A STUDY OF AUDITORY DIFFERENTIAL
SENSITIVITY T0 INTENSITY IN CHILDREN
AND ADULTS

Thesis for the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
ROBERT M. SCREEN
1968

 

 

 

”.m- ‘1‘“

LIBR A i’ 3"
Michigan Stat-e 3
University ~

    

This is to certify that the

thesis entitled

A STUDY OF AUDITORY DIFFERENTIAL
SENSITIVITY TO INTENSITY IN CHILDREN
AND ADULTS

presented by

ROBERT M. SCREEN

has been accepted towards fulfillment
of the requirements for

Audiology and
_B.h_JL_ degree in _3 Beech SCiences

I
Date _JU_1L-

I
0-169

 

 
  

 
 
   
       

BINDING BY ‘
HDAE & suus'
aunt mom mc.

LIBRARY BINDE RS
IMIII'DI'I. Ilcllm

L '

ABSTRACT

A STUDY OF AUDITORY DIFFERENTIAL SENSITIVITY TO
INTENSITY IN CHILDREN AND ADULTS

by Robert M. Screen

The purpose of this research was two—fold. The
first purpose was to determine if the Short Increment
Sensitivity Index (SISI) Test as now used in diagnostic
audiology could be employed with children five years of
age and above. The second purpose was to compare the size
of the intensity difference-limen (DL) in adults and child—
ren by using the quantal psych0physical method.

Seventy-two subjects in six age groups were
selected for this study. One group was composed of twelve
adults. The other groups consisted of twelve five-year
old, twelve seven-year old, twelve nine-year old, twelve
eleven-year old, and twelve thirteen—year old children.
Each group consisted of 50 per cent males, and was equally
divided into either an "Average IQ" or a "High IQ" category.

The intelligence quotients of the adult subjects
were determined by administering an abbreviated form of the
Wechsler Adult Intelligence Scale (WAIST. The intelligence
quotients of the children were determined by administering

Form A of the Peabody Picture VOcabulary Test. The scores

Robert M. Screen

achieved by adults and children were employed to place
each subject into the average or high IQ category.

Each subject was given a conventional pure tone
hearing evaluation in a commercial, sound—treated room.
Pure tone auditory thresholds were obtained for the fre—
quencies 500, 1000, and 4000 Hz by air conduction, and 500
and 4000 Hz by bone conduction.

Once the subject's auditory threshold had been
determined with conventional audiometry, the SISI test was
then administered. The test was presented at 4000 Hz at a
20 dB sensation level. At this frequency and sensation
level, each subject was tested with the increments of 0.50,
0.75, 1.0, 1.25, 1.50, 1.75, and 2.0 dB. The order in
which the increments were presented was randomized. After
the seventh increment presentation, an increment was
selected to repeat the test at a level closest to the sub—
ject's 50 per cent score in order to ascertain reliability.

The Method of Least Squares was used to calculate
each subject's DL from the derived data. The significance
of differences among mean DLs and mean SISI scores at the
1 dB increment was determined with three-dimensional
analyses of variance. Duncan's New Multiple Range Test
was applied to both analyses in order to make individual
comparisons among means. Reliability was determined by
employing a correlation coefficient and the standard error

of measurement (SEm).

Robert M. Screen

The following conclusions were drawn: (1) Children
nine, eleven, and thirteen years of age could be tested
with the Short Increment Sensitivity Index (SISI) Test,
and when situations warrant it, this test should be used
more frequently with children of these age levels; (2) A
modification should be made of the original SISI test in-
structions when employing the test with five and seven-
year old children; (3) Greater care should be taken with
the conditioning procedure when the SISI test is employed
with five and seven-year old children; (4) The sex of a
subject apparently does not affect his ability to perform
on the Short Increment Sensitivity Index (SISI) Test; (5)
Children and adults with IQ scores higher than 90 can per—
form on the Short Increment Sensitivity Index (SISI) Test;
(6) The criterion score for a clinically positive SISI
score for children five years of age and above should be
80 per cent; (7) The DL values obtained when using the
procedure in this study do not appear different from the
DL values reported in the literature using other methods,
and the quantal procedure is thus recommended for DL
testing; (8) Above the age of five, there are no statist-
ically significant differences in DLs as a function of age
when using the quantal psychOphysical method; (9) The sex
of a child or adult apparently does not affect his ability

to perform on the DL test procedure used in this study;

Robert M. Screen

(10) Children and adults with IQ scores higher than 90
seem to perform adequately on the DL test used in this

study and produce reliable DL results.

A STUDY OF AUDITORY DIFFERENTIAL SENSITIVITY TO

INTENSITY IN CHILDREN AND ADULTS

BY

Robert M. Screen

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Audiology and Speech Sciences

1968

 

ACKNOWLEDGMENTS

This research was supported in part by a pre-
doctoral Rehabilitation Research fellowship from the Office
of Vocational Rehabilitation.

Special gratitude is given to Dr. Edward J. Hardick,
Associate Professor of Audiology and Speech Sciences, for
the interest he di3played and the guidance he gave as thesis
advisor.

Recognition is given to Mr. Paul Niswander, graduate
assistant in Audiology and Speech Sciences, Dr. Bradley
Lashbrook, Assistant Professor of Audiology and Speech
Sciences, and Dr. Joseph Saupe, Director of Institutional
Research, Michigan State University,for their assistance.

Recognition is also given to Dr. Herbert J. Oyer,
Chairman of the Department of Audiology and Speech Sciences,

for the interest he displayed as academic advisor.

ii

TABLE OF CONTENTS

ACKNOWLEDGMENTS O O O O O O O O O O O O O O O 0

LIST OF TABLES O O O O O O C O O O O O O O O 0

LIST OF

FIGURES O O O O O O O O O O O O O O O 0

LIST OF MPENDICES O O O O O O O O O O O O O 0

CHAPTER

I

II

III

IV

INTRODUCTION 0 O O O O O O O O O O O 0

Purpose of the Study . . . . . . . .
Importance of the Study . . . . . . .
Definition of Terms . . . . . . . . .
Limitations of the Study . . . . . .

REVIEW OF THE LITERATURE . . . . . . .

Methods Employed in DL Measurements .
The Short Increment Sensitivity Index

(SISI) Test . . . . . . . . . . . .
Findings of DL Measurements . . . . .

sumary O O O O O I O O O I O O O O 0

EXPERIMENTAL PROCEDURES . . . . . . . .

subj eats O O O O O O O O O O O O O 0
Equipment and Sta dardized Tests . .

Peabody Picture Vocabulary Test . . .
The Abbreviated WAIS . . . . . . . .
Procedure . . . . . . . . . . . . . .
The Quantal Theory . . . . . . . . .
Analysis of the Data . . . . . . . .

RESULTS AND DISCUSSION . . . . . . . .

The Intensity Difference-Limen (DL) .
The SISI Score at the 1 dB Increment
Reliability . . . . . . . . . . . . .
Discussion . . . . . . . . . . . . .
Summary . . . . . . . . . . . . . . .

Page

ii

vi

vii

[—4

oombm

10
ll

18
27
33

36

36
4O
45
46
48
53
57

60

60
72
79
8O

CHAPTER Pag e

V SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS . . 86

smary 0 O O O O . O I O O I O O O O O O O O 8 6
Conclusions . . . . . . . . . . . . . . . . 88
Recommendations for Further Research . . . 89

BIBLIOGRAPHY . O O O O . O O O O O O O O O O O O O O O 92

APPENDICES O O O O O O O O 0 O O O O O O O O O O O O 97

iv

LIST OF TABLES

Table
1 Mean IQ Scores for Each Group of Subjects .
2 Mean DLs and Standard Deviations by Age,
Sex, and IQ for the Six Age Levels . . .
3 Summary of Analysis of Variance Comparing
Performance on the DL Test as a Function
of Age, Sex, and IQ . . . . . . . . . . .
4 Duncan's New Multiple Range Test Applied
to the Differences Between DL Means . . .
5 Ranges of SISI Scores at the 1 dB Increment
Showing Percentage Score Response by
Number of Subjects at Each Age Level . .
6 Mean Percent SISI Scores at the 1 dB
Increment and Standard Deviations
by IQ, Sex, and Age . . . . . . . . . . .
7 Summary of Analysis of Variance Comparing
Differences of Age, Sex, and IQ and
SISI Score for the 1 dB Increment . . . .
8 Duncan's New Multiple Range Test Applied
to the Differences Between SISI Means
at the 1 dB Increment . . . . . . . . . .

Page

39

69

70

71

73

76

77

78

LIST OF FIGURES

Schematic Diagram of the Testing Rooms

and Placement of Equipment . . . .

The Quantal and Phi-Gamma Functions

The Mean Percentage ReSponse at Each dB
Increment Level and the Mean DL for

Five-year Old Subjects . . . . . .

The Mean Percentage Response at Each
dB Increment Level and the Mean DL
for Seven-year Old Subjects . . .

The Mean Percentage Response at Each
dB Increment Level and the Mean DL
for Nine-year Old Subjects . . . .

The Mean Percentage Response at Each
dB Increment Level and the Mean DL
for Eleven—year Old Subjects . . .

The Mean Percentage Response at Each
dB Increment Level and the Mean DL
for Thirteen-year Old Subjects . .

The Mean Percentage Response at Each

dB Increment Level and the Mean DL
for Adult Subjects . . . . . . . .

vi

Page

43

56

62

63

64

65

66

67

LIST OF APPENDICES

Appendix Page
A Computed DL Values Per Subject . . . . . . 97
B SISI Scores at the 1 dB Increment Per
Subject . . . . . . . . . . . . . . . . 99
C Intelligence Quotient Scores Per
subject 0 O O O O O O O O O O O O O O O 101

vii

CHAPTER I

INTRODUCTION

It has become evident that in recent years, while
there have been significant advances made in differential
diagnosis in audiology, very little has been done in this
area with children. This research is concerned with one
of the tests used in differential diagnosis in audiology,
the Short Increment Sensitivity Index (SISI) Test, and in
determining whether or not this test, as currently admin-
istered, can also be successfully employed with children.

It has been pointed out by a number of investiga-
tors that there are certain kinds of hearing losses that
are apparently accompanied by a very keen differential
sensitivity to intensity. The employment of difference-
limen (DL) tests, therefore, has contributed to a deter-
mination of the site of lesion in hearing losses. The
Short Increment Sensitivity Index (SISI) Test employs one
of the techniques of DL testing-~the quantal psychOphysical
method. Yet, there seems to be no report in the literature
of the values of the difference-limen (DL) for normal
hearers obtained with the quantal psychoPhysical method.
Even where difference-limen values have been found with

normal hearing adults, the variations in these values were

so great that investigators found it difficult to evaluate
the results.1 The relationship of the difference—limen
for adults and children, as well as a comparison of their
performance on the SISI test as currently administered,

were important concerns of this research.

Purpose of the Study

 

The major purpose of this investigation was to
determine if the Short Increment Sensitivity Index (SISI)
Test as currently employed in diagnostic audiology can be
employed with children from the ages of five through
thirteen. Specifically, this research was concerned with
the increment amplitude now employed in the administration
of the SISI test, and the size of the difference-limen
(DL) for intensity using the quantal psychophysical method.
A question posed was whether the presently used increment
amplitude of 1 dB presented at a 20 dB sensation level can
be used with children with normal hearing, and produce
results similar to those of adults with normal hearing.

Another major concern of this research was the
comparison of the difference-limen (DL) in adults and
children using the quantal psychOphysical method. The

literature has stressed the contribution made by DL testing

1Lauritz Lund-Iversen, "An Investigation on the
Difference-Limen Determined by the Method of Lfischer-
Zwislocki in Normal Hearing and in Various Forms of Deaf-
ness," Acta Oto-Laryngologica, 42 (1952), pp. 219-223.‘

 

to a determination of the site of lesion in hearing losses.
Yet, the literature does not show evidence of systematic
research with any specific DL procedure for adults or
children. A question posed, therefore, was whether the

DLs of children with normal hearing differed significantly
from the DLs of adults with normal hearing. This research
also investigated whether there was any relationship be-
tween IQ scores at a given age level and a subject's DL
score as well as his performance on the SISI test at the

1 dB increment. Consideration was also given as to whether
or not the sex of a subject affected his performance on the
DL test and the SISI test using a 1 dB increment.

A final goal, of which this research is just a
beginning, is to determine if acoustically handicapped
children will re5pond to both the SISI test and the DL
procedure used in this study. Before such a goal can be
accomplished, however, it should be determined whether or
not these procedures are feasible with a normal hearing
p0pu1ation of children. A review of the literature ind-
icates an obvious dearth of information in this regard.

In order to gain information concerning the above
questions, the following null hypotheses were formulated:

1. There is no significant difference between the
responses of normal hearing adults and children
between five through thirteen years of age to the

1 dB increment currently employed in the Short
Increment Sensitivity Index (SISI) Test.

There is no significant difference in the difference-
limen (DL) of normal hearing adults and that of

children between five through thirteen years of age.

There is no significant difference in the DL and

SISI scores obtained employing the 1 dB increment
between subjects who have an average intelligence

quotient and those who have a high intelligence
quotient for each of the six age levels tested.

There is no significant difference in the DL and

SISI scores obtained employing the 1 dB increment

between male and female subjects at each of the
six age levels tested.

Importance of the Study

A review of the literature reveals that the great
majority of research studies concerning standardized audio-

metric procedures have utilized adults as samples.

The
response of children to conventional pure-tone audiometry

has been studied and reported extensively in the literature,

as well as the response of children to unconventional audio-
metric procedures. With regard to Special tests used in

differential diagnosis, however, very little, and in some

cases nothing at all, is reported concerning the reSponse
of children to these procedures. Among the special tests
used in differential diagnosis, Bekesy audiometry with

children has been investigated most frequently in recent

years, and this has not been done systematically. In this
same regard, almost nothing is reported concerning the use
of the SISI test with children. Further, there is a paucity
of information regarding the difference-limen (DL) for

children, since most studies concerned with difference-limen

data have used adults as subjects. In 1952, Jergerl
reported the DL for 39 normal ears for a 15 dB sensation
level, but the quantal psychophysical method was not used
in establishing the DL in his study, and there is no
Specific reference to the ages of the subjects. In the
1959 study by Jerger, et. a1.,2 at which time the SISI
procedure was first reported in the literature, there was
no concern for the DL per se, but only with finding an
increment size which would successfully dichotomize normal
ears from ears with cochlear pathology.

Since 1960, the SISI test has been a valuable
clinical tool in helping the audiologist to localize the
site of damage to the cochlea. Indeed, children as well
as adults suffer hearing problems caused by damage to the
cochlea. It is unfortunate, however, that attempts have
not been made to utilize it in differential diagnosis of
children. It would seem, however, that as a clinical tool
for localizing the site of damage to the cochlea, the SISI
test would certainly be as important for children as it is
for adults. A parallel can be drawn in this regard to the

DL values for normal hearing children and adults, then

. 1James F. Jerger, "A Difference-Limen Test and its
Diagnostic Significance," Laryngosc0pe, 62 (1952), pp.
1316-1332.

2James F. Jerger, Joyce Lassman Shedd, and Earl
Harford, "On the Detection of Extremely Small Changes in
Sound Intensity," AMA Archives of Otolaryngology, 69 (1959),
pp. 200-211.

 

these values would begin to have more meaning when compared
to those of the hearing impaired. Only with this kind of
concerted research effort can we ascertain the diagnostic

significance of DL values.

Definition of Terms

The following definitions are used in this invest-
igation.

The Short Increment Sensitivitinndex (SISI) Test.—-
A test which differentiates between patients with normal
hearing or middle ear lesions, sensory, or VIIIth nerve
disorders. The patient is presented with the desired con—
tinuous discrete frequency at a 20 dB sensation level.
Every five seconds the intensity is increased 1 dB for 0.2
second. Whenever the patient hears the intensity increase,
he pushes the signal key. Twenty increases are presented
in a series. The test is scored in terms of the percentage
of the number of intensity increases that are perceived.

The:guantal PsychOphysical Method.—-One of the
variations of the Constant Methods. It derives from an
hypothesis by Stevens, Morgan, and Volkmanl that increments

on a sensory-response continuum are not essentially

 

1S. S. Stevens, C. T. Morgan, and J. Volkmann,

"Theory of the Neural Quantum in the Discrimination of
Loudness and Pitch," The American Journal of Psychology,
54 (1941), pp. 315-335.

continuous but occur in small units or quanta. The
detection of a change in making comparative judgments
depends upon this unit or quantum. The standard stimulus
SS arouses a given number of quanta of neural activity.
It also provides varying amounts of a surplus that is not
sufficient in itself to arouse one more quantum, but in
conjunction with an increment in the stimulus S, it may
arouse one more quantum. As the surplus or residual excita—
tion varies, it requires a stimulus change of varying
amounts to complete the extra quantum and thus give rise
to a reportable change.

Difference-Limen (DL) for Intensity.--In any sen-

 

sory process a difference-limen is defined as a "just
noticeable difference" (jnd) in whatever aspect of the
sensation is under investigation. The DL for intensity

is the amount of change in intensity required to produce a
jnd in loudness.

Sensation Level.--Indicates the number of decibels

 

that a sound is above the threshold of audibility of a
Specific ear.

Average IQ.--In this study average IQS are defined

 

as intelligence quotients from 90 to 109 obtained with the
Peabody Picture Vocabulary test or the Wechsler Adult
Intelligence Scale (WAIS).

High IQ.—-In this study, an intelligence quotient
of 125 or higher obtained with the PPVT, or an intelligence

quotient of 120 and above obtained with the WAIS.

Conventional Audiometry.--Indicates the use of a

 

pure-tone audiometer in order to obtain a person's hearing
threshold for Specific frequencies. In this study, the
procedure was to have the subject listen for pure tones
through earphones and he was asked to indicate when he
heard the tone by depressing a signal key, and taking his

hand off the key when the tone disappeared.

Limitations of the Study

 

Several limitations were introduced into the study
in order to make it practicable. First, it was felt that
the children, in order to qualify as subjects, should
demonstrate the ability to respond to conventional audio-
metry. Certainly, if the child could not respond to con-
ventional audiometry he would not be able to rSSpond to
the task required to obtain his DL and SISI scores.
Therefore, if the investigator was unable to obtain a
reasonable auditory threshold by a conventional pure tone
test, the child did not qualify as a subject.

A second limitation was that all subjects were
required to have normal hearing, as indicated by conven-
tional audiometry. This was done to eliminate clinical
variables, such as contralateralization, which may arise
when testing an individual with pathological hearing.

Another limitation was imposed by the desire to

establish a clear relationship between intellectual ability

and the ability to perform the task required for this
study. In other words, children having an IQ score between
110 and 124, and below 90, did not qualify for this study.
Also, adults having an IQ score between 110 and 119, and
below 90, did not qualify for the study.

A final limitation was imposed by the physical
parameters actually involved in the SISI and DL tasks,
i.e., the frequency, sensation level, and the increment
magnitudes investigated. This study confined itself to an
investigation of the SISI scores at 4000 Hz and at a 20 dB
sensation level. In order to compute the DL, the increment
magnitudes investigated were 0.50, 0.75, 1.0, 1.25, 1.50,
1.75, and 2.0 dB. The selection of 4000 Hz and the 20 dB
sensation level was made for two reasons. First, the in-
vestigator wanted to approximate, as closely as possible,
the manner in which this test is given clinically so that
some comparison of results could be made. In the second
place, time would not permit the investigation of SISI
scores at more frequencies and sensation levels, although
it is hOpeful that future research will study responses at
other frequencies and sensation levels. The particular
increment magnitudes studied were also selected because
the investigator wanted a basis of comparison with the
increment magnitudes that had been investigated previously

with adults.

CHAPTER II

REVIEW OF THE LITERATURE

The review of the literature concerning this study
will include the pertinent investigations of differential
sensitivity to intensity change. Specifically, the dis-
cussion of these investigations will view (1) the manner
in which the difference-limen (DL) is measured, (2) the
psychOphysical methods employed, and (3) some of the re-
ported results of these investigations with normal hearers.
A further concern of this chapter will be the research
employing the Short Increment Sensitivity Index (SISI) Test
with normal hearing subjects. Both concerns of this chap-
ter should provide a general overview and a firm groundwork
for the understanding of the present study for two impor-
tant reasons. First, an investigation of the literature
in this area reveals a dearth of information showing the
results of the SISI test with children. Secondly, little
is apparently known of the normal DL, either with adults
or children, when the quantal psychOphysical method is

used.1

1Letter from James Jerger, Director of Research,
Houston Speech and Hearing Center, Houston, Texas, Dec. 4,
1967.

10

11

Methods Employed in DL
Measurements

The smallest detectable change in the intensity of
a tone determines the intensive differential sensitivity
of the ear. Stevens and Davis1 state that this sensitivity
may be strictly defined as the reciprocal of the just-
noticeable change, or DL (difference-limen).

According to Dallos and Carhart,2 the work of
Fechner some hundred years ago paved the road to the appre-
ciation of the difference—limen (DL) for intensity change
as a psychoPhysical indicator of the listener's experience
in the loudness domain. A great deal of research has been
devoted to such tasks as: (1) testing the validity of
E. H. Weber's original concept that the DL is prOportional
to presentation level; (2) ascertaining whether loudness
increase can be considered a simple cumulation of the sub—
jective magnitudes of the DL'S; and (3) clarifying the
effects on the DL of modifying the stimulus frequency, the
presentation level, and the pattern for modulating the
intensity fluctuation of the stimulus. This work has
yielded a general understanding of the nature of the

differential sensitivity to intensity change in the normal

 

1S. S. Stevens and Hallowell Davis, Hearing: Its
Psychology and Physiology (New York: John Wiley and Sons,
Inc., 1938), p. 136.

2Peter J. Dallos and Raymond Carhart, "Cumulation
of DL's for Intensity Change at Low Sensation Levels,"
figurnal of the Acoustical Society of America, 35 (1963),
pp. 848-855.

12

ear. It has also led experimental psychologists to view
meticulous measurement of DL's as an unproductive and
somewhat Spurious approach to defining the basic phenomenon
of the loudness experience.

One of the major concerns of most investigators of
differential sensitivity to intensity change is the manner
in which the difference-limen (DL) is measured. H. C.
Montgomeryl states that in measuring differential intensity
sensitivity a number of different factors which may influ—
ence the results must be taken into account. As is the
case with nearly all psychOphysical measurements, account
must be taken of variations among individuals who may be
used as subjects. The physical characteristics of the
sound used as a stimulus, such as the frequency, intensity,
and harmonic composition for a musical tone, must be
specified. Furthermore, Montgomery states that the follow-
ing elements are influential upon the results of testing
differential sensitivity to intensity change: monaural
versus binaural observation, duration of tones, transition
between tones, number of comparisons, control of instant

of presentations, and the type of judgment required.

 

1H. C. Montgomery, "Influence of Experimental
Technique on the Measurement of Differential Intensity
Sensitivity of the Ear," Journal of the Acoustical Society
9§ America, 7 (1935), pp. 39-43.

13

Dimmick and Olson1 are critical of Montgomery's
claim that differences in the technique of measuring
differential sensitivity are directly reflected in wide
divergencies of the results. While they appear to agree
with his conclusion, they are dissatisfied that Montgomery's
considerations of "type of judgment" in his own experiments
makes no mention of the standard psychOphysical methods.

In other words, they feel that Montgomery's measurements
of the detection of changes produced by the observer him-
self even defy simple naming, and they do not believe that
any uniformity or stability of results can be eXpected
until a standard procedure is adOpted. Support for this
point of view is found in Jerger,2 who states that the size
of the BL is extremely dependent on the method by which it
is measured. For this reason, the absolute sizes of the
DLS can be compared only if they have been obtained by the
same method.

Lfischer and Zwislocki3 employed a DL test that was

administered at a sensation level of 40 dB. It involved

lForrest L. Dimmick and Ruth M. Olson, "The Inten-
sive Difference Limen in Audition," Journal of the Acoust-
ipal Society of America, 12 (1941), pp. 517-525.

2James F. Jerger, "DL Difference Test," AMA
figchives of Otolaryngology, 57 (1953), pp. 490—500.

3E. Lfischer and J. Zwislocki, "A Simple Method for
Indirect Monaural Determination of the Recruitment Phenome-
non (Difference Limen in Intensity in Different Types of
Deafness)," Acta Oto-larypgology Supplement, 78 (1949),
pp. 156-168.

14

the presentation of a tone that "wobbulated" in intensity,
that is, it varied in intensity so the patient heard intens-
ity beats. The examiner then reduced the wobbulation
gradually until the patient reported the tone sounded
"steady." The amount of intensity variation occurring at
the point at which the patient signaled the tone was no
longer fluctuating is his DL.

Denes and Nauntonl differed from Lfischer and
Zwislocki in that no attempt was made to interpret the
absolute size of the DL. Denes and Naunton were interested,
however, in the difference in the value of the DLs at two
sensation levels. Their technique involved a comparison
of the size of a patient's DL at sensation levels of 4 and
44 dB. In determining the size of the DL at each sensation
level they used two separate tones of the same frequency,
one tone held constant in intensity and the other varied.
These tones were presented to the patient's ear alternately
with an interval of Silence between the two. In the be-
ginning the tones were of equal intensity, but as the test
progressed, one tone was varied until the patient reported
he could detect a difference in intensity between the tones.

The DL was defined as the intensity difference between the

_ 1P. Denes and R. F. Naunton, "The Clinical Detec-
tlon of Auditory Recruitment," Journal of Laryngplogy and
Otology, 64 (1950), pp. 375-398.

15

two tones when the variable tone is just perceptibly
"louder than" or "softer than" the fixed tone.

Dimmick and Olson1 gave their subjects a period of
training in loudness judgments with the psychOphysical
method of limits. After the training period, the psycho-
physical method of constant stimuli was followed in which
seven stimulus intensities were compared with the standard
in random order. The judgments of loudness difference
between the two tones was not based on change within a
continuous stimulus, but on the perception of difference
in intensity of two separate tones.

Knudsen2 used as a source of sound a telephone
receiver actuated by a current from a vacuum tube oscill-
ator. The intensity or the pitch could be changed period-
ically, once a second or so, by automatically changing the
resistance in the oscillating circuit. The method of
observation, then, was to change AE or AN continuously
until the threshold of perception of fluctuation was
reached. Separate observations usually checked within 10

per cent.

 

lDimmick and Olson, "Intensive Difference-Limen,"
pp. 519-520.

2Vern O. Knudsen, "The Sensibility of the Ear to
Small Differences of Intensity and Frequency," Physical
Review, 21 (1923), pp. 84-102.

l6

Lidén and Nilssonl used the method for determination
of difference-limen prOposed by Lfischer and Zwislocki. The
method was tried with normal individuals as well as with
patients suffering from various forms of deafness. The
variations in the values of difference—limen found, however,
were so great that it was difficult to evaluate the results.

In Jerger's2 original difference-limen test, a 15
dB sensation level was chosen. This was done because in-
vestigation showed that levels below 15 dB were too low
for accurate judgment by normal hearing subjects; there-
fore, 15 dB was chosen as the sensation level at which to
test.

In the test administration the subject was given a
steady tone, which was gradually changed to a beating tone.
He was to raise his fingers as soon as he heard the beat.

AS soon as he signaled, the beat was taken out, and a re-
trial was done. It was done three or four times in this
manner to make sure the patient did not make any mistakes.
After two practice trials, the test was begun. There were
three or four trials made at each frequency. The mean
value of the trials was taken as the difference-limen, or

DL for that frequency. Frequencies tested in order were

lGunnar Lidén and Gunnar Nilsson, "Differential
Audiometry," Acta Oto-lapyngplogica, 38 (1950), pp. 521-527.

2James F. Jerger, "A Difference-Limen Recruitment
Test and Its Diagnostic Significance," Laryngosc0pe, 62
(1952), pp. 1313-1332.

 

17

1000, 2000, 4000, 500, and 250 Hz, so as to maintain

consistency with common audiometric procedure.

Jergerl later develOped a test procedure in which

the DL was measured at two levels above threshold accord-

ing to the Denes and Naunton method, but in which these

DLS were actually obtained by the Lfischer technique. In

this method, two DLS were obtained: one at 10 dB, the

other at 40 dB above the patient's threshold. These DLS

were measured in the following manner. When the patient's

threshold at a given frequency had been determined audio-

metrically, the hearing loss dial of the audiometer was
set to a level 10 dB above threshold. Then, with the tone

"continuously on," the periodic variation was gradually

introduced until the Subject signaled his awareness of an

intensity beat. The average of three to five trials was

recorded as the patient's DL at the frequency under test.

This procedure was then repeated at a level 40 dB above

the patient's threshold. The difference between the two

DLS (DL at 10 dB minus DL at 40 dB) was recorded as the

"DL Difference" for the frequency under test.

 

lJerger, "DL Difference Test," pp. 493-494.

18

The Short Increment Sensitivity
Index (SISIY Test

The procedure of particular concern for this study

was introduced by Jerger, Shedd and Harfordl in 1959.

These authors pointed out that while there was disagree-

ment among early investigators in their attempts to eXplain

the observations of abnormally acute differential sensit-

ivity in terms of the loudness recruitment phenomenon, this

did not alter the significance of the observation that

certain patients (generally those with loudness recruitment)
were often able to detect smaller intensity changes than
normal ears at comparable levels above threshold. Recog-
nizing that a patient's intensity difference-limen might
be entirely normal when defined by a methodology not in-
volving sustained stimulation (e.g., the method of constant
stimulus differences), the authors supposed that patients
with cochlear lesions might manifest behavior simulating
abnormally keen differential sensitivity only when the
measurement is performed in a particular and unique manner.
The task, then, was that of designing a new proce—
dure with two goals in mind: (1) to employ a stimulus
whose temporal pattern would not require a beating

judgment; (2) to make the test procedure as objective
as possible by limiting the number of decisions required

of the tester. ‘

 

1James F. Jerger, Joyce Lassman Shedd, and Earl
Harford, "On the Detection of Extremely Small Changes in
Sound Intensity," AMA Archives of Otolaryngology, Vol. 69

(Feb., 1959), pp. zoo-211.

19

To meet these goals, the basic temporal pattern
used in the quantal psychophysical method was adOpted.
In this method, short amplitude increments are super-
imposed on a signal of constant amplitude at relatively
widely Spaced intervals. The patient's task is to
respond whenever he hears a momentary change in the
loudness of the signal. The tester's task is merely
to record the presence or absence of a response to the
presentations of each increment. The method is charact-
erized by (1) use of relatively sustained stimulation
over time; (2) a simplification of the patient's task
be complete avoidance of a "beating" judgment; and (3)
simplification and objectification of the tester's role
in the procedure.

In the present test sequence an increment occurred
once every five seconds. Each increment rose to a
maximum amplitude in 50 msec., remained at maximum
amplitude for 200 msec., then decayed to the steady-
state level in 50 msec. Thus, the duration of the
increment was exactly 1 dB.

The initial design of the procedure was guided by
the desire to construct a task so difficult that sub-
jects with normal differential sensitivity could not
perform it with any degree of success. To this end,
an increment magnitude of 1 dB and a presentation level
10 dB above the subject's threshold were intially
chosen. Subsequent experience suggested, however, that
10 dB may have been unnecessarily low. Even at levels
of 30 to 40 dB, the task is sufficiently difficult, and
the extreme faintness of the 10 dB tone makes the pro—
cedure somewhat frustrating for patients with a conduc—
tive loss. In view of these considerations, a sensation
level of 20 dB was later adOpted. The difference between
results obtained at the two levels was found to be min-
imal.

All results have been eXpressed in terms of the
percentage of 1 dB increments to which a correct
response was made. The resultant score is termed a
”Short-increment sensitivity index" (SISI). Thus, a
Patient who responded to 10 of the 20 increments re-
ceived a SISI score of 50%, etc.1

Results of the above investigation reveal that, in
general, purely conductive losses yield very low scores,
while losses presumed to be localized in the sensory

structure of the inner ear tend to show very high scores.

lIbid., p. 203.

20

In order to draw a specific dichotomy regarding
characteristic responses on the SISI test, it would be
important, indeed, to know what kinds of responses are
made by normal hearers to this procedure. Further, such
a consideration is necessary because of inter- and intra—
subject variability which could exist among normal hearing
subjects and thereby affect conclusions about the responses
to the SISI test of a patient with a pathologic ear.

Jergerl presented a general discussion of the SISI
test as one of a battery of useful techniques in otologic
diagnosis. He mentioned that his eXperience indicated
scores between 0 and 20 per cent for those with normal
hearing, with conductive losses, and with VIIIth nerve
involvement, and scores between 60 per cent and 100 per
cent (at frequencies above 4000 Hz) for patients with
cochlear pathology.

Yantis and Decker2 administered the SISI test to
normal hearers at 500, 1000, 2000, and 4000 Hz, and at
intensity increments of l, 2, 3, and 4 dB. These authors
also found the mean scores to be below 20 per cent. The

mean SISI scores become progressively greater with increased

1James F. Jerger, "Hearing Tests in Otologic
Diagnosis," AMA, 4 (1962), pp. 139-145.

2Phillip A. Yantis and Robert L. Decker, "On the
Short Increment Sensitivity Index (SISI Test)," Journal of
§peech and Hearing Disorders, 29 (1964), pp. 231-246.

 

21

intensity of the automatic tone pulse, and the average
normal ear is least sensitive to the 1 dB increment. At
4000 Hz, however, the range of scores was found to be 0 -
85 per cent; seven of the 25 subjects tested scored higher
than 20 per cent at this frequency. The authors concluded
that their data indicate that an intensity of 1 dB at the
sensation level (20 dB) employed in the SISI test is suff-
iciently small in amplitude to make this increment size
relatively difficult to detect by the normal ear, at least
as compared to increments of 2, 3, and 4 dB. However,
sensitivity to amplitudes even of this small size tends to
increase in the average normal ear with increasing fre—
quency.

Hanley and Uttingl investigated the SISI scores
of normal hearers at 4000 Hz, and at increment magnitudes
of 1 dB, 0.75 dB, and 0.50 dB. For the 1 dB increment the
mean SISI score for normals was 42 per cent, with one-third
of the 45 subjects scoring as high as 60 per cent. For
the 0.75 dB increment, the mean SISI score for normals was
only 15 per cent and with only two of the 48 subjects
scoring as high as 60 per cent. Even for the 0.50 dB
increment the mean SISI score for normals was 5.4 per cent.

The authors suggest that if a SISI score of 60 per cent or

 

lClair N. Hanley and Jack E. Utting, "An Examina-
tion of the Normal Hearer's Response to the SISI," Journal
9: Speech and Hearing Disorders, 29 (1964), pp. 231—246.

22

higher is to be accepted as indicative of cochlear
pathology, the use of a 0.75 dB increment rather than a

1 dB increment might more definitely isolate cochlear in-
volvement from other types of pathology.

Owensl introduced another variable in studying the
normal hearer's response to the SISI test. He studied the
effect of changing the sensation level at which the test
is administered. Owen's concern was not the mean score
nor the range of scores at different sensation levels, but
only determination of the lowest presentation level at
which subjects could score 60 per cent or higher. At a
20 dB sensation level only one of the 27 normal hearing
subjects was able to meet the 60 per cent criterion and
only at one frequency, 4000 Hz. At this sensation level
there was characteristically no response. At 40 dB sensa-
tion level, three subjects were able to achieve a score of
60 per cent or better at 500 Hz, four subjects at 1000 Hz,
and one subject at 4000 Hz. According to his data found
for normal hearers, Owens concluded that a 25 dB sensation
level may be employed for the SISI test with good assurance

that a positive score will indicate cochlear involvement.

 

1Elmer Owens, "The SISI Test and VIIIth Nerve Ver-
sus Cochlear Involvement," Journal of Speech and Hearing
Disorders, 30 (1965), pp. 252+262.

 

23

One of the more recent investigations of the
response of normal hearers to the SISI test was done by
Blegvad.l In this study the test was presented at 250,
1000, and 4000 Hz, and at 10 dB, 20 dB, and 40 dB sensa—
tion levels. At each individual frequency and sensation
level the subjects were tested with increments of 0.50,
1.0, 1.50, 2.0, 2.5, 3.0, 3.5, and 4.0 dB. The order in
which the different frequencies and sensation levels were
presented was randomized, as was the order of the differ-
ent increment magnitudes. The results of Blevgad's study
indicate that at low sensation levels scores increase with
frequency. At 40 dB sensation level the scores are grossly
independent of frequency. AS a result, the increase in
score with sensation level is most marked at low frequency.
The data in Blevgad's study agree quite closely with that
of Yantis and Decker,2 both for the 1 dB and 2 dB incre-
ments. On the other hand, Blevgad's values are signifi-
cantly lower than those obtained by Hanley and Utting.3

This applies to the mean score of 1 dB, as well as to the

number of subjects with scores that are equal to or greater

 

lB. Blevgad, "The SISI Test in Normal Listeners,"
Acta Oto-laryngologica, 62 (1966), pp. 201-212.

2Yantis and Decker, "On the Short Increment," p. 235.

3Hanley and Utting, "Normal Hearer's Response,"
p. 62.

24

than 60 per cent. In these studies the test was always
presented at a 20 dB sensation level. There is only fair
agreement with the Owens1 study with regard to the number
of subjects scoring 60 per cent or higher. At a 40 dB
sensation level there was a tendency for Owens' subjects
to score better at 1000 Hz than at 4000 Hz, which was the
Opposite case in the Blevgad study. The author concluded
that his data allow him to confirm Jerger's statement that
scores for 1 dB increments are generally below 20 per cent,
but in agreement with Yantis and Decker he found that about
four subjects were able to score higher at 4000 Hz, and two
subjects had scores equal to or greater than 60 per cent.
For 1.50 dB increments the number of subjects with positive
scores increased to 14, and the mean scores increased from
10 per cent to 40 per cent. Blevgad believed that although
the results of his study might have been influenced by a
certain practice effect, they do tend to imply that even
minor changes of increment magnitude will cause substantial
changes in SISI scores.

Sanders and Simpson2 tested the prOposal of Hanley

and Utting that the SISI test should employ an increment

 

lOwens, "The SISI Test," p. 226.

2Jay W. Sanders and Mary E. Simpson, "The Effect
of Increment Size on Short Increment Sensitivity Index
Scores," Journal of Speech and Hearing Research, 9 (1966),
pp. 297-304.

25

magnitude of 0.75 dB rather than the 1.0 dB increment
originally prOposed by Jerger, Shedd, and Harford. The
SISI test was given with three increment magnitudes, 1.0
dB, 0.75 dB, and 0.50 dB, to a group of normal hearing
subjects and to a group of subjects with cochlear lesion
hearing loss. The findings of this study disagreed with
the results of Hanley and Utting. The mean scores obtained
with the 0.75 and 0.50 dB increments in the two studies
were not Significantly different at the 5% level. At the
0.50 dB increment the mean score reported by Hanley and
Utting was 5.4% and 1.8% by Sanders and Simpson. At the
1.0 dB increment, however, the mean score reported by
Hanley and Utting was 42.4% as compared to 19.4% by Sanders
and Simpson. The mean scores reported at the 0.75 dB in-
crement were 15.5% by Hanley and Utting and 9.2% by Sanders
and Simpson.

The major conclusion of this study was that the
SISI test should be continued with the 1.0 dB increment
magnitude originally prOposed. The test results indicate
that the 1.0 dB increment is large enough to reveal coch-
lear pathology when it exists, yet small enough to prevent
Spuriously high scores in ears with normal cochlear struc—
tures. Morever, the findings presented in this study
suggest the possibility that the SISI test with the 1.0

dB increment might have sufficient sensitivity to detect

26

the presence of subliminal cochlear pathology in at least
some ears having pure-tone thresholds within normal limits.

The investigations just discussed are some of the
more pertinent ones with regard to the normal hearer's
response to the SISI test. They do indicate some intra-
subject variability which does not appear to be significant.
In general, however, they tend to agree with Jerger's
original statements regarding the parameters of increment
magnitude, sensation level, and the resultant SISI score
with normal hearers. On the other hand, these studies are
also evidence of the lack of information on this procedure
with children, which again demonstrates the need for
systematic research in this area.

The review of the literature thus far has estab-
lished a relationship between the SISI scores for normal
and pathologic ears, to the extent that normals, for the
most part, are not eXpected to achieve "positive" SISI
scores using this procedure. This research, then, would
like to know what would be this relationship if the SISI
test were administered to children. It is also apparent
that the investigations of the SISI test thus far discussed
have not been concerned with the DL per se, but rather with
finding an increment Size which would successfully dichoto-
mize normal ears from ears with cochlear pathology. Thus,
another question becomes of important concern: what is

the relationship between normal and pathologic ears that

27

can be seen from the actual DL values reported in the
literature? And, of course, this research would like to
know in what way the DL values of children compare with
those of adults, and what is the significance of this

comparison.

Findings of DL Measurements

 

A number of writers have claimed that increased
differential sensitivity (decreased DL size) occurs when
there is pathology of the cochlea, and therefore, that
measurement of the intensity DL in clinical situations can
serve as a major diagnostic indicator. Most evidence tends
to support this contention, though some investigators have
argued the Opposite. According to Dallos and Carhart,l
this consequent uncertainty has encouraged otologists and
clinical audiologists to treat the intensity DL as a sub-
ject for arguments rather than as a phenomenon which can
be explored to practical advantage.

There are discrepancies from one study to another
in the size of the DLS reported in the literature. It
should be remembered, however, that experimental conditions,
variations among subjects, and differences in psychophysical
method all influence the absolute size of the just notice—

able difference (jnd). Because of the number of factors

 

lDallos and Carhart, "Cumulation of DLS," p. 850.

28

that influence the measurement of intensity change, some
discrepancies would occur even if the psychOphysical method
were standardized from one study to another.

Hirsh, Palva, and Goodman1 cite a study by Doerfler
who compared the differential sensitivity of a group of
normal subjects with that of a group of patients with sen-
sorineural deafness, the latter group consisting of many
types of hearing loss that were non-conductive and pre-
sumably including both recruiting and nonrecruiting ears.
He employed the Sinusoidal modulation of Riesz by putting
two tones together, and showed that the DL for patients was
always between 0.5 and 0.1 dB lower than that of the normal
subjects.

Montgomery2 showed that the DL for one normal sub—
ject in a two-tone comparison was 0.8 dB for a 1000 Hz,

40 dB tone when the two tones were separated by an interval
of 0.5 second. When this interval was eliminated, the DL
decreased to 0.6 dB. When more than one two-tone compari-
son was allowed, the DL decreased to 0.4 dB. When the
subject himself controlled a switch that turned one or the
other tone on, the DL went down to 0.2 dB. This study by
Montgomery is a demonstration of the dependence of the

difference-limen upon eXperimental technique.

 

lMontgomery, "Influence of EXperimental Technique,"
pp. 42—43.

2Jerger, "A Difference-Limen," pp. 1321-1325.

29

Jergerl obtained the DL for 39 normal ears at a
15 dB sensation level. The figures listed by Jerger are
average DLS for each of the frequencies tested. At 250
Hz the average DL was 2.0 dB; at 500 Hz, 1.8 dB; at 1000
Hz, 1.7 dB; 2000 Hz, 1.7 dB; and at 4000 Hz, 1.6 dB. A
comparison of these DLS is made with those with pathologic
ears. For those patients with conductive losses, the
average DL at 250 Hz was 2.0 dB; at 500 Hz, 2.0 dB; 1000
Hz, 1.9 dB; 2000 Hz, 1.9 dB; and 4000 Hz, 1.8 dB. As for
the sensorineural group, all cases showing a sensorineural
component were included, whether there was a conductive
component or not. The DLS exhibited by these cases, at
the frequencies at which there was a sensorineural compo—
nent, were divided into three categories: abnormally small
DLS, normal DLS, and abnormally large DLS. At 250 Hz, the
average DL for the abnormally small DL was 1.0 dB; the
normal DL was 2.0 dB, and the abnormally large DL was 4.2
dB; at 500 Hz, the DL values were 0.9 dB, 1.8 dB, and 4.2
dB; at 1000 Hz the DL values were 0.9 dB, 1.8 dB, and 3.6
dB; at 2000 Hz, the DLS were 0.9 dB, 1.8 dB, and 4.3 dB;
at 4000 Hz the DLS were 0.9 dB, 1.6 dB, and 3.3 dB. The
ages of the subjects used to obtain these values, and the

values stated by Montgomery, were not Specified.

 

lJerger, "A Difference-Limen," PP- 1321‘1325-

30

The author states that these findings Show that in
pure conductive hearing loss the BL is normal, indicating
the absence of recruitment, but in sensorineural loss the
DL may be abnormally small, indicating recruitment, or
normal, indicating the absence of recruitment. It was
pointed out further that a small number of cases tested
showed abnormally large difference limens. In each one of
these cases there was sufficient concomitant evidence to
lead to the tentative conclusion that the patient's appar—
ent hearing loss was not on a peripheral organic basis.

No Specific conclusions regarding this apparent relation-
ship were drawn.

Harrisl used the forced—choice method in studying
the intensity difference-limen. Two tones, each 0.5 second
in duration were presented monaurally, with a 40—msec.
rise-fall time, and a 0.5 second inter-stimulus interval.
The subject was requested to press a silent microswitch
whenever he felt his physiological noise level was low and
he was in a maximally receptive attitudinal condition. He
was forced to judge the second tone "louder" or "softer."
If in doubt, he was allowed to press the switch at liberty
until a judgment was forthcoming. Each variable was pre-

sented 75 times, over a period of months to wash out

 

1J. Donald Harris, "Loudness Discrimination," JSHD
Hgnograph Supplement, 11 (1963), p. 6. _

31

incidental factors. Each DL75% was thus derived from a
total of 900 judgments. The frequency 1000 Hz was studied
with four normal hearing subjects at sensation levels of
5, 10, 20, 40, 60, and 80 dB. The results of this study
indicated that under Optimal conditions the human ear can

yield a DL of the order Of 0.5 dB and less over a major

75%
portion of the auditory area, and is thus almost completely
independent of frequency and loudness.

Stevens and Davis1 cite the eXperiments of Riesz
in measuring differential sensitivity to intensity. Riesz
presented his tones monaurally by means of a special moving
coil receiver designed to be especially free of distortion.
The receiver was connected to the outputs of two oscill-
ators in such a way that both oscillators activated the
receiver simultaneously and produced beats when the fre—
quencies of the two impressed tones were close together.
First, the tone from one oscillator was presented at a
definite sensation level, and then the intensity of the
tone from the other Oscillator was increased, from a point
near zero, until the observer was just able to detect a
beat. From the intensities of sound needed to obtain this
heat, the intensity at the maximum and at the minimum of

the beat could be calculated. The difference between the

maximum and the minimum was taken as defining the DL.

 

1Stevens and Davis, Hearing, p. 137.

32

The size of the DL was found to be a function of
the rate of the fluctuations in intensity. A representa-
tive curve showing the size of the relative DL as a func—
tion of the rate at which the beats were presented is
shown by the author.1 It is characterized by a broad
minimum in the neighborhood of 3 cycles, and this rate was
adOpted for the experimental determination of the DL for
intensity.

Average curves giving the size of the relative DL
as a function of intensity (sensation level), with fre-
quency as a parameter, are also presented by the author.2
At a given frequency the relative difference-limen approaches
a constant value for intensities above 50 dB, but increases
rapidly as the intensity is reduced toward the auditory
threshold.

The author presents another curve3 that shows the
behavior Of the relative DL as a function Of frequency for
different values of the parameter intensity. The relative
DL is a minimum at a frequency Of about 2500 Hz, although
the minimum is less sharply pronounced at high intensities

than at low. The region of the greatest differential

 

lIbid.

 

2Ibid., p. 138.

31bid., p. 139.

33

sensitivity Of the ear corresponds to the frequency range
of greatest absolute sensitivity.

The values for the DL in Riesz' study are not
cited, but they can be plotted from the above-mentioned
curves in Stevens and Davis.1 Although there are varia-
tions in these studies, the reasons which have been pre-
viously stated, there does seem to be a basic similarity
that unites all these studies, and this similarity can be
summed up in the following conclusions:

1. The differential sensitivity of the ear in-

creases with the intensity at the same time

as its dependence upon the frequency decreases.
2. Differential sensitivity is a minimum as a

function of frequency at about 2500 Hz. The

minimum is less sharply pronounced at high

intensities than it is at low intensities.

3. At any given frequency the differential sen-

sitivity approaches a constant value for high
intensities but increases rapidly as the inten-

sity is reduced toward the auditory threshold.

Summary

It is evident that the literature does contain

reports on the methods of DL measurements and their findings.

 

lIbid., pp. 137-139.

34

Nevertheless, the audiologist is still confronted with a
major problem when he looks at these findings with DL
measurements and attempts to interpret them clinically.
For example, what is the significance of a normal DL of
1.8 dB at 400 Hz, and how much would this DL vary with
psychOphysical method, with frequency, and sensation level?
An even more important concern is the comparison of this
DL with that of a pathologic ear. If some systematic work
is not done in this area, there is no way of knowing how
to interpret results obtained from DL measurements. In-
deed, these measurements would have little significance.
Another concern for the audiologist is the relationship
of the DL for children to that of adults. Certainly, the
audiologist's role with children cannot be de-emphasized,
and children suffer the same kinds of hearing problems as
do adults. Unfortunately, however, there is little evi-
dence of extensive work on DL measurements with children.
It can be said that standardization of methods used to
Obtain DL measurements is needed in order to give the
results of these measurements greater significance clin—
ically. In other words, extensive work Should be done
employing particular psychOphysical methods. In this way,
some basis would be established for classifying a DL as
normal or abnormal, as small or large.

The SISI test also suffers from a lack of systematic

investigations of the responses Of children. This seems

35

unfortunate for two reasons. First, at the present time
the SISI test has gained considerable use as an important
diagnostic instrument in audiology. Both in the United
States and EurOpe, different makes of SISI equipment are
now available. Secondly, all audiologists are well aware
of the need for improving diagnostic measures used with
children. It is not enough to assume that, on the basis

of past experience with conventional audiometry and other
special tests of hearing, children cannot be expected to
perform well on the SISI test. Because the SISI test is

an important diagnostic instrument, the capacity of its
employment with children should be thoroughly investigated.
Audiologists should know at what ages children can and
cannot be expected to perform on this test. Additional
data on the effects of a child's Intelligence Quotient
(IQ) and sex on his performance would give a more complete
picture to the investigation of children's responses to

the SISI test.

CHAPTER III

EXPERIMENTAL PROCEDURES

This chapter presents information regarding the
selection Of subjects, their assignment to categories by
age, sex, and intelligence quotient, and a discussion of
the testing procedures employed. In brief, a total of 72
subjects were used in the study. Twelve of the subjects
were adults, 50 per cent Of whome were males. These sub-
jects were equally divided into average and high IQ groups.
The remaining 60 subjects were children, ranging in age
from five to thirteen years. These subjects were divided
into five groups, consisting of twelve five, seven, nine,
eleven and thirteen-year olds. Each group was equally
divided according to sex, and was further equally divided
into average and high IQ groups. All subjects were ad-
ministered a conventional pure tone hearing test and the
Short Increment Sensitivity Index (SISI) Test, administered

at seven intensity increments.

Subjects

Two groups of subjects participated in the study.
One group was composed of twelve adults, ranging in age

from 19 to 21 years, with a mean age of 19.9 years.

36

37

Subjects were equally divided as to sex, and were also
equally divided with reference to IQ into "Average IQ" or
"High IQ" categories. A second group of subjects was com—
posed Of 60 children, ranging in age from five years to
thirteen years, eleven months. They were sub-divided into
five equal sized age groups (5 years-0 months to 5 years-
11 months, 7 years-0 months to 7 years-ll months, 9 years—
0 months to 9 years-ll months, 11 years-0 months to 11
years-ll months, and 13 years—0 months to 13 years-ll
months of age) and each group consisted of a 50 per cent
male and a 50 per cent female pOpulation. These subjects
were similarly divided into either an "Average IQ" or "High
IQ" category.

In order to Obtain IQ scores for the children, the
investigator administered Form A of the Peabody Picture
Vocabulary Test. This test was used because it can be ad-
ministered easily and quickly, and it has good demonstrated
validity. Scores from the PPVT were used to place subjects
into an average or high IQ category. The average IQ cate-
gory consisted Of those subjects scoring between 90 and 109,
and the high category consisted of subjects scoring 125 and
higher. Those subjects whose scores fell between 110 and
125, as well as those whose scores were below 90, were not
accepted for the study. If any other intelligence test
information was available, i.e., from the subject's school

folder, it was used as a validity check on the PPVT results.

38

Such information was available on eight of the 5-year old
subjects who met the IQ criteria for the study. None of
the eight subjects had to be rejected because of a Signif-
icant discrepancy between the test scores. These eight
subjects had previously been evaluated with the Wechsler
Pre-School and Primary Scale of Intelligence (WPPSI).

To determine the IQ scores of adult subjects, the
investigator administered an abbreviated form of the
Wechsler Adult Intelligence Scale (WAIS), which will be
discussed later in this chapter. This abbreviated form
of the WAIS was selected because it permits an accurate
estimate of the Full Scale score in 35 to 40 minutes of
testing. For the adult age group in this study, the
abbreviated form of the WAIS which was used has demonstrated
a correlation of 0.96 with the Full Scale score. Scores
from the WAIS were also used to place subjects into an
average or high IQ category. The average IQ category
consisted of subjects scoring between 90 and 109, and the
high IQ category consisted of subjects scoring 120 and
above. Those subjects whose scores fell between 110 and
119, as well as those whose scores were below 90, were not
accepted for the study. No other intelligence test inform-
ation was available on these subjects. The mean IQ score

for each group of subjects is illustrated in Table l.

39

TABLE l.--Mean IQ scores for each group of subjects

 

 

 

AVERAGE IQ Group HIGH IQ Group

Male Female Male Female
5 years 107 106 133 129
7 years 104 103 138 139
9 years 98 102 138 133
11 years 102 96 133 132
13 years 106 103 134 127
ADULTS 96 96 154 134

 

The subjects were Obtained from various sources.
Twenty per cent Of the children were Obtained from the
Michigan State University Laboratory Preschool and the
Spartan Nursery Of Michigan State University. The remain-
ing 80 per cent of the subjects were children of professors
and other graduate students living in the University commun—
ity. Some of the children in this 80 per cent category were
obtained from the investigator's friends who taught in the
Wainwright and Pleasant View Schools of Lansing, Michigan.
Even though there were sixty children employed in this
study, a larger number had to be contacted. Some Of the
children did not fit into either of the IQ categories de—
sired for the study, and a few did not meet the criteria
for normal hearing. NO child had to be eliminated from the
study because he could not attend to the task, i.e., taking
the earphones off or refusing to participate.

The adult subjects were undergraduate students at

Michigan State University and young adults outside the

40

college pOpulation who were contacted through friends and
acquaintances Of the investigator. As was the case with
the children, a large number of adult subjects had to be
contacted even though only twelve were needed for the
study. It was difficult to find undergraduate students
who scored in the "Average IQ" category on the WAIS. This
was true even of so-called "average" students who had been
recommended to the investigator by graduate assistants and
graduate advisors in the dormitories. Therefore, a non-
college pOpulation had to be used in order to fulfill this
category. In addition, several of these subjects did not
meet the hearing criteria and had to be eliminated from
the study.

A subject was not included in the study if he
revealed an auditory threshold which indicated an average
loss (500, 1000, and 4000 Hz) of 10 dB or greater (1964
ISO standard). Likewise, any subject whose conventional
pure tone test showed an air-bone gap was excluded from the
study. The pure tone test was administered at frequencies
of 500, 1000, and 4000 Hz for air conduction testing, and
at frequencies of 500 and 4000 Hz for bone conduction

testing.

Eguipment and Standardized Tests

The equipment listed below includes the major

instruments employed for this study.

41

Clinical Audiometer (Beltone, model 15—C)
Earphones (Telephonics, model TDH - 39)
Earphone cushions (model MX - 41/ AR)

Bone Vibrator (Radioear, Model B70-A)

SISI Adapter (Stowe, model 1259)

Sound Level Meter (Bruel and Kjaer, model 2203)
Artificial Ear (Bruel and Kjaer, model 4152)
Artificial Mastoid (Beltone, model M5B)

OSCillOSCOpe (Model 502A with Polaroid Camera
attachment)

Commercial Test Room (Industrial Acoustics Company,
Inc., model 10-1052)

Peabody Picture Vocabulary Test

Wechsler Adult Intelligence Test

The sound-treated room and all audiometric equip-
ment were located in the Speech and Hearing Clinic in the
Michigan State University Auditorium building. The ambient
noise of the sound-treated room had been previously
measured with the sound level meter on the C scale. This
noise level was found to be 42 decibels SPL. Also, an
octave analysis of the ambient noise level in the sound-
treated room had been done previously, and the results
indicated that the greatest amount of ambient noise (40 dB
average) was found in the octave bands below 100 Hz. For
those octave bands from 100 to 8000 Hz, the ambient noise

level averaged 14 dB SPL.

42.

The SISI adapter was located in an adjoining test
room which communicates with the sound-treated room by a
window and a two-way electronic communications system.
This adapter was designed as a "plug-in" unit for the pure
tone audiometer to allow the administration of the SISI
test. With this self-contained unit, extensive and perma-
nent modifications Of the audiometer circuit are not
necessary for the administration of the test.1 The audio-
meter used with the SISI adapter in this study was the
Beltone, model 15-C. The testing rooms and equipment were
situated as schematically diagrammed in Figure 1.

According to the manufacturer's specifications,
the Stowe SISI Adapter permits an intensity increment of
controlled magnitude with a 50 msec. rise-decay time, a
duration of 200 msec at peak intensity, and a total dura-
tion of 300 msec. These were the specifications Of the
SISI stimulus described by Jerger, Shedd, and Harford.l
Upon measuring these increments for this study, this parti—
cular adapter was found to provide an increment with a 50
msec rise-time and a 25 msec decay-time. The duration at
peak intensity was 200 msec, and the total duration was
275 msec. These measurements were made at regular intervals

with an OscillOSCOpe, model 502A, with a Polaroid camera

1Instruction Manual for the SISI Adapter, Model
1259 (Northbrook, Illinois: Gordon N. Stowe and Associates).

2
p. 202.

Jerger, Shedd, and Harford, "On the Detection,"

43

Dawsoocam can mEOOH mcflumwu ogu mo Scum

.ucmfimflnwm mo

 

 

 

a mi pmumumfleHSpe
a mMS B>mm OHO£3
muwmco pee wanes

pep owumswromii.a musmflm

 

IHHmmI AII

 

 

I

. HOOD .

 

 

 

mambo
.uomnnsm

D

_m

Uima Mom
mwconmnmm Q

ATmuooo
6.

 

mooa..
~33me /

 

 

 

 

 

 

uflmno
Hocwemxm

 

Enom
ucoemflnvm

 

 

\ \

44

attachment. Initially, the equipment did not produce a
Signal that would meet the specifications required for this
study. After a number of instrumental modifications and a
series of measurements, the above-stated Specifications
were reached. Once these Specifications were obtained,
three additional measurements were made; one at the beginn-
ing of the study, one at the mid-point of the study, and a
final measurement at the end of the study. It was found
that these increment magnitudes remained the same through—
out the study.

Calibration Of the SISI stimulus was made each day
of testing. This was done by placing the earphone over
the artificial ear assembly, which was attached to the
sound level meter. With the SISI stimulus turned on, the
increment magnitudes were read and measured from the sound
level meter. Calibration of the air conduction stimulus
from the Beltone 15—C was also done each day subjects were
seen. This was done by connecting the earphone to the
artificial ear assembly, which in turn was connected to the
sound level meter. The Beltone lS-C audiometer was then
set to produce a 60 dB Signal, and the output was measured
at octave intervals between 250 and 8000 Hz. The earphone
used in this study was a TDH-39 earphone in an MX 41/ AR
cushion. Only one earphone was used in the study, and the
same earphone was used for all subjects. The earphone

covering the Opposite ear was employed to keep the headset

45

in place on the subject's head.

Peabodnyicture Vocabularerest

 

The Peabody Picture Vocabulary Test is an intelli—
gence test designed to give an estimate of a subject's
verbal intelligence.1 The materials consist of 150 plates
contained in a spiral-bound booklet, each plate containing
four pictures. The test has been standardized to provide
norms for ages two years-six months through eighteen years.
Two equivalent forms of the test, A and B, are available,
each with a separate table of norms.

Various validity results are presented in the
manual. Some of the tests with which the PPVT has been
correlated are the Stanford Binet, Columbia Test of Mental
Maturity, Wechsler Intelligence Scale for Children, and
the California Test of Mental Maturity. As presented in
the manual, these correlations are 0.88, 0.82, 0.86, and
0.82.

The Wechsler Adult Intelliggnce
ScaIe (WAIS)

 

The Wechsler Adult Intelligence Scale (WAIS) is
designed to measure adult intelligence. It is an exten~

sion and modification of the Wechsler—Bellevue Intelligence

 

. 1Lloyd M. Dunn, Expanded Manual for the Peabody
Picture VOcabulary Test (MinneapoIis: American Guidance
SerVice, Inc., 19657, p. 25.

 

 

46

Scale, Form 1.1 The WAIS consists of eleven tests. Six
of these are grouped into the Verbal Scale; the remaining
five comprise the Performance Scale; all eleven tests are
combined to make the Full Scale. The Verbal tests are:
Information, Comprehension, Arithmetic, Similarities, Digit
Span, and Vocabulary. The Performance tests are: Digit
Symbol, Picture Completion, Block Design, Picture Arrange-

ment, and Object Assembly.

The Abbreviated WAIS

 

There are many instances when users of the WAIS
would want a reasonably accurate estimate of a subject's
IQ without giving all eleven sub—tests of the Scale.
DOppelt2 investigated the effectiveness of a sub-group of
tests in predicting the Full Scale Score, which is the sum
of scores on the eleven tests. One of the major problems
was the decision as to the number and type of sub-tests to
be included in the predictor group. This decision, to some
extent, was arbitrary. It was decided to select the group
of four sub-tests which correlated highest with the Full

Scale Score. Although the prediction of the Full Scale

 

lDavid Wechsler, Manual for the Wechsler Adult

Intelligence Scale (New York: The Psychological Corpora-
tion, 55), p. 4.

 

2Jerome E. DOppelt, "Estimating the Full Scale
Score on the Wechsler Adult Intelligence Scale from Scores
on Four Subtests," Journal of Consulting Psychology, Vol.
20, NO. l (1956), pp.-63-66.

 

47

Score was the goal, it was felt that the best approach
would be to select the two verbal sub-tests which were
most highly correlated with total Verbal Score and the two
performance measures which were the best predictors of the
total Performance Score.1 The four subtests selected were
Arithmetic, Vocabulary, Block Design, and Picture Arrange-
ment. The selection of these tests was also based on the
data Obtained in the national standardization of the WAIS.

In order to obtain the subject's IQ, a Simplified
regression equation was presented in which the sum of
scaled scores on the four selected subtests is multiplied
by 2.5, which is the coefficient of the predicting variable
in the regression equation. A constant is then added to
this figure, depending upon the subject's age. For the
age group used in this study, the constant was 10. This
abbreviated form permits an estimate of Full Scale Score
after 35 to 40 minutes of testing.

The author warned that in making any prediction
one should always have some idea of the error involved.2
In his study the standard deviation of Full Scale Scores
was about 25 and the correlation coefficient between the

sum of the four selected tests and the Full Scale Score

 

lIbid., p. 63.

2Ibid., p. 65.

48

was approximately 0.96. Consequently, the standard error
Of estimate in predicting Full Scale Score from the four
tests is about seven scaled score points, which is equiva-
lent to 4.2 IQ points. Thus, an estimated Full Scale
Score would be within seven scaled score points of the

actual score about 68 out of 100 times.

Procedure

 

The adult subjects were seen first in the study.
The investigator made appointments to meet each subject
either in a residence hall classroom or a classroom at the
Michigan State University Auditorium building. The non-
college subjects were seen in their homes or in the
Auditorium building. At this time, a subject was admin-
istered the abbreviated form of the Wechsler Adult
Intelligence Scale (WAIS).

Once the IQ had been determined, the subject came
to the sound-treated room in the Speech and Hearing Clinic,
and a pure tone hearing evaluation was administered. The
examiner determined thresholds for the frequencies 500,
1000, and 4000 Hz by air conduction, and 500 and 4000 Hz
by bone conduction. These thresholds were determined by

the method of Carhart and Jerger.l

lRaymond Carhart and James F. Jerger, "Preferred
Method for Clinical Determination of Pure-Tone Thresholds,"
Journal of Speech and Hearing Disorders, 24 (1959), pp.
330—345.

49

Once the subject's auditory threshold was obtained
by conventional audiometry, the SISI test was administered.
The instructions and test procedure of Jergerl were
essentially the same, with one exception. The familiariza—
tion run which precedes the actual SISI testing consisted
of five 5 dB increments and five 3 dB increments. The
investigator believed, as did Hanley and Utting,2 that ten
increments would help the subject to settle down and become
more familiar with the procedure, and reducing the famil-
iarization increment from 5 dB to 3 dB might not present
much contrast to the smaller increments of the test run.
The subject was then given the following instructions:

You will hear a steady sound through the earphone for
about two minutes. The sound will be very faint.
During the time it is on you may occasionally hear a
little jump in loudness. Whenever you are positive
that you have heard one Of these short loudness jumps,
press the button which you have in your hand. If you
think you heard a jump but you are not certain, then
do not press the button. Only press it when you are
sure you heard a jump in loudness.3

The test was presented at 4000 Hz and at a 20 dB
sensation level. It was given in only one ear, and the
ear chosen was the better of the two ears at 4000 Hz, as

determined by the pure tone hearing test. At this fre-

quency and sensation level, the subject was tested with

lJerger, Shedd, and Harford, "On the Detection,"
p. 203.

2Hanley and Utting, "An Examination of the Normal,"
p. 60.

3Jerger, Shedd, and Harford, "On the Detection,"
p. 203.

 

50

increments of 0.50, 0.75, 1.0, 1.25, 1.50, 1.75, and 2.0
dB. The order in which the increments were presented was
randomized according to a table Of random numbers.1
Following each Of the test runs at each increment, the
subject was allowed a one minute rest period before beginn-
ing a run at another increment. This was done to avoid
fatiguing the subject, and to keep him alert to the present-
ations at each increment. Randomization was also employed
to reduce the practice effect. After the seventh increment
presentation, an increment was selected to repeat the test
at a level closest to the subject's 50 per cent score for
a reliability check. Even with the reliability check, no
subject required more than 45 minutes to complete the test.
Thirty-two of the children were administered Form
A of the Peabody Picture Vocabulary Test in their homes
prior to coming to the Auditorium building for the hearing
test. The remaining twenty-eight children were administered
the test at a small table outside the sound-treated audio-
metric test room. Only 10 to 15 minutes were required to
give this untimed teSt. The scale is administered only
over a critical range of items for a particular subject.
The starting point, basal and ceiling, vary from testee

to testee.2 The starting points are listed in the manual,

lHubert M. Blalock, Jr., Social Statistics (New
York: McGraw Hill Book Company, Inc., 1960), pp. 437-440.

2

 

Dunn, Manual for the Peabody, p. 5.

 

51

and they vary as a function Of age and assumed ability.
From the indicated starting point the examiner works for-
ward until the subject makes his first error. In the event
that eight correct reSponses have not been made to this
point, the examiner goes back immediately to the starting
point and works backward consecutively until a total of
eight consecutive correct responses have been obtained.
This is the basal score. TO Obtain the ceiling score, the
examiner continues testing forward until the subject makes
six errors in eight consecutive presentations. The last
item presented is considered as the subject's ceiling.
The test is discontinued when a basal and ceiling have
been established. In order to determine the subject's
intelligence quotient, the sum of the incorrect responses
is subtracted from the ceiling score. This gives the raw
score which is used to Obtain the IQ score from the tables
in the manual.1
After the Peabody test had been given, a conven-
tional pure tone hearing evaluation was administered. As
with the adults, auditory thresholds were determined for
the frequencies 500, 1000 and 4000 Hz by air conduction,
and 500 and 4000 Hz by bone conduction. These thresholds

were also determined by the method of Carhart and Jerger.2

 

lIbid., pp. 8-10.

2Carhart and Jerger, "Preferred Method," pp. 330-345.

52

Once the subject's auditory threshold had been
determined with conventional audiometry, the SISI test was
administered. The procedure was essentially the same as
that used with adults, including the use of five 5 dB
increments and five 3 dB increments in the familiarization
run. There was a slight deviation in the manner in which
instructions were stated, however, so as to make the task
clear for the children. The subjects were given the
following instructions:

Now you will listen for tones that are just a bit
different from the first tones you heard. This time
you will hear a steady tone in your ear for a few
minutes. It will sound much like the tone you heard
in the first test. While the tone is on, however, you
will begin to hear a little jump in loudness. I want
you to listen very carefully for the jump in loudness
(the examiner often demonstrated this verbally). When
you are sure that you heard the jump in loudness, press
the button you have in your hand. Remember, press the
button only when you are certain that you heard the
tone jump. If you are not sure, do not press the
button.

There was one other deviation in the previously
outlined procedures as they applied to the children. After
the fourth increment run the child was allowed to come out
of the sound-treated room and take a ten minute break
before resuming the experiment. They were motivated to
continue by rewards which the investigator placed in their
Sight, even though most of them admitted they were tired
at the half-way point. Nevertheless, none of the children
refused to continue the experiment, and the elapsed time

required to complete the task for any child did not exceed

one hour.

53

The Quantal Theory

 

The quantal theory derives from the assumption that
the basic neural processes mediating a discrimination are
of an all-or-none character. The advantage Of a quantal
theory of sensory discrimination lies not solely in the
fact that it makes eXpliCit the role of neural processes
that are all-or-none, but also in the fact that it enables
us to predict the form and the SlOpe of certain psychometric
functions. Instead of psychometric functions resembling
the probability integral, there are rectilinear functions.
Instead of curves of unpredictable SlOpe, there are lines
whose slopes are prOportional to the differential sensit-
ivity of the observer. Instead of failure of even the
smallest stimulus-increments to produce perceptions of
increase, a critical value is found below which no incre-
ment is ever perceived.l The theoretical argument for the
quantal theory is as follows.

It is assumed that the neural structures initially
involved in the perception of a sensory continuum are
divided into functionally distinct units. Bekesy2 thought
of these units as single afferent fibers, but the evidence

indicates that the functional units are larger than fibers,

 

1S. S. Stevens, C. T. Morgan, and J. VOlkmann,

"Theory of the Neural Quantum in the Discrimination of
Loudness and Pitch," The American Journal of Psychology,
54 (1941), pp. 315—335.

2Stevens and Davis, Hearing, p. 147.

54

and that they are probably centrally located. A stimulus
of a given magnitude excites, at a particular instant, a
certain number of these quantal units, and, in order for

an increment to be noticeable, it must excite at least one
additional quantum. That is the basic picture; but here
enter some additional considerations. The stimulus which
excites a certain number of quanta will ordinarily do so
with little to spare; it will excite these quanta and leave
a small "surplus" insufficient to excite some additional
quantum. This surplus stimulation will contribute, along
with the increment AI, to bring into activity the added
quantum needed for discrimination. Consequently, at any
instant, the size of the increment necessary to add another
quantum to the total number excited must depend upon the
amount of "left-over" stimulation.

The frequency with which a given stimulus-increment
will excite an additional quantum depends upon the fre-
quency with which the surplus stimulation exceeds a certain
crucial amount, and this occurs a prOportion Of the time
which is dependent directly upon the amount to be exceeded.
From these considerations, it follows that, if the incre—
ment is added instantaneously to the stimulus, it will be
perceived a certain fraction of the time, and this fraction
is directly prOportional to the Size of the increment itself.

This argument can be rendered more precise with the

aid of mathematics. AS already stated, it is assumed that

 

55

at a given moment a steady stimulus excites completely a
certain number of quanta and leaves a small surplus, p,
which goes part way toward exciting the quantum next in
line; and that the stimulus increment, AI, which is re—
quired to complete the excitation of this quantum, is
smaller when the surplus, p, is larger. The size of a
quantum can be measured in terms of the increment, Q, which
will just succeed always in exciting it. Then the AI just
sufficient to complement the surplus, p, and thereby excite
an additional quantum is given by the equation, AI = Q - p.1
An example of the psychometric function is Shown by
the graph in Fig. 2. This graph is a plot of the equation,
R = (AI/Q - l) x 100,2 where R represents the percentage
of the increments which an 0 should be able to detect, and
R varies between zero and 100. This equation is derived
from the assumption that the addition of two quanta is
required for a discrimination. In the plot shown, the value
of Q is used as the unit for measuring the stimulus incre-
ment. In these units the slope of the straight psychometric
function is exactly determined. Furthermore, when a dis—
crimination requires the addition of two quanta, it is
noted that stimulus increments of less than one quantum are

never detected, whereas those greater than two quanta are

always detected.

lStevens, Morgan, and Volkmann, "Theory of Neural
Quantum," p. 318.

21bid.

 

56

100

 

 

A

 

 

 

 

4‘ ' . 1
o 1 2 3

 

SIZE OF INCREMENT (In Quantal Units)

Figure 2.-—The Quantal and Phi-Gamma Functions.
The straight line shows the results eXpected on the basis
of the quantal theory. It is the graph Of the equation
R = (AI/Q - 1) x 100. The S-shaped curve was constructed

by fitting the phi-function of gamma to the rectilinear
quantal function.

According to classical theory, the convention then
is to define as the difference-limen that increment which
is noticed 50 per cent of the time. It is sometimes
argued that when the difference-limen is small, the SlOpe
of the psychometric function must be steep (its b must be
high); but precisely what SlOpe is to be expected, the

theory is unable to disclose.-

 

lIbid., p. 319.

57

Analysis of the Data

 

The data derived from each subject's DL were
computed according to the formula from the Method of Least
Squares.1 A desk calculator was used for the computations.
After the computations were made, the DL data were plotted
on a composite graphy, using the line of best fit accord-
ing to the principle of least squares. The line of best
fit according to the principle of least squares is that
line from which the sum of the squares of the residuals is
a minimum. A residual is a discrepancy between an obtained
Y value and the Y value that could be predicted from its
corresponding X on the basis Of the equation which is being
used. According to Blalock,2 Yp may be used to indicate
that the Y value has been predicted from a least-squares
equation. The least-squares line, therefore, will be the
best estimate of the true regression if the regression
actually is linear.

In order to determine the significance of differ—
ences among the variables in the principal comparisons

- - - n
(H01, H02, H03, and H04), a "Three-Dimen51onal De51gn

analysis of variance was employed. The actual analysis of

 

1J. P. Guilford, Psychometric Methods (New York:
McGraw-Hill Book Company, Inc., 1954), p. 64.

2Blalock, Social Statistics, p. 284.

58

variance was conducted on a Control Data Corporation 3600
Digital Computer. The F-ratio was used in testing the
significance of the variations. Specifically, two differ-
ent three-way analyses were computed; one with the DL
score as the dependent variable, and the other with the
SISI score at the 1 dB increment as the dependent variable.
The source Of variation for each of these analyses, was
age, sex, and IQ.

After having tested the DL means and the SISI means
for each age group by an analysis of variance, Duncan's New
Multiple Range Test was applied to determine specifically
which means differed when a significant F was obtained.

The test itself was conducted on a Control Data Corporation
3600 Digital Computer, employing the program, "Calculation
of Basic Statistics on the BASTAT Routine."l

The mean DL scores for the variables of sex and IQ
were computed for each of the six age levels studied. An
examination of the variability around these means was made
by computing the standard deviation at each age level for
the variables of sex and IQ. Likewise, the mean SISI

scores at the 1 dB increment and their standard deviations

were computed for the variables of sex and IQ, at each of

the Six age levels. The means were computed on a Control

 

1Michigan State University, Agricultural EXperiment
Station, "Calculation on the BASTAT Routine, STAT Ser1es

Description #5 (March 1966).

 

 

59

Data Corporation 3600 Digital Computer. The standard
deviations were computed on a desk calculator.

The ranges of the SISI scores at the 1 dB increment
were tabulated at each age level for the variables of sex
and IQ. These ranges provided a comparison with previous
research at the 1 dB increment employing normal hearing
adults. Moreover, careful investigation of these ranges
Opened the possibility for a re-evaluation of the diagnostic
criteria previously established for the normal hearer's
response to the SISI test at the 1 dB increment.

During the DL experiment, an increment was selected
to repeat the test at a level closest to the subject's 50
per cent score in order to ascertain reliability. A
Pearson Product—Moment correlation coefficient was computed
on this test-retest data, and a standard error Of measure—
ment (SEm) was computed for the test-retest scores. A

desk calculator was used for these calculations.

CHAPTER IV

RESULTS AND DISCUSSION

This chapter is divided into four sections. The
first two sections are devoted to the presentation of the
results of the study relative to the hypotheses which were
tested. The third section presents reliability data rela—
tive to the SISI test procedure used in this study. The

fourth section presents a general discussion Of the results.

The Intensity Difference-
Limen (DL)

 

In order to gain information concerning the ques-
tions originally posed for this study, four null hypotheses
were formulated.' A re-statement of these hypotheses would
be apprOpriate at this time:

1. There is no Significant difference between the
responses of normal hearing adults and children
between five through thirteen years of age to the
1 dB increment currently employed in the Short
Increment Sensitivity Index (SISI) Test.

2. There is no significant difference in the differ-
enceelimen (DL) Of normal hearing adults and that
of children between five through thirteen years
Of age.

3. There is no difference in the DL and SISI scores
Obtained employing the 1 dB increment between sub-
jects who have an average intelligence quotient
and those who have a high intelligence quotient
for each of the six age levels tested.

60

61

4. There is no difference in the DL and SISI scores
obtained employing the 1 dB increment between male
and female subjects at each of the six age levels
tested.

Each subject's difference—limen (DL) was computed
according to the procedures outlined by Guilfordl and pre—
viously discussed on page 57. In addition, the DL was
computed for each age group. The line of best fit was ob—
tained by mathematically locating two points following the

procedures outlined in Blalock.2

Figures 3 through 8 show
the psychometric function derived from the application of
the quantal method. Each graph presents the mean response
at each increment for each age level. The resultant DL,
therefore, is the mean DL for that particular age level.
Examination Of Figures 3 through 8 indicates some
discrepancies between the visual diSplay of the mean DLS
and the mean DLS shown in Table 2. These discrepancies,
for the age groups over five years, are minor, and are
caused by the mathematical manipulation required in order
to plot the line of best fit. The mean DL for each age
group may be calculated one of two ways. By means of the

first method, the Slope and Y-intercept Of the line of

best fit for each subject are used to find the DL for that

 

1J. P. Guilford, Psychometric Methods (New York:
McGraw-Hill Book Company, Inc., 1954), p. 63.

 

2Hubert M. Blalock, Jr., Social Statistics (New
York: McGraw—Hill Book Company, Inc., 1960), p. 284.

 

% Correct Responses

 

62

[04

075‘

 

 

 

‘/.1_5' .J‘O .7: 1.0 u: up 1.7: 1.0

Increment Size in dB

Figure 3.--The mean percentage response at each
dB increment level and the mean DL
for 5-year old subjects.

% Correct Responses

 

63

  

b.--—-—----------—_-.-—-—

 

 

 

  
 
 

[.0 I43" mo 1.7! 1.0

Increment Size in dB

Figure 4.--The mean percentage response at each
dB increment level and the mean DL
for 7-year Old subjects.

% Correct Responses

 

64

  
  

aQ----——-—-----

X

41! X
/x( .m .1: 1.0 u! no Mr 1.0

   
 

 

 

Increment Size in dB

Figure 5.--The mean percentage response at each
dB increment level and the mean DL
for 9-year old subjects.

% Correct Responses

 

65

[to .1»

JV}

“if"

 

 

X

 

.00.
.1f .40 .V 1.0 m! [.30 1.11’ 2.0

Increment Size in dB

Figure 6.--The mean percentage response at each
dB increment level and the mean DL

for ll-year Old subjects._

% Correct Responses

 

66

X

 

 

.3'0 .9! no at! up mf‘ 1.0

 

Increment Size in dB

Figure 7.--The mean percentage response at each
dB increment level and the mean DL

for 13-year old subjects.

% Correct Responses

 

"O 4»

051.

.S‘D.

.th

 

.004

  

67

 

   
  

tut mu’ A0

Increment Size in dB

Figure 8.--The mean percentage response at each

dB increment level and the mean DL
for adult subjects.

68

subject. Thus, DLS may then be averaged to find the mean
DL for that group. Conversely, the SlOpes and Y—intercepts
for all subjects within a group may be averaged and the
mean DL calculated from the averaged values. It can be
Shown mathematically that these two methods are not exactly
equivalent, and may, in some cases, yield slightly differ-
ent values for the mean DL. This may also account for small
discrepancies between calculated values of the mean and
their graphical representations. In order to Show the
mean DL values graphically, however, it was necessary to
use the second method of calculation. The largest discre-
pancy appeared with the 5-year Old group, and this was
caused by the excessively large DL (12.58) on the part of
one subject, which affected the mean DL for that group.

The means used in comparing differences of age,
sex, and IQ for the DL test are presented in Table 2 along
with their respective standard deviations.

Inspection of the mean DLS in Table and in Figures
3 through 8 indicates that the "largest" DLS occur with

the five-year old children and with adults. Extreme cau-

tion should be used in applying the terms "large" DL or

"small" DL in this regard. Such caution is advisable be-

cause these terms are difficult to interpret; there is no

evidence of other research findings of the normal intens1ty

difference-limen (DL) using the procedure that was used in

this study. Until such research is available, then, these

terms would tend to have little or no clinical significance.

69

 

 

 

 

 

 

 

 

Ho.m ma.a A~.a mm.a oo.a om.m mma An
an. cmmz
Anmm.vammv.v Ammo.VAmem.v Awme.VAwmo.v Ammo.v1mom.v Aonm.VAhnv.v Ammm.VAnHm.v
va.m ma.m hm.a SN.H mm.a mm.o mm.a em.a mn.a em.H mm.a mm.m cmflm
Amam.VAan.v Avem.VAmon.v Ammm.VAmnH.v Amoo.VAeom.v Ammo.VAnmm.v Ath.VAomw.v
oo.m mn.a mn.a mm.H mm.a SH.H mm.H ha.H ov.a mm.a ma.~ ma.e 30A
.m E .m 2 .m 2 .m 2 .m 2 m 2
Dance ma Ha m h m
xmm 024 mod OH
.mao>oa mom xflm one
How 0H ocm .xom

.omm ma Amomocucoumm SH omv chHDMH>oo oumocmum one man cmozii.m mqmda

70

In order to compare the Significance of differences
as a function of age, sex, and IQ on the DL Test, a "Three-
Dimensional Design" analysis of variance was employed.1
The data from the eXperiment were entered in a three-
dimensional table. In this instance, it consisted of six
rows representing age, two columns representing sex, and
two slices representing IQ. Therefore, an A x B x C three-
entry table was employed to organize the data for the
three-factor eXperiment computation. The sources of varia-
tion for this analysis were age, sex, and IQ. A summary
of this analysis is presented in Table 3.

TABLE 3.--Summary of analysis of variance comparing per-

formance on the DL test as a function Of age,
sex, and IQ.

 

 

Source of Sum of Mean
Variance Squares df Square F-Statistic
Age 36.152 5 7.230 3.695*
Sex 2.229 1 2.229 1.139
IQ 0.011 1 0.011 0.005
Age x Sex 22.620 5 4.524 2.311
Age x IQ 1.069 5 0.213 0.109
Sex x IQ 0.300 1 0.300 0.153
Age x Sex x IQ 1.832 5 0.366 0.180
Within Cells 93.926 48 1.956 —--
(error)
TOTAL 158.141 71 —-- —--

 

iSignificant beyond the 0.01 level.

 

1E. F. Lindquist, Design and Analysis of Experiment

ipPsychology and Education (Boston: Houghton Mifflin CDT?
1956), pp._220—253.

71

Inspection of Table 3 reveals that the only factor
showing statistical Significance, at the 0.01 level of
confidence, was age. Levels within sex, levels within IQ,
and interactions, were not significant. The analysis of
variance, however, does not Show individual comparisons
among the means. It is not known, therefore, which means
differ from each other and how they differ.

In order to determine which of the differences
between means are Significant and which are not, the mean
DL score for each age level was subjected to multiple
comparisons using Duncan's New Multiple Range Test.1 The
results Of this test, at the 0.01 level of confidence, are
diSplayed in Table 4.

TABLE 4.--Duncan's New Multiple Range Test applied to the
differences between DL means. ‘

 

 

Age Groups
5 Yrs 7 Yrs 9 Yrs 11 Yrs 13 Yrs
Age and Means 3.30 dB 1.60 dB 1.25 dB 1.27 dB 1.49 dB

 

 

5 Yrs. 3.30 dB
7 Yrs. 1.60 dB 1.70 dB*
9 Yrs. 1.25 dB 2.05 dB* 0.35 dB
11 Yrs. 1.27 dB 2.03 dB* 0.33 dB 0.02 dB
13 Yrs. 1.49 dB 1.81 dB* 0.11 dB 0.24 dB 0.22 dB
Adults 2.01 dB 1.29 dB 0.41 dB 0.76 dB 0.74 dB 0.52 dB
*Significant beyond the 0.01 level.

 

 

1Allen L. Edwards, Experimental Design in Psycho-

logical Research (New York: Holt, Rinehart and Winston,
1960), pp.I136;157.

72

Inspection of Table 4 reveals the following
differences: (1) The mean DL score for five-year Olds
differed significantly from the mean DL scores of seven,
nine, eleven, and thirteen-year Old children; (2) The mean
DL scores of seven, nine, eleven, and thirteen-year old
children did not differ significantly from each other; (3)
The mean DL score for adult subjects did not differ sign-
ificantly from any of the other age groups.

The SISI Score at the 1 dB
Increment

 

 

The regular SISI test procedure as suggested by
Jerger, Shedd, and Harfordl was employed throughout this
experiment. The response at the 1 dB increment was a major
concern in this study. This concern was based largely on
the manner in which the SISI test is employed clinically.
Further, emphasis has already been given to the dearth of
information pertaining to children's responses to the SISI
test. In order to make a clinical comparison of children's
responses to adults then, it would be important to know
what these reSponseS were at the 1 dB increment.

Another factor which is important to the study of
re5ponses to the SISI test is the range Of SISI scores.

Careful inspection of ranges of SISI scores helps to

 

1James F. Jerger, Joyce Lassman Shedd, and Earl

Harford, "On the Detection of Extremely Small Changes in
Sound Intensity," AMA Archives of Otolaryngology, Vol. 69
(Feb., 1959), pP. 200-21I.

73

establish criteria for evaluating responses to the SISI
test. Though the literature has revealed cases that refute
pathological category based on their percentage SISI score,
this does not discount the value of carefully inspecting
the ranges of SISI scores for certain groups of subjects.
For the most part, clinical observation reveals that group-
ing cases on the basis of the range of their scores has
displayed an accuracy that has to be regarded with import-
ance. The ranges displayed for the six age levels in this
study are shown according to the number of subjects scoring

at the indicated percentage score levels.

TABLE 5.--Ranges of SISI scores at the 1 dB increment show-
ing percentage score response by number of sub-
jects at each age level

 

 

Percentage SISI Score

Age
Level 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 76 80

 

5 Yrs 2 l 2 l 2 2 2

7 Yrs 2 l 2 2 1 l l l l

9 Yrs 2 2 1 3 2 l 1

11 Yrs 1 1 1 1 3 l 2 l l
13 Yrs 3 l l l l l 1 l 2

Adults 1 6 2 l 1

 

74

Examination of Table 5 shows that the ranges for
the adults in this study are similar to those previously
reported in the literature.1 The highest score among the
sdults was 35 per cent, and this was scored by only one
subject. Six Of the twelve subjects scored only 5 per
cent. Inspection of the ranges of five— and seven-year
Olds shows that 11 of these 24 children scored above 30
per cent. Therefore, it would seem that the criteria
established by Jerger2 (0—30%) for the normal hearer's
response to the SISI test at the 1 dB increment should not
be applied to these age groups. The ranges of children
nine and above give further evidence that other criteria
should be employed when referring to the normal hearing
response to the 1 dB increment of the SISI test at these
age levels. Closer examination of these ranges shows that
31 of the 36 children in these three age groups scored
from 25 to 80 per cent, and 21 of these 31 children scored
from 25 to 50 per cent. These scores very closely parallel
the results of Yantis and Decker,3 who found a range of
scores for normal hearing adults to be from zero to 85 per
cent, and prOposed 80 per cent as the criterion for a

positive score.

 

lJerger, Shedd, and Harford, "On the Detection,"
pp. 200-211.

2Ibid.

 

3Phillip A.-Yantis and Robert L. Decker, "On the
Short Increment Sensitivity Index (SISI) Test," Journal of
§peech and Hearing Disorders, 29 (1964), pp. 231-246.

 

75

In order to get a comparison of SISI scores at the
various age levels, the mean SISI scores and standard de-
viations were computed for each age level by sex and IQ.

A diSplay of this comparison is shown in Table 6.

Inspection of Table 6 reveals that the mean per-'
centage SISI scores of children nine and above are higher
than the upper limit of the range for the normal hearer's
reSponse reported by Jerger.l Moreover, the mean percent-
age SISI score for eleven-year OldS is higher than the
mean percentage SISI score reported by Hanley and Utting,2
which is the highest mean percentage SISI score (42.4%) at
the 1 dB increment that has recently been reported in the
literature. The trend of these results seems to suggest
that the SISI test should be used more frequently with
children above the age of nine, and that 80 per cent should
be considered as the minimal criterion for a positive score
with these age levels.

The mean SISI scores at the 1 dB increment Shown
in Table 6 were also tested by an analysis of variance.

Again, a three-dimensional design3 was computed for the

 

lJerger, Shedd, and Harford, "On the Detection,"
pp. 200—211.

2C. N. Hanley and J. F. Utting, "An Examination of
the Normal Hearer's Response to the SISI," Journal of Speech
39d Hearing Disorders, 30 (1965), pp. 58-65.

 

3Lindquist, Design and Analysis, pp. 220-253.

 

 

 

 

 

 

 

 

 

O I
m
Am ~.p~ woe mm
wHOom
Hwnm
new:
/
fl
nom.~110m.me 1m.oaene.mno no.5H110m.mv Ion.eenom.mv 1m.mavnm. . .
op.n om.m m.mm 0.0m m.mm 0.0m o.pm m.ms m.HH p.mmo IWMSMVAM.WMO pm
we
IV.NHVAmN.mV ne.m~vno.mmv 1N.NNVAA.SHV ne.mv Av.mav In.av Am.mnv A . .
m.ma w.HH o.mm m.ma p.pm o.mm m.am m.ma m.mm m.mm m.mwvnw.wwv Son
.m 2 .m E .m z .m z .m E .m 2
pence mH an m A m
xmm oza mom OH

 

 

mom can .xmm .OH an Ammmmnusonmm
cw Omv mcowumfi>oo onmocmum cam Dawsonocfl mo H can no mouoom HmHm unwoumm cmozii.m memes

77

variables of age, sex, and IQ. A summary of this analysis
is shown in Table 7.
TABLE 7.--Summary of Analysis of Variance comparing differ—

ences Of age, sex, and IQ and SISI score for the
1 dB increment

 

 

 

 

 

 

 

Source
of
Variance Sum Of Squares df Mean Square F-Statistic
Age 10827.777 5 2165.555 6.403*
Sex 734.722 1 734.722 2.172
IQ 68.055 1 68.055 0.201
Age x Sex 498.611 5 99.722 0.294
Age x IQ 898.611 5 179.722 0.531
Sex x IQ 138.888 1 138.888 0.410
Age x Sex x

IQ 1061.111 5 212.222 0.627
Within Cells

(error) 16233.333 48 338.194 -—-
TOTAL 30461. 111 71 --- ‘ ---

 

*Significant beyond the 0.01 level.

Inspection of Table 7 reveals that the only factor
Showing statistical significance, at the 0.01 level of
confidence, was age. Levels within sex, levels within IQ,
and the interactions, were not significant. The analysis
Of variance computed on the SISI score means at the 1 dB
increment gave results similar to the analysis of variance
computed on DL means. Both showed significance only with

age.

78

Duncan's New Multiple Range Test was applied to
the SISI score means to determine if there were any signif-
icant differences between them. The results of this test,

at the 0.01 level Of confidence, are summarized in Table 8.

TABLE 8.--Duncan's New Multiple Range Test applied to the
differences between SISI means at the 1 dB incre-
ment

 

 

 

 

Age Groups
5 Yrs 7 Yrs 9 Yrs 11 Yrs 13 Yrs

Age and Means 26.2% 27.5% 40.0% 48.7% 34.5%

5 Yrs 26.2%

7 Yrs 27.5% 1.3

9 Yrs 40.0% 13.8 12.5

11 Yrs 48.7% 22.5* 21.2 8.7
13 Yrs 34.5% 8.3 7.0 5.5 14.2
Adults 9.5% 16.7 18.0 30.5* 39.2* 25.0*

 

*Significant beyond the 0.01 level.

ferences:

Examination of Table 8 reveals the following dif-

(l) The mean SISI scores for nine, eleven, and

thirteen-year Old children differed significantly from the

means of adults;

(2) The mean SISI score of eleven—year

olds differed significantly from that of five—year olds;

(3) The mean SISI score of adults differed significantly

from those of nine and thirteen—year olds;

(4) The mean

SISI scores of adults and five-year olds differed signif—

icantly from that of eleven—year olds.

79

Reliability

During the course of this study, a re-test was done
on the increment magnitude that was closest to the incre-
ment of each subject's original 50 per cent score. This
was done to ascertain reliability. A Pearson Product—
Moment correlation coefficient was computed on this test-
retest procedure.1 The correlation coefficient, however,
was applicable only to the DL test procedure, and not to
the SISI test at the 1 dB increment. Three of the subjects,
all five years of age, would not respond to re-test of the
increment because of restlessness and fatigue. The corre-
lation coefficient, therefore, was based on sixty-nine
subjects rather than the seventy-two in the original experi-
ment. The correlation was 0.57.

The correlation coefficient, however, does not
provide a measure Of a subject's absolute consistency, or
absolute variability in performance from test to re-test.
This absolute consistency is a major concern of the audiol-
ogist, and it may be expressed by the standard error of
measurement (SEm). The standard error of measurement may
be defined as an estimate of the standard deviation that

would be obtained for a series of measurements of the

1Allen L. Edwards, Statistical Methods for the
Behavioral Sciences (New York: Rinehart and Company, Inc.,
1960), pp. 145-148.

 

80

individual.1 The standard error of measurement for the
test-retest scores in this study was 8.59 per cent. This
means that for slightly more than two-thirds of the ob-
tained scores (about 68 per cent), retest scores were

within $8.59 per cent of the initial SISI score.

Discussion

 

The results of statistical analyses of the DL test
tend to support some Observations that were noted during
the experiment. For the most part, nine, eleven, and
thirteen-year old children responded to the procedure well.
They demonstrated an understanding of the procedure, and
they attended to the task consistently. The seven-year
olds, as a group, were less consistent, as some were more
difficult to condition to respond to the increment magni-
tudes than others. The majority of the five-year olds, on
the other hand, were erratic in their responses. They were
difficult to condition and their attention wandered fre-
quently. Even though they COOperated by remaining with
the task to its conclusion, they appeared restless and
bored toward the end, and many of them were curious to

know how long it would be before the task would end.

 

1Robert L. Ebel, Measuring Educational Achievement
(Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1965),
pp. 332-333.

 

81

A perusal of the literature indicated that the DL
values Obtained by other psychophysical methods do not
appear to be significantly different from the DL values
obtained by the quantal method used in this study. The
greatest difference occurs with the mean DL of the five—
year Old group, but it Should be remembered, however, that
the mean DL for five-year olds was strongly affected by
the DL (12.58) of one subject. It would seem, therefore,
that some five-year olds can be tested with the DL proce-
dure used in this study. These Observations tend to
indicate that the quantal method Of obtaining DL measure-
ments can be employed with children between five and
thirteen years Of age. When employing the procedure with
five and seven-year olds, however, a modification of in-
structions might be recommended to make sure that these
subjects understand the procedure clearly. A second
suggestion for improving the performance of these children
would be to employ greater care in familiarizing them with
the conditioning procedure. It would seem that this would
improve the consistency of their reSponseS. It was noted
during the experiment that five and seven-year olds
appeared to reSpond more consistently to the task after a
rest period., Shorter work periods might be suggested,
therefore, as another way in which the performance of these
children might be improved. Finally, rewards could be used

to motivate these children and provide additional incentive

82

for them to attend to the task to its conclusion. Rewards
were employed with all children during this research, and
it is believed that they contributed to the fact that all
of the five and seven-year olds remained with the task to
its conclusion.

The above-mentioned suggestions for improving the
performance of five and seven-year olds on the DL procedure
used in this study would also apply to improving the per-
formance of these children on the SISI test at the 1 dB
increment, which is used clinically. Since five and seven—
year olds are much more likely to be inconsistent in their
response behavior, it would seem that revising the instruc-
tions to make them more lucid, taking greater care to
familiarize the subjects with the conditioning procedure,
working in Shorter periods of time, and presenting rewards
for motivation, would tend to decrease the possibility that
these children give Spuriously high or low SISI scores at
the 1 dB increment. An example of revised instructions is
shown in Chapter III of this study. During the experiment,
the investigator did not have to repeat instructions to a
child of any IQ level in any age group. Inconsistent
reSponse patterns noted were apparently due to factors
already discussed. It was also noted that at no time
during the experiment did a child at any age level complain

that he did not understand the task.

83

The ranges of SISI test scores at the 1 dB
increment provide some observations that would tend to
have clinical importance. First, it seems apparent that
when referring to children nine years of age and above,
the criterion for a positive SISI score at the 1 dB incre—
ment should be 80 per cent, as proposed by Yantis and
Decker.1 Second, children nine and above do tend to Show
more sensitivity to small changes in intensity than do
adults, and this is further support for employing the 80
per cent criterion. Finally, children nine and above do
Show the ability to respond well to the SISI test, and it
is believed that the test should be used more frequently
in clinical situations with children at these age levels.
If the test is to be employed with five and seven—year
old children, the previously stated suggestions for improv-
ing their performance on the SISI test should be carefully

considered.

Summary

In reviewing the statistical information presented
in this chapter, it can be seen that one of the null hypo-
theses cited earlier can be rejected, whereas three of them

cannot be rejected.

 

lYantis and Decker, "On the Short Increment,"
pp. 231-246.

84

Specifically, Ho1 was concerned with the responses
to the SISI test at the 1 dB increment made by adults and
children between five and thirteen years of age. This null
hypothesis was rejected. There was a difference, and upon
statistical inSpection it was found that nine, eleven, and
thirteen-year Old subjects gave SISI scores that were
significantly different from adults. Moreover, some of
the SISI scores differed significantly from each other.

The scores of adults differed significantly from those of
nine and thirteen-year olds, and the scores of adults and
five-year olds differed significantly from those of eleven—
year olds.

The second null hypothesis (H02) was concerned with
the differences between adults and children five through
thirteen years of age and their responses on the difference—
1imen (DL) test. The results indicated that there were
significant differences between five-year olds and all the
other children's age groups. Nevertheless, there were no
significant differences between the DL Of adults and that
of the children, and the null hypothesis could not be re—
jected.

The third null hypothesis (H03) was concerned with
the difference in response to the DL test and the SISI
test at the 1 dB increment by bright and average subjects

at the six age levels studied. The results indicated no

85

Significant differences to either of these procedures as a
result of differences in intelligence quotients and the
null hypothesis was not rejected.

The final null hypothesis (H04) was concerned with
the difference in the response to the DL test and the SISI
test at the 1 dB increment by male and female subjects at
the six age levels studied. The results indicated no
significant differences to either of these procedures as
a result of differences in sex, and the null hypothesis

was not rejected.

CHAPTER V

SUMMARY, CONCLUSIONS, AND RECOMMENDATION

The purpose of this research was essentially two-
fold. The first purpose was to determine if the Short
Increment Sensitivity Index (SISI) Test as now used in
diagnostic audiology could be employed with children five
years of age and older. A question posed was whether the
presently used increment Of 1 dB presented at a 20 dB
sensation level could be used with normal hearing children
and produce results similar to those of adults with normal
hearing. Questions were also posed concerning the effect
of age, intelligence, and sex on the performance on the
SISI test. The second purpose of the research was to com-
pare the size of the intensity difference-limen (DL) in
adults and children using the quantal psychophysical method.
Consideration was also given to the effect of age, intell—
igence, and sex on the size of the difference-limen (DL)

for intensity using the quantal psychOphysical method.

Summary

Seventy-two subjects in six age groups were selected
for this study. One group was composed of twelve adults.

The Sixty remaining subjects were children of five, seven,

86

87

nine, eleven, and thirteen years of age. These subjects
were sub-divided so that twelve children were in each Of
the age groups. All groups consisted of a fifty per cent
male and a fifty per cent female pOpulation, and were
equally divided into either an "Average IQ" or a "High IQ"
category.

Each subject's auditory threshold was determined
with conventional audiometry for the frequencies 500, 1000,
and 4000 Hz by air conduction, and 500, and 4000 Hz by bone
conduction. Once the subject's auditory threshold had been
determined with conventional audiometry, the SISI test was
then administered. The test was presented at 4000 Hz and
at a 20 dB sensation level. At this frequency and sensa—
tion level, the subject was tested with increments of 0.50,
0.75, 1.0, 1.25, 1.50, 1.75, and 2.0 dB. The order in
which the increments were presented was randomized accord-
ing to a table of random numbers.

The results of this study indicated that there is
a Significant difference in the ability to respond to the
1 dB increment on the SISI test as a function of age. The
results relative to the effect of sex and intelligence on
performance on the SISI test at the 1 dB increment revealed
no significant differences. The results also indicated
that the performance of five-year olds on the DL test was
Significantly different from the performance of seven,

nine, eleven, and thirteen-year olds. The performance of

88

adults on the DL test did not provide any significant
differences between adults and any of the other age groups
studied. No Significant differences were revealed regard-
ing the effect of sex and intelligence quotient on per-.

formance on the DL test.

Conclusions

 

Within the limits imposed by the design of the
study, the following conclusions appear warranted:

1. That children nine, eleven, and thirteen years
of age can be tested with the Short Increment Sensitivity
Index (SISI) Test, and that when situations warrant it,
this test should be used more frequently with children of
these age levels.

2. That a modification should be made of the
original SISI test instructions when employing the test
with five and seven-year old children.

3. That greater care Should be taken with the
conditioning procedure when the SISI test is employed with
five and seven-year Old children.

4. That the sex of a child or adult apparently
does not affect his ability to perform on the Short Incre-
ment Sensitivity Index (SISI) Test.

5. That children and adults with IQ scores higher
than 90 can perform on the Short Increment Sensitivity

Index (SISI) Test.

89

6. That 80 per cent should be employed as the
criterion for a clinically positive SISI score for children
five and above.

7. That the DL values obtained using the procedure
in this study do not appear grossly different from the DL
values reported in the literature using other methods, and
the quantal procedure is thus a recommended procedure for
DL testing.

8. That above the age Of five, there is no
statistically significant difference in DLS as a function
of age using the quantal psychophysical method.

9. That the sex Of a child or adult apparently
does not affect his ability to perform on the DL test pro-
cedure used in this study.

10. That children and adults with IQ scores higher
than 90 would seem to perform adequately on the DL test

used in this study and produce reliable DL results.

Recommendations for Further
Research

More information should be Obtained on the normal

DL using the quantal method by expanding some of the para-

meters that were used in this study. Specifically, 0.50

dB could be eliminated because of the large number of 0

per cent responses at this increment, and additional incre—

ments above 2.0 dB could be explored. A further suggestion

would be to use finer divisions along the intenSity scale

90

rather than the divisions of 0.25 dB that were used in
this study.

This study on the Short Increment Sensitivity
Index (SISI) Test with children should be eXpanded to in-
clude other frequencies and sensation levels. Parameters
other than 4000 Hz and the 20 dB sensation level have been
reported in the literature using normal hearing adults.

A better comparison of hearing sensitivity is possible,
therefore, if additional parameters are also eXplored with
normal hearing children.

The procedure used in this study for Obtaining the
intensity difference-limen (DL) should be replicated using
larger samples of children and adults. Consideration Should
be given to wider age ranges and to IQ levels that were not
investigated in this study.

Since this study indicated that children nine,
eleven, and thirteen years of age can respond to the Short
Increment Sensitivity Index (SISI) Test, additional re-
search should be conducted to compare the responses of
children at these same age levels with cochlear and retro-
cOchlear lesions, with those of adults who also suffer
these lesions. Any additional evidence to aid in differ-
ential diagnosis with children is certainly desirable.

Once more knowledge of the normal DL is Obtained

using the procedure in this study, then research can be

91

conducted on children and adults with cochlear and
retrocochlear lesions for a comparison of results. It is
only with this kind of comparison that a DL score can
meaningfully carry the clinical label of "normal" or

"abnormal" DL.

BIBLIOGRAPHY

Books

Blalock, Hubert M. Jr. Social Statistics. New York:
McGraw-Hill Book Company, Inc., 1960.

 

Ebel, Robert L. Measuring Educational Achievement.
Englewood Cliffs, New Jersey: Prentice-Hall, Inc.,
1965), pp. 332-333.

Edwards, Allen L. Experimental Design in Psychological
Research. New York: Holt, Rinehart and Winston,
1960.

. Statistical Methods for the Behavioral Sciences.
New York: Rinehart and Company, Inc., 1960.

Guilford, J. P. Psychometric Methods. New York: McGraw-
Hill Book Company, Inc., 1954.

Hirsh, Ira J. The Measurement of Hearing. New York:
McGraw—HilITTInc.,71952.

 

Lindquist, E. F. Design and Analysis of Experiments in
Psychology afid”Education. Boston: Houghton
Mifflin Company, 1956.

 

Newby, Hayes A. Audiology. 2d ed. New York: Appleton—
Century-Crofts, 1964.

 

Stevens, S. S. and Davis, Hallowell. Hearing: Its Psycho-
logy and Physiology. New York: John Wiley Sons,
Inc., 1938.

Stevens, S. S. (ed.) Handbook of Expsrimental Psychology.
New York: John Wiley Sons, Inc., 1966.

"The Psychophysics of Sensory Function," Sensory
Communication. W. A. Rosenblith, editor. New York:
John WiIey Sons, Inc., 1961.

 

92

93

Periodicals

Bangs, J. L., and Mullins, C. J. "Recruitment Testing in
Hearing and its Implications," AMA Archives of
Otolaryngology. 58 (1953), 582-592.

 

Blegvad, B., and Terkildsen, K. "Contralateral Masking
and the SISI Test in Normal Listeners," Acta—

Otolaryngologica. 63 (1967), 557-563.

 

. "The SISI Test in Normal Listeners," Acta-
Otolaryngologica. 62 (1966), 201—212.

 

Carhart, Raymond and Jerger, James F. "Preferred Method
for Clinical Determination Of Pure-Tone Thresholds,"
Journal of Speech and Hearing Disorders. 24 (1959),
3304345.

Dallos, Peter J., and Carhart, Raymond. "Cumulation of
DL'S for Intensity Change at Low Sensation Levels,"
Journal Of the Acoustical Society of America. 35
(1963T, 848-855.

Denes, P., and Naunton, R. F., "The Clinical Detection of
Auditory Recruitment," Journal Of Larypgglogy and
Otology. 64 (1950), 375-398.

Dimmick, Forrest L., and Olson, Ruth. "The Intensive
Difference Limen in Audition," Journal of the
Acoustical Sociepy of America. 12 (1941), 517-
525.

 

DOppelt, Jerome E. "Estimating the Full Scale Score on
the Wechsler Adult Intelligence Scale from Scores
on Four Subtests," Journal of Consulting Psychology.
Vol. 20, NO. l, 1956, 63—66.

Fletcher, H., and Wegel, R. L. "The Frequency-Sensitivity
of Normal Ears," Physical Review. 19 (1922), 553-
565.

Garner, W. R., and Miller, G. A. "Differential Sensitivity
to Intensity as a Function of the Duration of the
Comparison Tone," Journal of Experimental Psychology.
34 (1944), 450-463.

Halm, T. "Determination of the Difference Limen and the
Latest Illustration Thereof in Audiometry," Journal
of Laryngology and Otology. 63 (1949), 464-366.

94

Hanley, C. N., and Utting, J. F. "An Examination of the
Normal Hearer's Response to the SISI," Journal of
Speech and Hearing Disorders. 30 (1965), 58-65.

 

 

Harris, J. D. "Loudness Discrimination," JSHD Monograph
Supplement. 11 (1963), 1-17.

 

 

. "The Effect of Sensation Levels on Intensive
Discrimination of Noise," American Journal of
Psychology. 63 (1950), 409-421.

 

 

Hirsh, I. J., Palva, T., and Goodman, Allan. "Difference
Limen and Recruitment," AMA Archives of Otolaryngol-
ogy. 60 (1954), 525-540.

Jerger, James F., Shedd, Joyce Lassman, and Harford, Earl.
"On the Detection of Extremely Small Changes in
Sound Intensity," AMA Archives Of Otolaryngology.
69 (1959), 200-211.

Jerger, James F. "A Difference Limen Recruitment Test and
Its Diagnostic Significance," Laryngoscgpe. 62
(1952), 1316-1332.

 

. "DL Difference Test," AMA Archives of Otolaryngol—
ogy. 57 (1953), 490-500.

. "Differential Intensity Sensitivity in the Ear
with Loudness Recruitment," Journal of Speech and
Hearing Disorders. 20 (195577 183-191.

 

 

. "Comparative Evaluation of Some Auditory Measures,"
Journal of Speech and Hearing Research. 5 (1962),
3-170

. "Hearing Tests in Otologic Diagnosis," Asha. 4

(1962), 139-145.

Knudsen, Vern O. "The Sensibility of the Ear to Small

Differences of Intensity and Frequency," Physical
Review. 21 (1923), 84-102.

Lidén, G., and Nilsson, G. "Differential Audiometry,"
Acta-Otolaryngologica. 38 (1950), 521-527.

 

Loeb, Michel, and Binford, John R. "Some Factors Influenc-
ing the Effective Auditory Intensive Difference

Limen," Journal of the Acoustical Society of
America.7'35 (1963), 884-891.

 

95

Loeb, Michel, and Dickson, C. "Factors Influencing the
Practice Effect for Auditory Thresholds," Journal
of the Acoustical Society of America. 33 (1961),
917-921.

Loeb, Michel, and Harokes, G. R. "Auditory Intensity Dis-
crimination as a Function of Stimulus Presentation
Method," Journal of the Acoustical Society of
America. 734 (1962), 1643-1647.

 

Lund-Iversen, L. "An Investigation of the Difference Limen
Determined by the Method of Luscher-Zwislocki in
Normal Hearing and in Various Forms of Deafness,"
Acta-Otolaryngelogica. 42 (1952), 219-223.

 

Lfischer, E., and Zwislocki, J. "A Simple Method for In-
direct Monaural Determination Of the Recruitment
Phenomenon (Difference Limen in Intensity in
Different Types of Deafness)," Acta-Otolaryngologica
Supplement. 78 (1949), 156-168.

 

Montgomery, H. C. "Influence of Experimental Technique on
the Measurement of Differential Intensity Sensit-
ivity Of the Ear," Journal of the Acoustical Sociepy
of America. 7 (1935), 39-43.

 

Owens, Elmer. "The SISI Test in VIIIth Nerve Versus Coch-
lear Involvement," Journal of Speech and Hearing
Disorders. 30 (1965), 252-262.

 

 

Riesz, R. R. "Differential Intensity Sensitivity of the
Ear for Pure Tones," Physical Review, 31 (1928),
867-875.

Sanders, Jay W., and Simpson, Mary E. "The Effect of Incre-
ment Size on Short Increment Sen51tiv1ty Index
Scores," Journal Of Speech and Hearing Research.

9 (1966), 297-304.

Sherrick, C. E. "Effect of Background Noise on the Auditory

Intensive Difference Limen," Journal of the
Acoustical Society of America. 31 (1959), 239—242.

Stevens, S. S., Morgan, C. T., and Volkmann, J. ."Theory
Of the Neural Quantum in the Discrimination of
Loudness and Pitch," The American Journal of

Psychology. 54 (1941777315-335.

 

96

Swisher, Linda P., Stephens, Myrna M., and Doehring, D. G.
"The Effects Of Hearing Level and Normal Variability
on Sensitivity to Intensity Change," Journal of
Auditory Research. 6 (1966), 249-259.

 

 

Yantis, Phillip A., and Decker, Robert L. "On the Short
Increment Sensitivity Index (SISI) Test," Journal
Of Speech and Hearing Disorders. 29 (1964), 23I-
246.

Unpublished Material

Rosenberg, P. E. "The Influence of Stimulus Duration Upon
Differential Intensity Sensitivity in Normal and
Impaired Ears," Unpublished Ph.D Dissertation,

Northwestern University, 1956.

Other Sources

Dunn, Lloyd M. Expanded Manual for the Peabody Picture
Vocabulary Test. Minneapolis: American Guidance

Service, Inc., 1965.

Michigan State University, Agricultural Experiment Station.
"Calculation on the BASTAT Routine," STAT Series

Description #5 (March 1966).

Wechsler, David. Manual for the Wechsler Adult Intelligence
Scale. New York: The Psychological Corporation,

1955.

Instruction Manual for the SISI Adapter, Model
Illinois: Gordon N. Stowe and

1259. Northbrook,
Associates.

on

Letter from James F. Jerger, Director of Research, Houst4
I

Speech and Hearing Center, Houston, Texas, Dec.
1967.

 

‘a

APPENDIX A

COMPUTED DL VALUES PER SUBJECT

97

Subject

koooqmmprI—I

Age
5

DL

1.89
1.33

12.58

3.62
3.82
5.11
1.29
1.43
2.17
1.43
2.42
2.54
1.16
2.22
1.26
2.12
1.67
1.19
1.06
1.87
2.45
1.37
1.41
1.43
1.09

.96
1.67

.92
1.18
1.42
1.21
1.42
1.36
1.25
1.25
1.27

98

Subject

37
38
39
4O
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
7O
71
72

Age
11

Adults

D_L_

.92
1.13

.93

.92
1.31
1.29

.98

.91
2.00
1.42

.95
2.52
1.07
1.12
1.62
1.26

.90
2.54
1.36
1.42
1.33

.95
1.26
3.09
2.81
1.91
1.68
1.55
1.83
1.92
2.25
1.65
2.53
2.72
1.76
1.52

APPENDIX B

SISI SCORES AT THE 1 DB INCREMENT PER SUBJECT

99

100

 

 

SISI SISI
Subject Age Score (%) Subject Age Score (%)
1 5 45 37 11 50
2 " 35 38 " 40
3 " 10 39 " 60
4 " 0 40 " 80
5 It 40 41 " 50
6 " 25 42 " 35
7 " 20 43 " 60
8 " 4o 44 " 70
9 " 20 45 " 30
10 " 45 46 " 50
11 " 35 47 " 55
12 " 0 48 " 5
13 7 60 49 13 50
14 " 5 50 " 45
15 " 45 51 " 15
16 " 15 52 " 55
17 " 15 53 " 65
18 " 55 54 " 10
19 " 35 55 " 25
20 N 0 56 " 10
21 u 0 57 " 35
22 " 3o 58 " 65
23 ll 40 59 " 30
24 n 30 6O " 10
25 9 45 61 Adults 5
26 u 60 62 It lg
2 n 40 63 n
29 " 30 55 "
30 " 35 2g 1: 18
31 " 40
32 n 40 68 : 5
33 " 30 59 u 0
34 n 25 70 5
35 " 25 71 " 15

36 ll 45 72 II 35

APPENDIX C

INTELLIGENCE QUOTIENT SCORES PER SUBJECT

101

Subject

\OGDQONU'Ith-WNI"

127
141
131
107
107
107
127
125
135
109
105
105
142
142
130
100
107
107
141

135.

143
107
108

95
129
143
144

91
109

96
133
137
129
107
105

102

Subject

37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72

Age
11

$9.

130
125
145
94
106
106
136
130
131
91
98
100
136
141
135
107
104
109
128
125
128
105
105
100
172
130
162
92
100
97
130
132
140
90
92
107