close

Вход

Забыли?

вход по аккаунту

?

00004519-197909000-00007

код для вставкиСкачать
0360-9294/79/0502-05$02 oO/O
JOURNAL OF T H t A M t R I C A N AUDITORY
SOCIETY
Copynght 0 1979 by The Williams & Wilklns Co
Vol 5 . No 2
Printed in
U SA
THE HEARING MEASUREMENT SCALE AS A PAPERPENCIL FORM: PRELIMINARY RESULTS
WILLIAM G. NOBLE
Department of Psychology. The University of New England
Received January 22. 1979, accepted June 7, 1979
tively if they know about the relative incidences
and degrees of handicap arising from hearing impairment in various communities within these cultures. A written-down form of self-report, based
on a paper-pencil inventory, allows large-scale
surveys to be conducted fairly economically.
Schein et al. (1970) recognized the value of such
an approach in the United States when they attempted to devise a brief questionnaire that would
give the equivalent information provided by tonal
threshold estimates at 0.5, 1, and 2 kHz. Their
attempt was not altogether successful, and in any
case, the questionnaire that they developed could
not be said to provide much detailed information
about everyday hearing handicap problems. I have
appraised this and similar ventures in more detail
elsewhere (Noble, 1978, Chapters 9 and 10) and
do not intend to comment further on related research in this field in the present paper.
What does need to be stressed, in concluding
these introductory remarks, is that the Hearing
Measurement Scale is favored above other selfreport inventories for the purpose of handicap
assessment, because of its breadth and detail of
coverage.' Hence, a program to develop a paperpencil form of this scale for population surveys of
hearing handicap would seem to be a worthwhile
exercise.
The Hearing Measurement Scale developed by
Noble (1969) and described by Noble and Atherley (1970) was originally designed for use as an
interview schedule. In many circumstances, the
advantages of an interview format outweigh the
cost in time required for individual appraisal. Interviews using the scale can take 10 to 20 min and
at times up to 45 min to conduct. The desirable
features of interviewing are that precise meanings
of questions and answers can be negotiated between interviewer and respondent; the relevance
of different situations can be more exactly pinpointed, and, in keeping with this whole style of
approach to assessment of hearing impairment, an
interview is a person-to-person interaction that
vividly emphasizes the primacy of the person who
suffers handicaps arising from that impairment.
An interest is being shown in the person with a
hearing impairment by the very fact that a personal approach is adopted to the appraisal of its
effects.
I would want to argue, therefore, that in rehabilitative or clinical contexts where the concern is
to make a full appraisal of problems that people
face as individuals in the world with a view to
mediating these problems, face-to-face interviews
are the appropriate modus operandi.
There are several circumstances where information about hearing ability needs to be obtained
on a larger population scale. In particular, the
populations of high technology cultures are becoming increasingly aged, and large numbers of
people in these same cultures possess a history of
exposure to occupational or armed service-connected noise. Agencies charged with the task of
rehabilitation can undertake that job more effec-
Preparatory Procedure
This research was partly supported by an InternalResearch
Grant from the University of New England.
The wording of some items in the original interview form of the Hearing Measurement Scale
is fairly loose, which is to say that although their
meanings are clear, the precise referents of the
items typically need to be elaborated upon in the
face-to-face interaction of the interview itself. Furthermore, the scoring categories are both arbitrary
and inconsistent, again a feature that has little
' This is the view of 2 independent investigators (Ward.
1976; Stephens, 1978) who surveyed a range of available
self-report instruments for the purposes of application in
rehabilitative contexts.
Send reprint requests to: Associate Professor William G.
Noble, Department of Psychology, University of New England,
Armidale. New South Wales 2351, Australia.
95
96
NOBLE
consequence in an interview where the exact degree or extent of difficulty is negotiated between
interviewer and respondent, responses then being
placed against the nearest equivalent scoring category (details of these procedures are given in the
revised edition of the Test Manual for the Scale,
Noble, 1979). Although such features are simply
characteristic of a scale designed for interview,
they are obviously unacceptable in a written-down
form that the respondent has to deal with unaided
by an interviewer’s elaborations. Wording both of
items and scoring categories must be clear-cut and
unambiguous.
As a first step in identifying the equivocal and
imprecise elements in the Hearing Measurement
Scale, the original interview form was presented
to the 14 members of a University Department.’
They were asked to complete the questionnaire
with an eye to picking out flaws and uncertainties
in its content. From the consistency of some annotations provided, it was clear in what ways a
number of items needed to be reworded. It was
clear also that the scoring categories needed to be
less ambiguous as well as more consistent across
items. A revised version, taking account of the
criticisms was presented a few months later to the
same group of commentators, and only sporadic
objections and questions were raised in response
to the new form. It was considered that the revised
version was fit for trial in a clinical sample.
The revised set of questions is given in Appendix
I. A 5-point scoring scale was adopted in the repeat
preparatory trial for all appropriate items. The
new scale is also shown in Appendix I. This scale
differs from the various scales in the original questionnaire because in the original the items were
given 2-, 3-, and 5-point scoring scales, depending
on their weight. Items in the Hearing Measurement Scale, now as then, are differentially
weighted (there are 9 categories of weighting such
that maximum score on the least weighted items
is I , and maximum score on the items with most
weight is 9). Items with maximum scores of I in
the original version of the HMS had 2-point scoring scales (see Noble and Atherley, 1970, for details of weighting and scaling). Some items in the
new version (which previously had only 2- or 3point scales) now have the new 5-point scale. As
mentioned, the new scale has been applied where
appropriate throughout the revised form?
The author wishes to thank Professor Gordon Atherley
and the members of the Department of Occupational Safety
and Hygiene, University of Aston in Birmingham, England, for
their generous assistance in participating in this exercise.
The appropriateness of items in this context means that
As a result of this change in scaling, but not in
weighting, responses are likely to fall midway
between scorable scale points. This is a potential
feature that I drew attention to in the test manual
for the original Hearing Measurement Scale (Noble, 1971), and one fairly obvious solution suggested there was to ascribe a midpoint score (for
example, 0.5) if respondents answered around the
midpoint (“half the time.” or “sometimes”) to a
question with 0 or 1 as the only scale points
available. That solution has been adopted in the
present research study, which accounts for the
presence of numbers rendered to 0.5 in the individual results given in Appendix 11.
Field Trial of Paper-Pencil Form
Sample Selection. Through the offices of an
otologist colleag~e,~
contact was established with
30 people who had been referred to this specialist
over the preceding year. The sample comprised
men whose hearing impairment the otologist had
diagnosed as noise induced, although that diagnosis was made with varying degrees of certainty
between cases. The reasoning behind concentration on such a diagnostic sample partly derives
from a point made in introduction to this paper,
that a significant subpopulation in societies like
ours is that of people with a history of occupational
and service-connected, noise-induced hearing disorder. Partly, also, use of such people provides
some continuity with the samples used in development of the original Hearing Measurement
Scale (Noble and Atherley, 1970). As it turns out,
response patterns of the present sample are different from those of the original reliability study
sample, no doubt because of the uncertainty in
diagnosis. Implications of this difference will be
considered later.
The initial approach by the otologist to his
clients sought their permission to let the author
inspect their audiological test records. Clients were
also informed that I would approach them independently, if they were willing, to seek their help
in a research project. Affirmative replies were
received from 28 of 30 clients, and 2 groups of 14
persons each were derived, matched as well as
the wording of the item is such that the new 5-point scale can
be used meaningfully by the respondent in answering the
question. Certain items (e.g.. Items 28. 30. 32, 40,41, q.v. in
Appendix I) cannot take the new scale, so the original scale
forms are used in these cases.
‘
I would like to thank Dr. Peter McArthur of Armidale. New
South Wales, and his staff for their unfailing help and care in
arranging for the successful contact that was made with his
clients.
HEARING MEASUREMENT SCALE AS A PAPER-PENCIL FORM
could be done on the basis of age and tonal
audiometric records.
Design. The aim of the clinical field trial of the
paper-pencil form was to ascertain the consistency
between such a form and the original scale. The
original HMS was tested for stability by repeated
application after a 6-month interval on a sample
of 27 foundrymen. A reliability coefficient (Spearman) of 0.928 was obtained, indicating a high
degree of stability over time. To test the consistency of the 2 forms, the present design entailed
one subgroup of 14 participants being interviewed
on the first test occasion and then being followed
up six months later with the paper-pencil version
mailed to them at their homes. The other subgroup
was mailed the paper-pencil questionnaire and
then interviewed 6 months later. The revised item
order and wording was used throughout, there
being no essential changes in item meaning or
content as a result of the revision and yet an
improvement in clarity that merits use of the revised form, even in interview.
By means of this design, interform consistency,
controlling for order of presentation, could be
appraised. A necessary further component in the
complete testing of the paper-pencil form will be
a stability study over time using the written-down
version on both occasions. This is planned as a
future project.
Procedure. Paper-pencil forms were mailed, together with an explanatory letter, to members of
the group whose characteristics are summarized
first in Table 1 (Group A). At-home interviews
were arranged with the second group (B) described
in the table. About 6 months later, the alternate
treatment was applied to the 2 subgroups. In the
intervening time, a summary of initial results was
sent to each participant, not only to keep them
informed about the research project but also
97
as a way of reminding them that a follow-up
would be undertaken. As can be seen, the final
numbers on initial and repeat testing are less than
14 in each group. Nonreplies, incomplete replies,
and changes of residence away from the local area,
plus one refusal to participate, account for both
the initial shortfall and subsequent attrition. Complete test-retest records are in fact available from
only 1 1 people in each group. These numbers are
just sufficient to permit statistical treatment.
Incomplete retest results from 2 people in Group
B (interview followed by paper-pencil form) arose
because pages of the questionnaire form were
missed by the respondents. In one case, the effect
is minimal (one scoring item overlooked); hence,
his results can be included with a suitable adjustment to the initial score (subtraction of the score
from initial test total on the item missed on retest).
Thus, final overall numbers for purposes of analysis are 23.
Results
It is apparent from data in Table I that despite
the dropout problem, the 2 subgroups remain reasonably well matched in terms of age structure
and audiometric performance. Group A is somewhat younger than Group B and displays less
dispersed ranges of threshold levels and lower
average threshold levels at 0.25 to 2 kHz than
Group B and threshold levels at variance with
those for Group B at 4 and 8 kHz. These differences in average tonal thresholds between the 2
groups are not too surprising given the smallness
of the samples and the inherent unreliability of
audiometric testing at lower and higher frequencies. In audiometric terms, the differences are quite
minor, and there are no statistically significant
differences in thresholds between the groups. Nevertheless, results from Group A are marginally at
TABLE 1
Audiometric and age characteristics 01both groups of participants in field study
kHz
Group A"
Mean AC threshold
(dB ISO)
Left ear
Right ear
Group B'
Mean AC threshold
(dB ISO)
Left ear
Right ear
0.25
0.5
I
2
4
8
14.5 f 5.7h
18.2 f 7.5
17.7 f 6.1
15.5 f 8.8
24.5 f 14.6
19.5 f 10.1
33.6 f 19.1
27.7 f 20.7
55.9 f 23.2
55.0 f 24.3
51.8 31.3
55.5 f 29.5
17.9 f 13.6
22.5 f 11.8
18.3 f 13.2
17.9 f 9.6
23.3 f 19.2
20.8 f 14.4
39.2 f 28.4
28.3 26.0
53.3 f 29.6
46.3 f 29.9
61.7 f 22.1
50.4 f 32.2
*
" N = I I. Paper-pencil form on initial test: interview on retest. Mean age. 48 years (S.D.
= 10.4).
Mean k S.D.
' N = 12. Interview form on initial test: paper-pencil on retest. Mean age. 50 years (S.D.
= 11.9).
*
98
NOBLE
a lower level and less dispersed than those from
Group B. Test-retest results from the Hearing
Measurement Scale for both groups are given in
Table 2. Results shown are total scale scores and
subsection scores. Individuals' total and subsection
results are given in Appendix 11. By way of explanation of the breakdown in Table 2, it should be
pointed out that the Hearing Measurement Scale
comprises 42 scoring items in 7 sections, the items
distributed as summarized in Table 3 and shown
in Appendix I.
In Section I, results have been analyzed for the
section as a whole and also for 2 overlapping
subsets of items (Items I to 6 and 6 to 11). The
reason for examining this subdivision is that in
analyzing results using the original scale (Noble,
1969), I found that Items I to 6 were reasonably
well intercorrelated. as were Items 6 to 11, but
there was not a close association between items
across these 2 sets (with the self-evident exception
of Item 6). I have argued (Noble, 1978) that Items
I to 5 are concerned with face-to-face interaction,
whereas Items 7 to 1 1 are concerned with listening,
and these are distinct aspects of speech hearing.
Item 6, on listening at a public gathering, may well
contain elements common to both activities, however. It is worthwhile in my view to pursue separate analyses of these subsets, therefore.
The most immediately obvious outcome shown
in the results in Table 2 is the marked change in
scores in both groups but in contrary directions on
retest. Whereas the difference between the groups'
initial average total scores is nonsignificant (t =
0.52), the retest average difference yields a I value
of 2.67 (p < 0.02). Furthermore, the downward
change in average score in Group A yields a t
value (correlated means) of 2.77 (p < 0.02). The
upward change in the average of Group B is
nonsignificant.
Because of the nature of the survey design, a
complete analysis cannot be undertaken to reveal,
for instance, the extent to which an order-of-test
effect may interact with a mode-of-test effect to
bring about this unanticipated result. The outcome
demonstrates that interviewing tends to provide
lower scoring than self-assessment by paper-pencil
means. This result in turn suggests a self-disclosure
element in the responses to items on the questionnaire; respondents revealed a greater degree of
self-assessed difficulty when answering questions
for themselves than when conversing with the
interviewer. What remains unknown is whether
particular orders of testing (interview followed by
paper-pencil, for instance) account for any of this
variability or whether mode of test (interview or
99
HEARING MEASUREMENT SCALE AS A PAPER-PENCIL FORM
TABLE 3
Summary of rhe Hearing Measuremeni Scale
Section
I
11
111
IV
V
VI
VII
Section title
No. of items
Maximum possible
score
Speech hearing
Hearing for nonspeech
sounds
Spatial localization
Emotional response to hearing impairment
Speech distortion
Tinnitus
Personal opinion of hearing
II
8
76
28
7
7
28
45
3
3
3
20
16
13
42
226
paper-pencil) is the sole factor governing the effect. If the latter is the case, then retest results from
Group A should be comparable with initial test
results from Group B and likewise for initial and
retest results from the 2 groups, respectively. However, that being so, we are faced with a markedly
greater degree of self-reported hearing difficulty
in Group B compared with Group A and one that
is scarcely paralleled by the audiometric test data,
even though these latter data are marginally in the
same direction. If however, we are faced with an
order-of-test effect lying alongside a mode-of-test
effect, then the divergence from initial test results
could be viewed as a differential effect acting on
groups that commence with similar kinds of responses irrespective of test mode. The differential
effect would have to be argued as an inhibition
arising from interview on retest following paperpencil form in one group and as a facilitation on
paper-pencil responding following interview in the
other group. This is a difficult case, both conceptually and theoretically, whereas the mode-of-test
idea has more straightforward appeal. Yet, the
latter possibility has no support from other information that one gathered about these participants.
It must be mentioned that no help is forthcoming
in trying to solve this puzzle from the original
stability study results (Noble and Atherley, 1970),
because these revealed virtually no change in selfreported hearing difficulty over a 6-month period
when respondents were interviewed on both occasions. Such an outcome at first seems to support
a mode-of-test effect operating in the present
study, but the particular conditions of the present
study could nonetheless have been productive of
differential effects unaddressable by a previous,
unimodal survey. It is necessary to consider these
outcomes in detail here because they bear on
feasible ways of interpreting analyses of the data
for consistency. The test-retest total scale scores of
the groups can probably only be compared independently inasmuch as the divergent changes on
retest intrude on an analysis of paper-pencil versus
interview irrespective of order-of-test. All possible
correlations of total scores have been calculated
nonetheless, and the Pearson product-moment
correlations of total scores of the 2 groups separately and for the 2 types of combined group scores
are given in Table 4.
These correlation values should be treated cautiously, but they do point to a possible order-oftest element in the outcome. The fairly close correlation obtained from test-retest total scores of
Group A contrasts with the rather looser relationship observable in the scores of Group B. The
“coefficients of determination” given alongside the
correlation values show the character of the relationships more clearly. Were order of test not a
factor in the results, one would anticipate greater
similarity in the correlations. The combined testretest r value also stands in contrast to the somewhat higher correlation between paper-pencil and
interviewing modes of administration, i.e., ignoring order of test. It is this latter analysis that, as
mentioned earlier, necessarily suffers interference
from a possible order-of-test factor. The low r
value associated with combined test-retest results
is expectable, given the divergent changes in group
scores across occasions of test. The higher r value
between test modes is a more gratifying outcome
suggesting some measure of consistency across test
modes in spite of the hypothesised order-of-test
and self-disclosure factors affecting the results.
It is worthwhile to explore results from the
viewpoint of consistency in more detail to identify
subsections and other groups of items that provided consistent results despite an overall lack of
high consistency. Individual total scale scores have
100
NOBLE
TABLE 4
Pearson product-moment correlations and coefjcients of determination (in brackets) between initial and retest total
scale scores f o r the 2 groups separate()>and combined and between paper-pencil testing and interview
lie., irrespective of order-of-test)
Paper-pencilinterview
Inirial-retest
Group A”
Group B‘
Groups A and B combined
0.853
0.732
0.61 I
(73)h
(54)
(45)
0.795
(63)
Paper-pencil test and then interview.
* Numbers in parentheses, coefficients of determination (4 X
100).
‘ Interview and then paper-pencil.
been rounded up for purposes of the foregoing
analyses, but individuals’ subsection scores may
be rendered to 0.5 of a whole number. This is due
to more uniform adoption of a 5-point scoring
category system, the effect of doing which was
described earlier in the paper. The rounding-up
procedure accounts for the minor inconsistencies
between average total scores and the sum of averaged subsection scores (Table 2; Appendix 11).
The effect of rounding up totals but not item or
subsection scores is of course negligible in the
statistical analysis.
Table 5 gives Pearson product-moment correlations between initial and repeat test subsection
scores for the 2 groups separately. Also shown are
equivalent correlation values for interviewing versus paper-pencil responding. The fourth possible
analysis (combined test-retest correlations) would
be unproductive given the expectably low total
scale score correlation.
In Table 6 are mean differences and estimates
of the standard deviations of differences between
each set of scores correlated in Table 5. These
reflect both the constant and random variability
of scores between the 2 test series and give information about the actual closeness of association in
numerical terms as well as clarify results where
restricted range effects distort some correlation
values in samples of this small size. It is evident
from Table 2 and Appendix I1 that the smallerscoring subsections show skewed and restricted
distributions of scores, making correlation values
subject to artifactual fluctuation. It is of course
undeniably the case that quite inconsistent results
were obtained in response to items in Table 3,
Section 111 and to some extent also in Sections V
and VI, and I will discuss these later. The results
to concentrate on are from Sections I, 11, IV, and
VII. Sections I and IV carry the bulk of scoring
(about two-thirds of the total), the scores on the 4
sections together account for roughly 85% of total
score. The unfortunate feature from the standpoint
of consistency revealed in the analyses in Tables
5 and 6 is the outcome on Section I scores across
samples. Whereas Items 6 to 1 I (“listening” items)
show fairly consistent responses in Group A and
Items 1 to 6 (“communication” items) do not, the
contrary is the case for Group B. Over both groups,
as a result, the correlations and standard deviations of differences are respectively depressed and
inflated on both subsets of items. No conclusion
can be drawn at this stage about the confidence
that can be placed on Section I scores obtained by
different forms. The contrariwise fluctuation in
scores on item subsets across samples may be a
sample size phenomenon that would wash out in
a larger body of data. On the other hand, it may
be a real consequence of the hypothesized orderof-test effect, although there is no clear way of
reasoning about what could underlie it. If it is a
real effect, then of course a conclusion would be
that scores on Items 6 to I I are liable to show fair
consistency if paper-pencil administration precedes interview; the same prediction being made
about Items 1 to 6 if interviewing is undertaken
first. However, such a conclusion would be hazardous at this stage, and the picture will only
become clearer with more extensive investigation.
A feature of the stability study using the original
version of the scale (Noble, 1969) is that scores on
the equivalents of Items 1 to 6 gave a test-retest
correlation of r = 0.825, whereas the equivalents
of Items 6 to 1 I gave a value of r = 0.525 ( N =
27). This finding is in line with that for Group B
in the present study provided an argument is made
that initial interviewing is critical in producing
some stability in scores on the first items of the
scale. A stability study using the scale as a paperpencil form is obviously required to discover
whether the pattern observed in Group A in the
present study is reproduced.
The test-retest scores on Sections 11, IV, and VII
HEARING MEASUREMENT SCALE AS A PAPER-PENCIL FORM
101
TABLE 5
Test-retest correlation coefficients for each group separatelv and correlations between interview and paper-pencilform
(both groups combined) on each section and total scale
Section
Group A
Group B
Combined”
I
1I.h
0.948
0.592
0.680
0.672
0.855
0.697
I,,,,
I1
111
0.858
0.486
0.606
0.817
0.733
0.742
0.327
0.644
0.341
IV
V
VI
0.886 -0.343 0.667
0.779
0.523 0.444
0.807
0.241 0.538
VII
Total
0.780
0.854
0.810
0.853
0.732
0.795
“ Interview correlated with paper-pencil form. Group A’s initial results and Group B’s retest results correlated
with Group B’s initial results and Group A’s retest results. N = 23.
provide more intelligible results, in the sense that
the moderate consistency evident in the data is
maintained across samples. The hypothesized selfdisclosure effect is quite obvious, however, in the
Section IV scores, which is not unexpectable because respondents are asked some potentially
rather threatening questions in that section. One
might, thus, expect a more defensive response in
a face-to-face interview encounter compared with
the less interpersonal, self-administered form.
The most inconsistent results emerge from
scores on Section 111. To digress, it should be
mentioned that a procedure adopted in testing of
Group A (interview on retest) was to take each
respondent’s first test (paper-pencil) questionnaire
form to the at-home interview. At the end of the
interview, I produced the first-test questionnaire
(unseen for 6 months) and briefly compared responses on the 2 forms. Where major inconsistencies were evident in scores, this was pointed out
to the respondent, with a view to trying to fmd out
what differences in approach to the questions
might have been operative on the first as against
second test occasion. Section I11 had clearly provided most problems for these men, and from
postinterview conversation, it emerged that many
of them had been faced with items demanding
quite a degree of sensory self-awareness. In casting
around for ways to tackle the questions, respondents had to rely on sporadic instances. Hence, the
marked variability could be put down to an inability to survey typical experience. As a result, no
aggregate self-assessed ability to locate sound was
forthcoming. A part of the problem too, in this
section and also in Sections V and VI was the
clinical composition of the groups. Sections 111, V,
and V1 are peculiarly relevant to people with
sensorineural disorder (Noble, 1978). These types
of respondents notice problems in location and
speech discrimination and often have noticeable
tinnitus. As mentioned in introduction to this paper, diagnoses of disorders in the present groups
were made with varying confidence, and there was
no doubt that mixed disorder was represented to
a greater extent than in previous samples tested
with the scale by this author. In consequence, for
many participants, items in Sections 111, V, and VI
had no clear meaning, referring to problems of
which a proportion of these respondents would
never have been aware. However, given an expectation on their part that some kind of response was
sought (why else would the questions be there?),
the men may have simply given whatever reply
they could. In the reliability study of the original
Hearing Measurement Scale (Noble, 1969), Section 111 scores gave a test-retest coefficient (Spearman) of 0.854. The sample ( N = 27) tested in that
study was of men with pure noise-induced hearing
disorders, and they had more noticeable difficulty
in localization than the present groups. Thus,
whereas test and retest mean total scale scores in
that noise-deafened sample were 35.3 and 35.2,
Section 111 scores were 5.6 and 5.7 (cf. Table 2 for
appropriate comparative scores). The point of
mentioning this is to try to give substance to the
argument that the low consistency in some sections
in the present study, in being at odds with previous
results, can arguably be said to reflect different
sample composition and not automatically be
thought of as an intrinsic defectiveness of the scale.
Conclusion
I believe it is safe to conclude that the Hearing
Measurement Scale in its revised form can usefully
be applied as a paper-pencil self-administered
questionnaire. In making that conclusion, I am
not, however, claiming that use of the scale in this
way is an adequate replacement for the scale used
in interview. The 2 forms in the present, smallscale and necessarily preliminary survey did not
provide closely consistent results. However, they
did provide fairly comparable outcomes, giving a
coefficient of 0.795 between total scores on paperpencil and interview administration. Subsections
102
NOBLE
+I
tl
within the instrument as a whole, notably Sections
11, IV, and VII gave results that maintained consistency irrespective of order of presentation of the
forms. An uncertain outcome was that for Section
I aspects of which gave differentially consistent
results in the 2 orders. More extensive survey work
will be needed to explore and clarify that outcomc.
Another way in which the scale as a paperpencil form differs from the scale as interview is
in actual score levels observed in the samples
tested with the 2 types. Respondents fairly consistently revealed somewhat higher levels of selfassessed hearing difficulty when tilling out the
questionnaire for themselves, and a hypothesized
self-disclosure effect could account for this outcome. However, an order-of-test effect was also
argued to be operating, revealing, therefore, a
limitation in the present design.
Notwithstanding these qualifications, use of the
scale as a screening device in large-scale population work would serve to replace the more costly
interview method and at that (population) level
would undoubtedly give intelligible results. The
Hearing Measurement Scale has a well-established
sensitivity to different degrees of hearing disturbance (Noble and Atherley, 1970 Mays, 1977) and
also seems to be sensitive to the different effects of
different types of disorder (Noble, 1978: and certain features of the present results). Given the
present results that show a fair level of consistency
between the scale as paper-pencil and interview
forms, it can be anticipated that in the former
mode it would continue to be valuable for identifying and differentiating the incidence and degree of hearing difficulty in various samples.
+I
I
Summary
t l t l t l
$ 5 4
I
4
:
- .i
I1
8
II
I 0 Z
m
1 1 .
a.-
'02
ip v4 -3 t" v v
;#;;
- -~
'
~
L.5
E .I
P L="
p .- a
no
:I=
$X 'j;
& ,c"
e > G
.Do.!?
G
"-x88
v)uI
I;.$!$g
g
L& $ % & &
:z2zz5
f
r
l
r
c
Following revision of item wording and scaling,
as determined from the responses and comments
of a group of independent observers, the Hearing
Measurement Scale was applied as a paper-pencil
(self-administered) form to a group of 1 I partially
deaf men with preponderantly sensorineural hearing disorders (mean age, 48 years). At the same
time, the revised version was used in the traditional
interview manner with a further 12 men (mean
age, 50 years) of comparable audiometric and
diagnostic status to the first group. After 6 months,
~
f
9
the reverse procedure was applied to the 2 groups.
Test-retest correlation in the first group (paperpencil followed by interview) of total scale scores
gave r = 0.853 and in the second group gave r =
0.732. Combined results to compare paper-pencil
with interview administration (i.e., irrespective of
test order) gave r = 0.795. Among features of the
HEARING MEASUREMENT SCALE AS A PAPER-PENCIL FORM
results, there was noted a trend to greater selfassessed hearing difficulty in paper-pencil responses, and a “self-disclosure” element was hypothesized to have operated in the study. Also, an
order-of-test effect was hypothesized to have occurred. Within sections of the scale, those covering
hearing for nonspeech sound, emotional response
to impairment, and personal opinion of hearing
gave fairly consistent results across samples. The
section on speech hearing gave partially consistent
(but across groups, differentially consistent) results. The sections on spatial location, speech distortion, and tinnitus gave inconsistent results, and
this was thought to be due to sample composition.
It is concluded that although the hearing measurement scale as a paper-pencil form may not be
wholly consistent with its use at interview, present
results encourage a view that it may successfully
be used in large-scale population work, for example, as a preexamination screening device. Further
research to examine the stability of the paperpencil form is required to establish its utility in
follow-up surveys.
References
Mays, D. A. 1977. The CID sentence list compared to
CID auditory test W-22, and in relationship to the
hearing measurement scale. M.A. dissertation, University of Cincinnati.
Noble, W. G. 1969. A scale for the measurement of
hearing loss and disability. Ph.D. thesis, University
of Manchester.
Noble, W. G. 1971. Test Manual for the Hearing Measurement Scale. University of New England, Department of Psychology, Armidale, Australia.
Noble, W. G. 1978. Assessment of Impaired Hearing: A
Critique and a New Method. Academic Press, Inc.,
New York.
Noble, W. G. 1979. Test Manual for the Hearing Measurement Scale. Revised ed., University of New
England Press, Armidale, Australia, in press.
Noble, W. G., and G. R.C. Atherley, 1970. The Hearing
Measurement Scale: A questionnaire for the assessment of auditory disability. J. Aud. Res. 10, 229250.
Schein, J. D., A. Gentile, and K. W. Haase. 1970. Development and evaluation of an expanded hearing
loss scale questionnaire. National Center for Health
Statistics, Series 2, No. 37, United States Department of Health, Education, and Welfare, Rockville,
MD.
Stephens, S . D. G. 1978. Evaluating the problems of the
hearing impaired. Paper presented to Fourteenth
International Conference of Audiology, Acapulco,
Mexico.
Ward, P. R. 1976. A research project to evaluate followup services for adults issued with hearing aids.
103
Report to the Institute of Biometry and Community
Medicine, University of Exeter, England.
Appendix I
Hearing Measurement Scale (revised form)
The form given here is of the scoring-item wording
and order used in the present study. Section headings
not given in the questionnaire form itself are given here
as an aid to textual description of the scale. Most items
were accompanied by a new 5-point scaling system, viz:
All the time
Most of the time
About half the time
Occasionally
Never
This scale was abbreviated to:
All
Mo Half Occ Ne
_ _
and a reminder about the meanings of these symbols was
printed at the top of each page on the questionnaire
form. Various instructions accompanied the paper-pencil
version to assist respondents in completing the questionnaire for themselves. These are not included here. The
complete revised form, is now available as a printed,
single-sheet (folded) questionnaire blank published by
the University of New England Press. These can be
purchased in batches of 50, together with a scoring key,
and a revised test manual for the scale containing edited
versions of the present paper, the earlier paper by Noble
and Atherley (1970), and the test manual for the scale
administered as an interview.5
Section I. Hearing for Speech
1. Do you have difficulty hearing in a conversation
with one other person when you’re at home?
2. Do you have difficulty hearing in group conversation
at home?
3. Do you have difficulty hearing in a conversation
when you’re with one person outside? (By “outside,”
1 mean some place outside the house where you
would be talking to others.)
4. Do you have difficulty hearing in group conversation
outside?
5 . At work, do you have difficulty hearing in a conversation? (If yes) Is this due to your hearing, due to
noise at work or a bit of both?
Circle 1, 2, or 3
I. Due to hearing
2. Due to noise
3. A bit of both
6. Do you have difficulty hearing the speaker at a
public gathering (assuming that you are standing or
sitting in a place that makes it possible to hear
properly)?
If you generally don’t attend public gatherings, mark
“ X * here. Tell me, however, whether that is because
hearing difficulty makes it pointless to attend and/
or for other reasons, circle whichever applies (both
if appropriate)
Send inquiries and orders to: University of New England
Press, Armidale, New South Wales, Australia 2351.
NOBLE
104
7.
8.
9.
10.
1. Hearing difficulty
2. Other reason
Do you have difficulty hearing what is said on the
news on television (assuming in this and in the next
3 questions that the ‘volume’ is up at its usual level
in your household)?
Do you have difficulty hearing what is said on TV
programs apart from the news?
Do you have difficulty hearing the radio news?
Do you have difficulty hearing radio programs apart
from the news?
If you generally don’t watch T V news, mark “ X
here.
If you generally don’t watch TV apart from the
news, mark “ X here.
If you generally don’t listen to the radio news, mark
“ X ’ here.
If you generally don’t listen to the radio apart from
the news, mark “ X here.
If you have marked “ X ’ anywhere above, please tell
me whether you generally don’t watch or listen
because of hearing difficulty and/or other reasons.
20. When you can hear the sound of people talking
outside the room you are in, do you have difficulty
telling whereabouts this sound is coming from?
21. If you are with a group of people and someone
behind you starts to speak, do you have difficulty
telling where that person is behind you?
22. If you hear a car horn or a bell and you can’t see it,
do you have difficulty telling which direction it’s
coming from?
23. Do you turn your head the wrong way when someone you can’t see calls out to you?
24. Do you have difficulty judging how far away that
person is, just from the sound?
25. Do you notice that cars you can hear but not see
turn out to be much closer than you thought?
26. Do you tend to move the wrong way to try to get out
of the path of someone or something coming up
from behind you?
Section 4. Emotional Response
27. Do you give the wrong answer to people because
you’ve misheard them? Some people don’t tend to
do this if they mishear what another person says, but
Don’t watch
Don’t listen to
TV news
Other TV
programs
Radio news
Other radio
programs
I. Hearing
difficulty
2. Other reason
1. Hearing
difficulty
2. Other reason
I. Hearing
difficulty
2. Other reason
I. Hearing
difficulty
2. Other reason
I I. Do you have difficulty hearing what’s said in a film
at the cinema? If you generally don’t go to the
cinema, mark “ X here and tell me why by circling
1 and/or 2
1. Hearing difficulty
2. Other reason
Section 2. Hearing for Nonspeech Sound
12. Are you surprised by the arrival of someone at the
house because you did not hear their footsteps?
13. If you have a cat or dog or bird as a pet, can you
hear it when it mews or barks or chirps? (What type
of pet?)
14. If someone calls at the house, can you hear them
ring the doorbell or knock on the door?
15. When you’re in the street, can you hear the sound of
a car horn?
16. When you are in a room facing away from the door,
can you hear it when someone opens the door to
come into that room?
17. Can you hear the clock ticking in the room? (Assuming you have a ticking clock; if not, mark “ X here.)
18. Can you hear the water running when you turn a
tap on?
19. Can you hear the water boiling in a pan when you
are in the kitchen?
Section 3. Spatial Location
rather ask the other person to repeat what they said.
If you tend to do this rather than give the wrong
answer, treat the question as: Do you have to ask
people to repeat what they’ve said because you have
misheard them?
28. If you give the wrong answer or ask people to repeat
what they said, do you treat this lightly or do you
get upset?
29 Do other people get irritated by your wrong answers
or by you asking them to repeat what they said?
30. Do you think other people are tolerant about any
difficulty in hearing you have or do they make fun
of you?
31. Do you get bothered or upset if you are unable to
follow a conversation?
32. Do you think you are generally more irritable than
other people, about the same, or less irritable?
33. Do you get a feeling of being cut off from things
because of difficulty in hearing? If yes, does this
feeling upset you?
Section 5. Speech Distortion
34. Do you find that people fail to speak clearly?
35. Do you find that announcers on TV/radio fail to
speak clearly?
36. In everyday conversation with friends or family, do
you find you cannot unravel the words being said
HEARING MEASUREMENT SCALE AS A PAPER-PENCIL FORM
even though you can hear what is being said?
Section 6. Tinnitus
37. Do you get buzzing or ringing noises inside your
head or ears? If “occasionally” or more often in
answer to question 37:
38. Does this head noise prevent you from getting to
sleep? If “occasionally” or more often in answer to
question 37:
39. Does this noise upset you?
Section 7. Personal Opinion
40. Do you think your hearing is normal?
41. Do you think any difficulty in your hearing is particularly serious?
42. Does any difficulty in hearing restrict your social or
personal life?
Complete questionnaire blanks, containing instructions for self-administration can be purchased, in batches
of 50 along with a revised test Manual and Scoring key,
from:
University of New England Press,
Armidale. New South Wales. Australia 2351.
105
Appendix I1
Individual section and total scale test-retest results
Individual results, section by section. are given to
show in a straightforward way the actual variations in
scoring observed across occasions of test. Data are from
a sufficiently small number of people to allow individual
results to be presented in an uncumbersome way. It also
permits future investigators to compare outcomes and
perform further analyses than the ones presented here.
Scores rendered to 0.5 of a whole number occur in
sections where respondents marked a category on the
new 5-point scale that does not carry a whole-number
score. (See text for full account of this effect of revision
of item scaling in light of unchanged item weighting.)
Total scores have been rounded up. where necessary, to
next whole number.
Scores under the section headings, ‘‘11-6.’ and “ 1 ~ 1 1 ”
are partial scores on Section I. These do not count,
therefore, toward the total score. (Again, see text for
explanation of the presentation of these partial scores.)
Group A: Initial test with paper-pencil form; retest with interview form
I1
Ill
IV
V
VI
2
I
I
I
3
I
15
10
2
2
14
14
5
4
0.5
2
0
0
4
2
16
7
15
7
2
I
0
0
0
0
4 Test
Retest
19
14
14
13
7
2
4
1.5
5 Test
Retest
32
29
26
24
9
9
0.5
2.5
6 Test
Retest
31
21
29
17
6
7
7 Test
Retest
48
28
28
15
22
14
8 Test
Retest
12
22
9
17
5
6
9 Test
Retest
41
28
31
26
14
5
10 Test
Retest
8
17
8
15
1
2
I 1 Test
Retest
11
8
I1
8
0
0
Respondent
Number
I
I Test
Retest
10
11-6
16-11
6
9
5
2 Test
Retest
17
16
3 Test
Retest
VII
Total
6
4
7
5
44
29
2
1
6
4
2
I
32
26
13
9
0
3
5
3
6
6
40
28
1
I
7
4
2
0
0
0
4
3
31
24
0
0
10
12
2
0
0
0
2
6
47
50
2
3
4
I1
9
4
I
0
3
0
5
7
46
15
6
10
2
14
9
3
2
0
2
4
5
94
54
5
3.5
1.5
2.5
13
4
1
0
4.5
5
5
42
38
5
1
5
8.5
25
19
6
0
0
II
I1
94
68
0
1.5
I .5
10
7
5
6
1
.5
I .5
4
1
27
31
1.5
0
.5
0
4
2
2
0
3.5
.5
2
I
25
12
.5
.5
55
NOBLE
106
Group B: Initial test with interview form; retest with paper-pencil form
I
1I.S
161I
I!
I Test
Retest
26
58
24
28
6
33
I1
2 Test
21
30
21
21
I
II
25
33
24
20
21
16
111
IV
V
VI
VII
Total
17.5
5
14
16
17
0
2
0
0.5
13
13
71
I22
4
4
0
8
9
5
0
2
0
0.5
6
6
40
56
4
14
0
I
3
1.5
9
2
6
2
6
4
9
5
58
49
21
16
0
I
0
I .5
0
0
9
15
5
7
0
0.5
4
4
39
44
29
21
28
21
I
0
2
2.5
0
0
12
3
3
I1
I
I
4
3
40
26
17
23
16
3
2
5.5
2
4
2
10
8
3
I
3
3
4
4
56
37
14
10
II
8
3
3
I.5
2
0
2.5
6
12
0
3
I
3
6
I
26
37
36
26
25
19
12
8
8
4
1
5
2
10
3
2
0
0
8
7
58
54
12
8
9
5
4
4
I
I
Retest
0
2.5
9
25
0
2
4
6
3
2
30
46
10 Test
Retest
51
47
33
29
19
20
3
5
13
16
4
2
I
3.5
7
7
82
95
I 1 Test
Retest
27
31
22
20
7
14
5
6
29
40
5
6
0
5
8
4
74
95
12 Test
25
35
22
26
5
I1
2.5
10.5
6
6
I
0
3
4
7
9
51
77
Respondent
Number
Retest
3 Test
Retest
4 Test
Retest
5 Test
Retest
6 Test
Retest
7 Test
Retest
8 Test
Retest
9 Test
Retest
3
14
0
2.5
6
I2
53
Документ
Категория
Без категории
Просмотров
2
Размер файла
822 Кб
Теги
00007, 00004519, 197909000
1/--страниц
Пожаловаться на содержимое документа