lurunal oJ Speech and Hearing Reserll I, \'olutne .5, 094-660, lune 1992
Reliability of Two Measures of
Speech Recognition in Elderly
People
Carol Geltman Cokely
Larry E. Humes
Indiana UmverstN
Bloomington, IN
The revised Speech Perception In Noise (SPIN) test and the Dichotlc Sentence Identification
(DSI) test have been used to help evaluate speech-recognition capabilities in elderly people. We
evaluated the test-retest reliability of these measures for 17 subjects aged 63-82 years. The DSI
and revised SPIN tests were administered at 65, 75, and 85 dB SPL, with a total of three
presentations at each level. Reliability was assessed using a repeated-measures analysis of
variance and 95% critical differences for each test. Results raise serious questions about the use
of these tests for diagnostic determinations or assessment of speech-recognition ability in
elderly people
KEY WORDS: aged, speech audiometry, reliability
For individuals aged 65 or older, hearing impairment is the third-ranked permanent
disability, affecting 25-40% of this population (Committee on Hearing, Bioacoustics,
and Biomechanics [CHABA], 1988; National Center for Health Statistics [NCHS],
1977). Furthermore, speech-recognition capabilities of elderly people often appear to
be poorer than would be predicted from the degree and nature of peripheral sensitivity
loss (Bergman et al., 1976; Dubno, Dirks, & Morgan, 1984; J. Jerger & Hayes, 1977b;
Orchik & Burgess, 1977; Shirinian & Arnst, 1982). Elderly people's speech-recognition deficits frequently mimic central auditory dysfunction that arises from focal lesions
to the ascending central auditory pathways (J. Jerger & Hayes, 1977a; J. Jerger & S
Jerger, 1974; Noffsinger & Kurdziel, 1979). Therefore, it has been proposed that
individual differences in speech-recognition capabilities of the elderly may result from
impaired processes of the central auditory system. It has also been proposed that the
decline in speech recognition with age is related to cognitive factors such as memory,
perceptual organization, and attention (CHABA, 1988).
Two tests that have been used recently to examine the speech-recognition
performance of elderly people are the Dichotic Sentence Identification (DSI) test and
the revised Speech Perception in Noise (SPIN) test. The DSI test (Fifer, J. Jerger,
Berlin, Tobey, & Campbell, 1983) is a closed-response test consisting of six synthetic
sentences taken from the standard Synthetic Sentence Identification (SSI) test with
pairs of sentences presented dichotically under headphones (Speaks & J. Jerger,
1965). The SPIN test is a 50-item test with 25 predictability-high (PH) and 25
predictability-low (PL) sentences delivered monotically in the presence of multitalker
babble using an open-response format (Kalikow, Stevens, & Elliott, 1977). Jerger and
colleagues (J. Jerger, S. Jerger, Oliver, & Prozzolo, 1989; J. Jerger, Oliver, &
Pirozzolo, 1990) have used both of these tests diagnostically to evaluate central
auditory processing capabilities in the elderly.
There is little information on the reliability of either of these tests when administered
to elderly listeners with hearing mpairment. Bilger, Nuetzel, Rabinowitz, and RzecD 1992, American Speech-Language-Hearing AsnoCli:atlon
654
0022!-46855 '9 2/3ii03i 005·$0
.0'
Cokely & Humes: Reliability of Speech Recognition Measures
zkowski (1984) examined the reliability of the SPIN test for
use with individuals younger than 70 years with hearing
impairment. When revised lists of only the 200 most reliable
sentences from the original set of 250 sentences were used,
the lists yielded equivalent scores, demonstrated homogeneity of variance, and were of equivalent and high reliability.
Moreover, Bilger et al. found no significant change in performance across two repetitions of the revised SPIN materials.
However, the subject population in the Bilger et al. study was
under 70 years of age, with the median age for this group
below 50 years. Thus, it remains to be seen whether the
SPIN test is a reliable measure for elderly listeners with
hearing impairment.
Fifer et al. (1983) evaluated the clinical application of the
DSI for listeners with hearing impairment. They found no
significant differences between test scores obtained at a
presentation level 50 dB above the pure-tone average and
those obtained at 50 dB HL. The variance for both presentation methods, however, increased as hearing loss increased. The reliability of the DSI has not been previously
evaluated; however, the reliability of the SSI (from which the
DSI is derived) has been evaluated with respect to its use
with individuals with hearing loss (Dubno & Dirks, 1983). In
the study by Dubno and Dirks, the SSI was administered six
times in succession on two separate occasions. Large intrasubject variability existed through the first three repetitions of
the SSI on both occasions. Although mean performance
stabilized with subsequent trials, large intrasubject variability
remained and alpha reliability coefficients were low and
nonsignificant. Given the questionable reliability of the SSI,
the reliability of a test based on the SSI, such as the DSI,
should be investigated.
Scores across trials for any given speech-recognition
measure are expected to vary about a 'true' score. For a test
to be useful, these variations must be small enough so that
each score is a reliable reflection of the true score. If
repeated measurements of an identical test condition result
in large test-retest differences, then such a test can not
illustrate reliable differences between populations or between
test conditions. If the SPIN test and DSI test are to be used
to evaluate speech-recognition capabilities in elderly listeners with hearing impairment, then the reliability of these
measures should be established for this population. The
purpose of this study was to evaluate the reliability of the DSI
and SPIN tests in elderly individuals.
Method
Subjects
Seventeen subjects aged 63-82 years (M = 70.6 years),
participated in the study. All subjects exhibited a symmetric
sensorineural hearing loss (interear differences < 15 dB HL)
with three-frequency pure-tone averages (PTA) (0.5, 1.0, and
2.0 kHz) ranging from 18 to 52 dB HL (ANSI, 1969). The
mean right-ear and left-ear PTAs were 34.3 dB HL and 32.7
dB HL, respectively.
655
Apparatus
The tapes of the revised SPIN test were provided by Dr.
Robert Bilger at the University of Illinois. The DSI tapes were
provided by Auditec of St. Louis. Tapes were played through
a two-channel cassette-tape deck (Sansui, D-W9). For the
SPIN test, the output of one channel of the tape deck was
routed to an attenuator and then to one channel of a
two-channel amplifier (McIntosh, C24), and was mixed with
the output of the other channel of the tape deck. Amplifier
output was delivered to earphones of a network of matched
TDH-39 earphones mounted in MX-41/AR cushions. All
presentation levels were verified using an NBS-9A 6-cm 3
coupler. The outputs of both channels were calibrated electrically prior to each test administration using a voltmeter.
Procedures
Subjects received practice lists of the SPIN test presented
to the test ear and the DSI test presented monaurally to each
ear, in quiet at 75 and 85 dB SPL. Test presentation levels for
the DSI and SPIN tests were 65, 75, and 85 dB SPL for all
subjects so that scores could be obtained at speech levels
spanning soft to loud conversation. For the SPIN test, babble
was delivered at an 8-dB signal-to-noise ratio. The DSI test
had two equivalent forms, and the revised SPIN test had
eight forms. There were a total of three presentations at each
level. The first and second presentations of each test were
separated by approximately 1 hr, with a third presentation
following 1-2 weeks later. For each presentation, the order of
presentation level and test was randomized. Answer forms
for both the DSI and SPIN tests were provided. For the DSI
test, all subjects had before them a list of the DSI sentences,
numbered 1-6. Subjects had to write down the numbers
corresponding to the two sentences delivered under earphones, without regard to ear specificity. For the SPIN test,
subjects were required to write the last word of the sentence.
Subjects were encouraged to guess when they were unsure
of an answer.
All testing was completed in an acoustically treated room
with ambient noise levels lower than those required for
threshold measurement with headphones (ANSI S3.1-1977).
Approximately 3 hr, including frequent breaks, were required
of each subject during the first test session (practice with
SPIN and DSI; Trials 1 and 2 of the SPIN and DSI), and
approximately 1 hr was required during the second test
session (Trial 3 of SPIN and DSI). Subjects were paid for
their participation.
DSI scores consisted of individual ear scores. SPIN scores
were obtained for the left ear only and consisted of percentcorrect scores for both PL and PH sentences. All percentcorrect scores were converted to rationalized arcsine units
(raus) prior to data analysis (Studebaker, 1985). The rau, like
the arcsine transform, stabilizes the error variance, but unlike
the arcsine transform, it is numerically similar in value to the
original percent-correct scores when scores are between
10% and 90%.
656 Journal of Speech and Heanng Research
35
Results
654-660
une 1992
20
SPIN
The mean PL and PH scores and their associated standard deviations across trials and levels are given both in
percent correct and in raus in Table 1. The large standard
deviations are a reflection of the large individual differences
between subjects.
The reliability of the SPIN scores was assessed through
the following: (a) calculation of the standard error of measurement (SEM) and the standard error of the difference
(SED) between repeated measures, and (b) analysis of
variance (ANOVA) for repeated measures. The standard
error of measurement was given by
15
r
cl
0
LL
w
10
C
a
5
SEM = SDgm (1 - a)1 2
where SDgm is the geometric mean of the standard deviations across the three trials for a given condition and a is the
coefficient alpha measure of reliability for repeated measures
calculated using the covariance matrix from the three trials.
The SEM reflects the precision or accuracy of the measurement by providing an estimation of how an individual's score
is expected to vary. SEMs were calculated for each condition
(3 Levels x 2 Materials). SEMs for PH and PL scores,
expressed in raus, are illustrated in Figure 1. In general, the
accuracy of the SPIN measurements is poorest at the lowest
presentation levels.
Next, the standard error of the difference (SED) was
calculated to establish significant differences in scores between repeated measures for identical conditions. The SED,
in raus, for SPIN lists was calculated as outlined by Studebaker (1985). First, the variance (V) of the arcsine transform
for each list was given by
o
II
I
65
75
Speech Leve
IIi
85
n dB SPL
FIGURE 1. Standard error of measurements (SEMs) for SPIN PL
and SPIN PH scores at 65, 75, and 85 dB SPL.
then determined by taking the square root of the sum of the
variances for the two samples being compared. For PH and
PL sentences, each with 25 items, the SED is 12.9 raus. A
95% critical difference, in raus, of 25.3 was determined for PL
and PH sentences by multiplying the SED by 1.96. Testretest differences for each subject and condition were then
compared to the 95% critical difference. The percentage of
subjects with test-retest differences exceeding the 95% critical difference is illustrated in Figure 2. As can be seen, when
scores for all trials were compared, approximately 40% of all
subjects displayed test-retest differences that exceeded the
95% critical difference, except for the PH condition at 85 dB
SPL. Even when the scores for Trial 1 were regarded as
practice and omitted (unfilled bars in Figure 2) and test-retest
differences were evaluated for only the latter two trials, a
V = [1/(N + 0.5)] K,
where N is the number of items in the list, and K is a
multiplicative constant used to yield equivalent values between transformed and untransformed scores. The SED was
TABLE 1. Means (M) and standard deviations (SD) of Speech Perception In Noise test PL and PH scores In percent correct and in raus
across trials and levels for all subjects.
Trial 1
Trial 2
raus
%correct
Trial 3
% correct
dB
SPL
M
SD
M
SD
M
65
75
18.4
19.4
23.3
24.3
11.8
13.7
29.1
30.9
16.8
26.7
85
41.2
29.8
38.4
32.8
28.8
65
75
85
39.3
55.6
72.0
31.6
39.6
35.4
32.8
53.7
76.4
34.3
45.6
39.5
54.0
62.8
64.0
SD
PL scores
21.6
25.9
25.6
PH scores
36.7
36.8
35.3
Note. PL = predictability-low sentences; PH = predictability-high sentences.
raus
% correct
raus
M
SD
M
SD
M
SD
10.4
23.7
27.8
29.9
17.4
39.3
22.9
27.8
12.3
38.4
27.5
28.8
23.7
31.7
36.2
30.2
32.8
34.2
53.5
63.5
61.8
42.8
42.7
36.4
52.2
70.8
76.7
36.7
35.5
29.9
51.7
73.9
81.9
41.9
40.7
35.2
Cokely & Humes: Reliability of Speech Recognition Measures
657
As with the PL sentences, the ANOVA for the PH sentences indicated significant effects of trial [F(2, 26) = 10.4, p
< .01] and of level [F(2, 26) = 19.54, p < .01], and a trial by
level interaction [F(4, 52) = 5.41, p < .01]. Univariate post
hoc contrasts revealed that at 75 dB SPL, PH scores for Trial
3 were significantly better than those scores for Trial 1 (t =
-3.01, p < .01) and Trial 2 (t = -3.02, p < .01). At 85 dB
SPL, PH scores for Trial 2 were significantly poorer than
scores for either Trial 1 (t = 3.56, p < .01) or Trial 3 (t =
-5.17, p < .01). Thus, for both PH and PL sentences at 75
and 85 dB SPL, subjects displayed significant variation
across trials. Subjects achieved significant improvements in
performance even beyond two trials.
80
60
13
U
o
0
40
Uo
Il
4)
a_
20
DSI
Mean values and corresponding standard deviations for all
DSI scores, expressed both in percent correct and in raus,
are in Table 2. Mean DSI-R (right ear) and DSI-L (left ear)
scores differed by less than approximately 10% for a given
condition.
The reliability of DSI scores was evaluated in the same
manner as were SPIN scores. The SEM, in raus, for DSI
scores across levels is illustrated in Figure 3. The accuracy of
the DSI-R score improved with level, whereas the accuracy
of the DSI-L score appears to be worse than that of DSI-R
and less affected by level. In general, the DSI scores appear
to be somewhat less accurate than the SPIN scores (Figure
1).
For DSI-R and DSI-L conditions, each with 30 items, the
SED was calculated to be 11.9 raus. The 95% critical
difference for DSI scores, in raus, was 23.3. Figure 4 shows
that typically 40-50% of these subjects displayed significant
test-retest differences when scores across all trials were
compared. When test-retest differences for only the latter two
trials were compared, approximately 20% of these subjects
continued to demonstrate excessive test-retest differences.
The ANOVAs for ear-specific DSI scores revealed significant level effects for both the DSI-R scores [F (2, 26) =
12.44, p < .01] and DSI-L scores [F(2, 26) = 16.94, p < .01].
Trial effects were significant for the DSI-L scores [F(2, 26) =
0
PL65
PL75
PL85 PH65 PH75 PH85
Condition
FIGURE 2. Percentage of subjects who displayed test-retest
differences for SPIN PL and PH sentences greater than the 95%
critical difference at 65, 75, and 85 dB SPL. Dashed line ndicates 5% level of error.
sizable percentage of subjects evidenced test-retest differences greater than the 95% critical difference.
A repeated-measures analysis of variance (ANOVA) was
completed to evaluate differences between scores across the
three trials and the three presentation levels. For PL sentences, the ANOVA revealed significant main effects of trial
[F(2, 26) = 6.59, p < .01] and of level [F(2, 26) = 13.12, p
< .01], as well as a trial by level interaction [F(4, 52) = 5.69,
p < .01]. Univariate post hoc contrasts revealed that at 75 dB
SPL, scores were significantly better by Trial 3 when compared to either Trial 1 (t = -5.56, p < .001) or Trial 2 (t =
-4.86, p < .001). At 85 dB SPL, subjects' performance was
significantly poorer on Trial 2 than on Trial 1 (t = 3.02, p <
.01).
TABLE 2. Means (M) and standard deviations (SD) of DSI-L and DSI-R scores in percent correct and in raus across trials and levels
for all subjects.
Trial 1
Trial 2
%correct
raus
% correct
dB
SPL
M
SD
M
SD
M
65
75
40.1
65.2
41.1
36.5
37.5
66.4
47.2
42.2
61.5
68.4
85
68.8
33.8
69.9
38.5
75.6
65
75
85
Note. DSI =
Trial 3
41.2
40.5
39.1
47.5
72.7
34.9
76.1
39.8
68.8
35.8
70.9
41.1
Dichotic Sentence Identification test; L = left
SD
DSI-L scores
41.0
40.6
33.1
DSI-R scores
65.6
45.1
66.7
42.9
74.5
34.3
ear; R = right ear.
raus
%correct
raus
M
SD
M
SD
M
SD
62.0
70.0
45.9
48.2
62.8
88.6
38.4
23.2
64.2
93.1
42.5
26.7
78.2
38.5
87.4
23.5
92.6
26.8
65.2
68.2
78.2
53.3
50.7
38.2
58.2
78.0
78.8
45.6
37.8
33.4
57.5
81.1
82.4
54.4
44.8
38.5
658 Journal of Speech and Heanng Research
35
20
D51
=
effect when comparing DSI-L and DSI-R scores [F(13, 1) =
0.09, p > .05].
LE
RE
Correlation with Age and Hearing Loss
15
C
10
C
0
5
0
:
--=
-
65
-
-
d
-
-
,&
-
75
-
85
Speech Level in dB SPL
FIGURE 3. Standard error of measurements (SEMs) for DSI-L
(left ear) and DSI-R (right ear) scores at 65, 75, and 85 dB SPL.
80
60
Across presentation levels and trials, SPIN and DSI scores
were not significantly correlated with age (p > .05). However,
both the DSI and SPIN scores were significantly correlated
with pure-tone sensitivity. Pearson product-moment correlations between both SPIN PL and PH scores and test-ear
PTAs were low to moderate (-0.43 < r < -0.71), but most
were not significant. Correlations were moderate to high and
significant (p < .01) between high-frequency pure-tone averages (HFPTA) (1.0, 2.0, 4.0 kHz) and both PL and PH
scores (-0.53 < r < -0.89). Regarding DSI performance
and pure-tone sensitivity, correlations were moderate to high
between both DSI-R and DSI-L scores and HFPTA (-0.54 <
r < -0.86), and the majority were significant. Whereas
correlations were low to moderate and chiefly not significant
between PTA and DSI-L scores (-0.25 < r < -0.67),
correlations were high and significant between PTA and
DSI-R scores (-0.7 < r < -0.86). Although right-ear and
left-ear PTAs were highly correlated (r = 0.87, p < .01),
correlations were low and nonsignificant between DSI scores
for the right ear and scores for the left ear (0.04 < r < 0.32,
p > .05). That is, despite symmetry of hearing loss, an
individual's test score for one ear was independent of the test
score from the other ear.
Discussion
1A
(a
,
0.
To
(n
a
54-660 J]une 1992
40
V
a
20
0
LE65
LE75
LE85 RE65 RE75
Condition
RE85
FIGURE 4. Percentage of subjects who displayed test-retest
differences for DSI-L (left ear) and DSI-R (right ear) sentences
greater than the 95% critical difference at 65, 75, and 85 dB SPL.
Dashed line indicated the 5% level of error.
9.03, p < .01] but were not significant for the DSI-R scores [F
(2, 26) = 2.85, p > .05]. Trial-by-level interaction effects were
significant for DSI-L scores [F (4, 52) = 3.25, p = .01].
Univariate post hoc comparisons showed significantly better
DSI-L performance when scores for Trial 3 were compared to
scores for Trial 1 at 75 dB SPL (t = -3.15, p < .01) and at
85 dB SPL (t = -3.44, p < .01). ANOVAs revealed no ear
A test's reliability is basic to its diagnostic utility. The
results of our analyses of the DSI and SPIN tests raise
serious questions about the usefulness of these tests in the
elderly population. Test-retest variability for both the SPIN
and DSI tests for our group of elderly individuals was large.
The significant intertrial differences for these listeners for the
DSI and SPIN scores is somewhat disconcerting, given the
suggested use of these tests as diagnostic measures of
central auditory processing deficits (CAPD) in elderly people
(J. Jerger et al., 1989; J. Jerger et al., 1990).
Bilger (1984) evaluated the SEM and standard deviations
of SPIN PL and PH scores. For lists administered at 50 dB
SL, standard deviations were 25-30% for his group of 32
young listeners with hearing impairment, whereas SEMs
across lists averaged 6.3%. These values compare well with
our results at 85 dB SPL. Bilger (1984) did not evaluate
individual test-retest differences across equivalent forms
Dubno, Dirks, and Morgan (1984) used an adaptive strategy
to find the signal-to-babble ratio (S/B) needed to achieve
50% recognition. For SPIN PL and PH sentences, they found
no significant effects of trial and small SEMs and critical
differences for elderly listeners with hearing loss. Quantifying
speech-recognition difficulties in terms of S/B, therefore, may
be more reliable than use of percent-correct scores in the
elderly hearing-impaired population. Whether this measure
has any diagnostic utility for the elderly, however, remains to
be seen. Humes, Espinoza-Varas, and Watson (1988), for
instance, have demonstrated that the S/B is strongly affected
Cokely & Humes: Reliabllity of Speech Recognilon Measures
by the amount and configuration of the peripheral sensorineural hearing loss.
For the DSI, F:ifer et al. (1983) reported standard deviations of approximately 10-20% for listeners with pure-tone
averages less than 50 dB HL when lists were presented at
either 50 dB HL or 50 dB SL. Standard deviations for our
elderly listeners were considerably greater (see Table 2).
Additionally, Fifer et al. reported no significant level effects
when comparing DSI performance at 50 dB HL to performance at 50 dB SL for their hearing-impaired subjects. The
present data, however, revealed a significant improvement in
mean scores for our elderly listeners with hearing loss when
higher presentation levels were used.
To evaluate further the relation between presentation level
and test findings, results from a subgroup of 9 subjects who
displayed a mild hearing loss were evaluated. Although tests
were administered in this study at fixed sound-pressure
levels rather than fixed sensation levels, we reviewed the
SPIN scores and DSI scores across trials at 85 dB SPL for 9
individuals whose PTAs were better than 30 dB HL. By these
means, data were evaluated separately for a subgroup of
subjects who received the tests at a presentation level of
40-50 dB SL. Mean scores and corresponding standard
deviations are given in Table 3. The SEMs for the SPIN and
DSI scores at 85 dB SPL for these subjects with mild hearing
impairment were comparable to the SEMs observed for the
whole group at this level. Approximately 30-60% of these
subjects displayed excessive test-retest differences for SPIN
and DSI tests when scores for Trial 2 were compared to Trial
3. Thus, the poor reliability of the DSI observed here does not
appear to be related to the sensation level of the speech
materials.
The high and significant correlations observed between
scores on both the SPIN and DSI tests and peripheral
hearing is also of concern if the tests are to be used to
assessing "central" aspects of speech-recognition ability in
elderly people, as others have suggested (J. Jerger et al.,
1989; J. Jerger et al., 1990). Clearly, peripheral hearing loss
affects performance for both the DSI and the SPIN.
Caution should be exercised when using the SPIN and DSI
tests for purposes of differential diagnosis in elderly listeners,
given the poor reliability for these tests with our elderly
listeners. For example, after evaluating the performance of
many elderly individuals on the SPIN or DSI tests, some
researchers have suggested that large numbers of elderly
TABLE 3. At 85 dB SPL, means (M) and standard deviations
(SD) of SPIN and DSI scores and corresponding standard
deviations expressed in percent correct for a subgroup of 9
individuals with mild hearing impairment.
Trial 1
Test
PL
PH
DSI-L
DSI-R
Trial 2
Trial 3
M
SD
M
SD
M
SD
52.4
24.0
36.5
25.9
42.7
30.8
94.8
15.4
86.2
74.7
86.7
22.8
21.6
22.8
75
86.2
92.6
25.8
18.1
12.7
84.9
97.4
21.3
5.6
Note. SPIN = Speech Perception in Noise test; PL = predictabilitylow sentences; PH = predictability-high sentences; DSI = Dichotic
Sentence Identification test; L = left ear; R = right ear.
659
individuals display CAPD that exist in isolation and occur
more frequently than general cognitive decline. J. Jerger and
colleagues (J. Jerger et al., 1989; J. Jerger et al., 1990)
reported that 40-50% of those over 50 years old have CAPD.
J. Jerger and colleagues employed both the DSI and the
SPIN tests to determine the presence or absence of CAPD.
The test-retest differences seen for our subjects for the DSI
and SPIN tests were large enough to result in changes in
diagnostic disposition within a given individual across trials.
That is, an individual's score on one trial was deemed
abnormal, whereas a score on a subsequent trial was
considered within normal limits. The excessive variability and
large test-retest differences observed for our elderly subjects
on the SPIN and DSI tests call into question the usefulness of
these tests for making categorical determinations or in assessing speech-recognition capability in elderly people. Further investigation of appropriate speech materials for use
with elderly people should be undertaken.
Acknowledgments
This research was supported in part by the National Institute of
Aging. We are grateful to Jeffrey Cokely, Diane Van Tasell, Jill
Preminger, and an anonymous reviewer for their comments on an
earlier version of this manuscript.
References
American National Standards Institute. (1969). Specifications for
audiometers (ANSI S3.6-1969, R-1970). New York: ANSI.
American National Standards Institute. (1977). Criteria for permissible ambient noise during audiometric testing (ANSI S3.1-1977).
New York: ANSI.
Bergman, M., Blumenfeld, V. G., Cascardo, D., Dash, B., Levitt,
H., & Margulies, M. K. (1976). Age-related decrement in hearing
for speech: Sampling and longitudinal studies. Journal of Gerontology, 31, 533-538.
Bilger, R. C. (1984). Manual for the clinical use of the revised SPIN
test. Champaign, IL: University of Illinois.
Bilger, R. C., Nuetzel, J. M., Rabinowitz, W. M., & Rzeczkowski,
C. (1984). Standardization of a test of speech perception in noise.
Journal of Speech and Hearing Research, 27, 32-48.
Committee on Hearing, Bioacoustics, and Biomechanics
(CHABA). (1988). Speech understanding and aging. Journal of
the Acoustical Society of America, 83, 859-893.
Dubno, J. R., & Dirks, D. D. (1983). Suggestions for optimizing
reliability with the synthetic sentence identification test. Journal of
Speech and Hearing Disorders, 48, 98-103.
Dubno, J. R., Dirks, D. D., & Morgan, D. E. (1984). Effects of age
and mild hearing loss on speech recognition in noise. Journal of
the Acoustical Society of America, 76, 87-96.
Fifer, R. C., Jerger, J. F., Berlin, C. I., Tobey, E. A., & Campbell,
J. C. (1983). Development of a dichotic sentence identification test
for hearing impaired adults. Ear and Hearing, 4, 300-305.
Humes, L. E., Espinoza-Varas, B., & Watson, C. S. (1988). Modeling sensorineural hearing loss: I. Model and retrospective evaluation. Journal of the Acoustical Society of America, 83, 188-202.
Jerger, J., & Hayes, D. (1977a). Diagnostic speech audiometry.
Archives of Otolaryngology, 103, 216-222.
Jerger, J., & Hayes, D. (1977b). Hearing and aging. In S. S. Hans &
D. H. Coons (Eds.), Special senses in aging. Ann Arbor, Ml:
University of Michigan Press.
Jerger, J., & Jerger, S. (1974). Auditory findings in brainstem
disorders. Archives of Otolaryngology, 99, 342-350.
Jerger, J., Jerger, S., Oliver, T., & Pirozzolo, F. (1989). Speech
660
Journal of Speech and Heanng Research
understanding in the elderly. Ear and Hearing, 10, 79-89.
Jerger, J., Oliver, T., & Pirozzolo, F. (1990). Impact of central
auditory processing disorder and cognitive deficit on the SelfAssessment of Hearing Handicap in the elderly Journal of the
American Academy of Audiology, 1, 75-80.
Kalikow, D. N., Stevens, K. N., & Elliott, L. (1977). Development of
atest of speech intelligibility innoise using sentence materials with
controlled word predictability. Journal of the Acoustical Society of
America, 61, 1337-1351
National Center for Health Statistics (NCHS). (1977). Prevalence
of selected impairments. DHHS Pub. No. 81-1562. Washington,
DC: U.S. Government Printing Office.
Noffsinger, D., & Kurdziel, S. A. (1979). Assessment of central
auditory lesions. In Rintelmann (Ed.), Hearing assessment. Baltimore: University Park Press.
Orchik, D.J., & Burgess, J. (1977). Synthetic sentence identification as a function of the age of the listener. Journal of the American
Auditory Society, 3, 42-46.
35
654-060
June 1992
Shirinian, J. M., & Arnst, D. J. (1982). Patterns in the performance
intensity functions for phonetically balanced word lists and synthetic sentences in aged listeners. Archives of Otolaryngology,
108, 15-20.
Speaks, C., & Jerger, J. (1965). Method for measurement of speech
identification. Journal of Speech and Hearing Research, 8, 185194.
Studebaker, G. A. (1985). A "rationalized" arcsine transform. Journal of Speech and Hearing Research, 28, 455-462.
Received June 13, 1991
Accepted October 7, 1991
Contact author: Carol Geltman Cokely, Indiana University, Department of Speech and Hearing Sciences, Bloomington, IN47405.