grammatical class of the stim ulus items and
unduly influenced by some particularly difficult items. Further research is needed to
investigate the effects of the carrier phrase, "Say th e again" on responses to
stim ulus items.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
r e f e r e n c e s
American National Standards Institute (1969). Specifications
for audiometers. ANSI S3.6-1969. New York.
American National Standards Institute (1977). Criteria for
permissible ambient noise during audiometric testing. ANSI
S3.1-1977. New York.
American National Standards Institute (1981). Specifications
for audiometers. ANSI S3.6-1969. New York.
ASHA Committee on Audiometric Evaluation (1988). Guidelines for
determining threshold level for speech. ASHA. 30, 85-89.
Bruning, J. L., and Kintz, B. L. (1977). Computational handbook
of statistics. Glenview: Scott, Foresman and Company.
Bzoch, K. R. (1989). Communication disorders related to cleft
lip and palate. Boston: Brown, Little and Company.
Causey, G. D., Hood, L. J., Hermanson, C. L., and Bowling, L. S.
(1984). The Maryland CNC Test: Normative studies.
Audioloav. 23, 552-568.
Causey, G. D., and Elkins, E. (Date Unknown*). Marvland speech
intelligibility materials - The CNC Test [Cassette
Recording]. Glendale: Olsen Distributors.
Chomsky, N. (1965). Aspects of the theory of syntax.
Cambridge: M.I.T. Press.
Craig, C. H. (1988). Effect of three conditions of
predictability on word recognition performance. Journal of
Speech and Hearing Research. 31, 588-592.
Elkins, E. F. (1970). Analyses of the phonetic composition and
word familiarity attributes of CNC intelligibility word
lists. Journal of Speech and Hearing Disorders. 35, 156-
160.
Gelfand, S. A. (1975). Use of the carrier phrase in live voice
speech discrimination testing. Journal of Auditory
Research. 15, 107-110.
Gladstone, V. S., and Siegenthaler, B. M. (1971). Carrier phrase
and speech intelligibility test score. Journal of Auditory
Research. 11, 101-103.
♦The date was unavailable from Olsen Distributors and Dr. Causey
could not be reached to provide this information.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
Gleason, j. B. (Ed.)(1989). The development of language (2nd
Ed.). Columbus; Merrill Publishing Company.
Hayes, D. (1984). Diagnostic applications of speech
recognition. In E. Elkins (Ed.), Speech recognition bv the
hearing impaired, ASHA Reports, 14, 49-51.
Kreul, E. J., Bell, D. W . , and Nixon, J. C. (1969). Factors
affecting speech discrimination difficulty. Journal of
Speech and Hearing Research, 12, 281-287.
Lynn, J. M. , and Brotman, S. R. (1981). Perceptual significance
of the CID W-22 carrier phrase. Ear and Hearing. 2, 95-99.
Martin, F. N., Hawkins, R. R. , and Bailey, H.A.T. Jr. (1962).
The nonessentiality of the carrier phrase in phonetically
balanced (PB) word testing. Journal of Auditory Research,
2, 319-322.
Miller, G. A., Heise, G. A., and Lichten, W. (1951). The
intelligibility of speech as a function of the context of
the test materials. Journal of Experimental Psychology. 41,
329-335.
Miller, G. A., and Isard, S. (1963). Some perceptual
consequences of linguistic rules. Journal of Verbal
Learning and Verbal Behavior. 2, 217-228.
Penrod, J. P. (1985). Speech discrimination testing. In J.
Katz (Ed.), Handbook of clinical audiology. (pp. 235-258).
Baltimore: Williams and Wilkins.
Thornton, A. R. (1984). Speech recognition testing and
understanding the effects of disease on function of the ear.
In E. Elkins (Ed.), Speech recognition bv the hearing
impaired. ASHA Reports, 14, 49-51.
Urdang, L. and Flexner, S. B. (Eds.)(1968). The Random House
college dictionary. New York; Random House, Inc.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
APPENDIX A: ARTICULATION SCREENING TEST
/p/
aPPle
/b/
baBy
/t/
mounTain
/d/ canDy
/k/
chicKen
/g/
waGon
It/
elePHant
/v/
shoVel
/e/
tooTHbrush
/é
/ feaTHer
/s /
bicycle
/z/
sciSSors
/ / /
diSHes
/3 /
television
/t;/
maTCHes
/d^ /
briDGes
/w /
sandwich
/I
/
baLLons
/j /
onlOns
/r /
aRRow
/m/
haMMer
/n/
baNana
/J/
haNGer
From: Bzoch Error Pattern Diagnostic Articulation Test (1989)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
APPENDIX B: SUBJECT INSTRUCTIONS AND PRACTICE STIMULI
The following instructions were given to each subject prior
to test administration:
Soon you will hear a man's voice on a tape. He
will always say the same sentence, but only one word
will change each time. He will say, "Say the blank
again". I want you to repeat the word after "the".
For example, if he says "Say the horse again", you
repeat horse. Do you understand? If you are not sure
of the word, please guess. Now let's try some
practice sentences :
1. Say the bat again.
2. Say the sad again.
3. Say the tub again.
4. Say the jug again.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
a p p e n d i x C: THE MARYLAND CNC SPEECH INTELLIGIBILITY TESTS
WORD LISTS
List 1-1
1.
jar
23.
wheel
2. boil
24.
fit
3.
tough
25.
patch
4. tooth
26.
make
5.
goose
27,
dime
6. toad
28.
bean
7.
rout
29.
thin
8.
mess
30.
seize
9.
kite
31.
hate
10. jug
32.
wood
11. pad
33.
check
12. salve
34.
ditch
13. van
35. rose
14. home
36. merge
15.
cape 37.
lease
16.
shore
38.
loop
17.
wreck
39. king
18.
shirt
40. dead
19.
knife
41. chore
20. hull
42.
boat
21.
yearn
43.
wish
22.
sun 44.
name
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
APPENDIX C (continued)
List 1-1 (continued)
45.
pick
46.
ripe
47.
fall
48. lag
49.
gale
50. sob
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
APPENDIZ C (continued)
List 1-3
1.
jail
26.
fade
2.
rat
27.
lake
3.
toss
28. gull
4.
soon
29.
rouge
5.
faith
30.
bar
6.
sung
31.
tone
7.
keg
32.
chin
8.
vote
33. piece
9.
size
34.
purge
10. numb
35.
bell
11.
dab
36. work
12. what
37. life
13.
room
38.
pod
14 .
kid 39. shine
15.
dike
40.
toll
16. mate
41. joke
17. well
42. head
18. rig
43.
with
19. four
44.
keen
20.
bush
45. more
21.
dip
46. leave
22 .
gap
47. hut
23 .
perch 48. noise
24.
sheep
49.
man
25 .
house
50.
yam
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
APPENDIX C (continued)
List 1-6
1.
whip
26.
door
2.
bud
27.
niece
3.
shone
28.
cat
4.
rug
29.
move
5.
cheese
30.
cool
6.
chain
31. web
7.
look
32.
knock
8.
dull
33.
jot
9.
pope
34.
cage
10.
calf
35.
mode
11. fire
36.
search
12.
turn
37.
gone
13. raise
38.
rush
14.
sour
39. pole
15.
bed
40. dig
16. lawn
41. bad
17.
sit 42. live
18.
tube
43. map
19. veal
44. wife
20.
get
45. fan
21. pace
46.
birth
22.
night
47. team
23.
hiss
48. howl
24.
shock 49.
hike
25 .
wing
50.
jam
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
APPENDIX C (continued)
List 1-7
1.
note 26.
reach
2.
doom
27.
face
3.
coke
28.
bet
4.
hole
29.
caught
5.
join
30.
laugh
6.
third
31.
shall
7.
mouth
32.
geese
8.
sure
33.
tape
9.
vague
34.
sack
10.
big
35. ridge
11. far 36.
cheek
12.
gun 37.
dumb
13.
pearl 38.
top
14.
loot
39.
young
15.
save
40.
led
16.
side
41.
rib
17.
heat
42. pass
18. bun
43.
wit
19. fish
44. did
20.
have
45. call
21. mole
46.
neck
22.
pine
47.
such
23.
nap
48.
lose
24 .
mine
49.
gem
25.
was
50.
tar
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX C (continued)
41
List 1-9
1
.
2 .
3.
4.
5.
6 .
7.
8.
9.
1 0.
1 1
.
1 2 .
13.
14.
15.
16.
17.
18.
19.
2 0 .
2 1 .
22 .
23.
24.
25.
lack
watch
power
mire
nail
thine
word
tool
mob
hen
got
sane
shout
pill
both
shade
jazz
lathe
catch
white
chair
loaf
pun
ham
lip
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
wrong
yes
sin
curve
haze
girl
time
book
reap
fudge
voice
rag
mud
ball
deck
cut
need
cheer
soap
feet
tick
roof
dog
beat
dish
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
APPENDIX C (continued)
List 1-10
1.
sub
26.
shack
2.
lot
27.
cone
3.
din
28. sell
4.
death
29. your
5. chill
30.
term
6. coin
31.
mood
7. cause
32. deep
8.
burn
33.
meek
9. loose
34. rope
10. palm
35.
witch
11. judge
36.
ride
12. wash
37. bake
13.
rob
38.
gore
14.
fine
39. fool
15. while 40. guess
16. chat
41. mouse
17. bit
42.
lung
18. nick
43. load
19.
neat
44. path
20.
hair
45. peak
21.
safe
46.
run
22. hit
47. sag
23.
jade
48. cave
24 .
hurt
49.
thatch
25.
pile
50.
towel
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
