DATA EVALUATION AND METHODS RESEARCH

Development of theSpeech Reception Test

Description of the development for use in the Health and NutritionExamination Survey of 1974-75 of a new test of hearing discrim-ination ability for speech sounds based on the revised Central In-stitute for the Deaf Sentences devised by a working group of theCommittee on Hearing and Bioacoustics of the National ResearchCouncil.

DHEW Publication No. (HRA) 78-1345

Series 2Number 71

US. DEPARTMENT OF HEALTH, EDUCATION, AND WELFAREPublic Health Service

Health Resources AdministrationHyattsville, Md. October 1977

Library of Congress Cataloging in Publication Data

Elkins, Earleen. Development of the speech reception test.

(Vital and health statistics: Series 2, Data evaluation and methods research; no. 71) (DHEW publication; no. (HRA) 781345)

Bibliography: p. 1. Audiometry, Speech. I. Causey, G. Donald, joint author. II. Roberts, Jean, joint author.

III. Title. IV. Series: United States. National Center for Health Statistics. Vital and health statistics: Series 2, Data evaluation and methods research; no. 71. V. Series: United States. Dept. of Health, Education, and Welfare. DHEW publication; no. (HRA) 78-1345. [DNLM: 1. Hearing tests. 2. Hearing disorders-Diagnosis. W2 A N148vb no. 711 RA409.u45 no. 71 [RF294.5.S6] 312.‘07’23s

ISBN o-84060088-7 [617.8‘9] 77-2800

For sale by tho Superintendent of Documents, U.S. Government Printing Offhx Washington, D.C. 20402

Stock No. 017-02240589-l

NATIONAL CENTER FOR HEALTH STATISTICS

DOROTHY P. RICE, Director

ROBERT A. ISRAEL, Deputy Director

JACOB J. FELDMAN, Ph.D., Associate Director for Analysis

GAU F. FISHE~ Associafe Director for the Cooperative Health Stat fitics System

ELIJAH L. WI-DTE,Associate Director for Data Systems

JAMES T. BAIRD, JR., Ph.D., Associate Director for International Statistics

ROBERT C. HUBER, Assoctite Director for Mawgement

MONROE G. SIRKEN, Ph.D., Assoa”ate Director for Mathematical Statistics

PETER L. HURLEY, Assoctite Directorfor Operations

JAMES M. ROBEY, Ph.D., Assoctite Director for Program Development

PAUL E. LEAVERTON, PhJ1., Associate Director for Research

ALICE HAYWOOD, Information Ofiicer

DIVISION OF HEALTH EXAMINATION STATISTICS

ARTHUR J. McDOWELL, Director

JEAN-PIHULE H&BICHT, M.D., Ph.D., Special Assistant to Director

JEAN ROBERTS, Chtif Medical Statistics ~rtmch

ROBERT S. MURPHY, Chief Survey Planning and Development Branch

COOPERATION OF THE U.S. BUREAU OF THE CENSLE

Under the legislation establishing the National HeaIth Survey, the Public Health Service isauthorized to use, insoh as possible, the m-vias or fzcilitics of other l?edcral, State, or privateagenacs.

h acmrdamcc with specifications ~tablished by the Division of Health Interview Statistics,the Bureau of the Census, under a conhaeturd arrimgcmentj pm-dtipate d in plauning tbe surveyand colkcling the data.

Vital and Health Statistics-Series 2-No. 71

DH EW Publication No. (H RA) 78-1345Library of tingrecs Catakg Card Number 77-2800

CONTENTS

Introduction ........................... .. .... ............ ................................ ...... ........................ ..........................................

Wc&omd ....... .............. ...... ..................... ................. .............. ................. ..... ...... .............................................

Development of Test Mttids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....Dmcription of Stiti ........................... ........................m............ .... .. .......... ..m... ....... .....................................Tdkm ..... ...... ........ .......................... ...................................... ................................. ......................................Recording Eqtipmcnt ............................ ..... ....... ................... ............ ............................... ...... ..... ..................Tape Assembly Rocedwes .... ........................ ............. .................................. .............. ...... ....... .... ..................Verification of Stitdi ............................................. ........ ....... .... .................... ..... ............. ..... ........... .. ......... ..Quality of Recording ....... ................................................................. .............. ............ ............ .......... ........ ....

Pilot Study ...... ...... .............. ....... ......................... ............ ............ ................................... ...................................

Study Among Normsl-Hearing Subje~s ............................. ................. ...... ................................... ...... ...............Subjects and Metiod ........ .................... ..... ....................................................... ...... ...... ....... ... .....................Ftid~ ..................................... ....................... ............. ........................................ .......................................

Study Among Sensorineursl-Hearing-Impaired Subjetis ............ ...... ........................................... ..................... ...Subjeti ........................................... ..... .................... ................. .................... ...... .... ... ..................................Mefiod ................... .............. ................. .................. ...... ............ .............................. ....................................Findings ...... .... ................................. ................................ ...... .... .. .... ........ ....................... ....... .......................

Intcrht Equivalency Smdy ......... .......................................... ...... ........................ .................................. .............Subject~ ....... .... . ...... .................................. ........................ ......................................... ..................................Metiod . . ...... . . .... . .. .. ... .. .. .. . .. .. .. .. . ... .. .... .. .... . . ... .. . .... . . .... .. .. .. . ... .... .. ... . .. .. . . .. . .. . .. .. . .. ... . .. ... ..................................FindirsgI.................................. ......... .... ....... ........................ ....... ..... ...................................................... ........

Discussion .. ..... ....... ..................................... ................................................ ........... ...................... ......................Normal U ..... ....... . ............. .......... .................................... ..... ....... ........... ...........................................Hearing Impaired .... ....... ................................... .............. ........... ...... ................................ .. ..... ..... ......... .......Intcrlist Eqtivden~ .. ...... ....................... ..... ...... ...... .... .................. ............................................. ................

s~ ...........................................................................................................................................................

References ............................ ................................ ...... ........... .............................. ...... ........... ....... ......................

1

1

2222444

5

556

7799

10101010

11111212

12

13

Appendm. Hearing Characteristics of the Sensorineural Hearing-Impaired Subjects in the Study ...................... 16

LIST OF FIGURES

1-1. schematic diagram of recording equipment and its placement .................. ................. ...... ....................... 3

1-2. Diagram of talker, examiner, snd microphones in recording &mber ... .... ................... .... ..... .................. 3

2. Performance-intensity functions derived horn mean discrimination scores on tbe RCID Sentence Gritsby 100 norsnahhcaring snd 55 impskcd-hearing mbjetis ..... ................................. ... ...................... ...... ... 6

...Ill

LIST OF TEXT TABLES

1. Mean, standard deviation, and standard error of the mean obtained with 10 lists of RCID Sentencesfrom 100 normid-hearing subjects at 8 presentation levels .. . .. . ... .. .... . .. ... ... . .. .... . ... .. . .... .. ... . .. .. ... .. .... ... . . .... . 6

2. Pearson correlation coefficient from 100 normal-hearing subjects’ scores on 10 lists of RCID Sentencesat 8 presentation levels ...................................................... .............. .................... .............. ....... ....... ........ 7

3. Mean, standard deviation, and standard error of the mean obtained on each of 10 lists of RCID Sen-tences from 10 normal-heming subjects at 8 presentation levels .. .. . ... .. . .... . ... . .. ... ... .. ... . ... .. . .. .. .. .. .. ... .. . .. .. .. 8

4. Mean, standard deviation, and standard error of the mean obtained with 10 lists of KID Sentences at12 sensation levels for the 55 subjects with sensorineural hearing impairments, and the number con-tributing to each mean ........ ................ .................. ........ ........ ...................................... ............ .... ........... 9

5. Mean, standard deviation, and standard error of the mean obtained on RCID Sentence lists by 10 sub-jects at 10 dB SL and 10 subjects at 20 dB SL, and by both groups combined .. . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 11

SYMBOLS

Data not available ------------------------------------ ---

Category not applicable ----------------------------- . . .

Quantity zero ------------------------------------------- -

Quantity more than O but less than 0.05---- 0.0

Figure does not meet standards ofreliability or precision ------------------------ *

iv

DEVELOPMENT OF THE SPEECH RECEPTION TEST

EarIecn Elkins, Ph.D., and G. Donald Causey, Ph.D., University of Maryland, andJean Roberts, Dizuk_onof Health Examination Statistics, NCHS

INTRODUCTION

The hearing component of the Health andNutrition Examination Survey in 1974-75 wasdesigned to obtain data on the discriminationability for speech sounds and the relationship ofthis ability to pure-tone air-conduction hearingthresholds among adults. These data, which willhave been obtained for the first time with stand-ardized methods on a national probability sam-ple of civilian noninstitutionalized adults 25-74years of age in the United States, will make itpossible to assess objectively the functional im-plications of hearing impairment in the adultpopulation. Prior to the start of the Heakh andNutrition Examination Survey there was generalagreement that one or more measures empIoyingspeech stimuli are necessary to quantify an in-dividual’s abiIity to understand suprathreshoIdspeech.1 Recent research has shown the extentto which specific types of hearing loss affect theintelligibility of speech and tlhat amplification(such as with hearing aids) does not restore fulldiscrimination when there is sensorineural IOSS.2

T& new speech reception test provides anobjective instrument, not previously avaiIable,for the determination within a short period oftime of the abiIity of adults to hear and under-stand everyday conversational speech. This de-velopmental study further provides some indica-tion of the limitation and precision of this typeof survey measurement in comparison withIonger clinical tests of speech discrimination.

BACKGROUND

The d hoc advisory committee for the hear-ing component of the examination had recom-mended for use in speech discrimination testingthe Revised Central Institute for the Deaf(RCID) Sentences, that had been developed by aworking group of the Committee on Heming andBioacoustics (CI-L4BA) of the National ResearchCouneiI. The R(XD Sentences, while not yet ona suitable recording, were recommended as pro-viding a better indication of functional hearingimpairment within the time limits available forthis component of the examination than wasposslMe with the test material commerciallyavailable.

Recorded materiak for speech discrimina-tion testig, then commercially availabIe, con-sisted of lists of 50 monosyllabic words. ThosedeveIoped at the Harvard Psychoacoustic Lab-oratory, known as PB-50 lists,3 were compiledwith strict adherence to phonetic composition.Later, Hirsh and his associates at the Central In-stitute for the Deaf4 recorded four 5O-word listscompiled with primary emphasis on a restrictedrange of word famdkity. These recordings areknown as the l\~-22 lists. A considerakde numberof research projects have empIoyed these tworecordings and speech discrimination ability hasbeen clinically assessed with them. No test or setof tests has been generaIIy recognized as accept-able to replace the IV-22 and PB-5 O commercialrecordings which are more than 20 years old.Rerecording and adaptation, incIuding shorten-

1

ing of these tests, would also have been neces-sary for use in the national survey.

Dr. Hallowell Davis, of the Central Institutefor the Deaf, made the lists of RCID Sentencesand developmental materials from CHABA avail-able to the Division of Health Examination Sta-tistics, to have adapted for their use.

Under contract with the Division of HealthExamination Statistics, Dr. G. Donald Causeyand Dr. Earleen Elkins, University of Maryland,recorded the RCID Sentences and developed thespeech discrimination test used in the Healthand Nutrition Examination Survey during1974-75. Their methods, findings, and the re-stitant test based on the RCID Sentences, aredescribed in this report.

DEVELOPMENT OFTEST MATERIALS

Description of Stimuli

The RCID Sentences supplied by Davis con-sisted of 10 lists of 10 sentences each. The fol-lowing criteria were employed by the workinggroup of CHABA in the development of theselists:

Vocabulary appropriate to adults.

Words that appear with high frequency in one ormore of the well-known word counts of the Englishlsnguage.

Exclusion of proper names and proper nouns.

Free use of common nonslang idioms and construc-tions.

Avoidance of phonetic loading and tongue twisting.

High redundancy.

Low level of abstraction.

Grammatical stiucture that varies freely.5

Each list of 10 sentences contains 50 keywords so that a discrimination score based on 2percentage points per key word may be ob-tained. These sentence lists were an attempt toprovide speech discrimination stimuli within aconstruct more closely associated with everydayspeech.

Talker

A 24-year-old male, determined by a panelof hearing and speech scientists to have normalspeech in relation to fundamental frequency,articulation, general American dialect, rate, andprosodic features, was selected as the talker. Hewas unsophisticated with regard to monitoredlive-voice technique for audiometric testing, butwas extensively trained in the technique prior tothe final recording session. This procedure in-volved the monitoring of his vocal output on thecarrier phrase of the item and allowing the stim-uli to occur naturally within the sentence, ire.,the level within each sentence varied as thespeech power of its component sounds varied.The carrier phrase provided a reference levelaround which the elements of the stimulus (oc-curred in their natural speech power relation-ships. The carrier phrase also served to preparethe listener for the cmcoming test item duringactual test presentation.

The voice monitoring was accomplished withthe phrase “Number__.” The talker maintainedoptimal needle deflection on the first syllable of“number” and on the first or only syllable ofthe digits 1 through 10.

All RCID Sentence stimuli were recordedduring one session in an effort to minimize day-today vocal variability of the talker. When astimulus sentence was judged unacceptable, itwas repeated immediately or at the end of a listand subsequently spliced into correct order.

The timing of each item was regulated by anautomatic device which triggered a light every6.5 seconds to alert the talker. This permitted ~anaverage response time of 5 seconds which hadbeen found adequate in a preliminary study ofmen from ages 38 to 78 years.

Remrding Equipment

The RCID Sentence recordings were made inan anechoic chamber (Industrial Acoustics Co’m-

pany, Inc.–IAC) with inside dimensions of 7feet by 7 feet by 7 feet. Four microphones(Bruel & Kjaer, Type 413 1) and cathode fol-10 wers (Bruel & K.jaer, Type 2619) weremounted in free air within the chamber. Microp-hone No. 1 was 24 inches from the talker’s lipsat a 30° angle to the ri~ghtof center; microphone

Microphones and 1 Microphone Tape

Cathode FollowersI

Powr Supplies Recorders

9

9 000 Q(2@Q–i @-m

000 Oomaoo

1Monitor I

RMS Aud10voltmeter i

Monitor

i

Speaker

*H AmplifierI

SpeakerMicrophone

1

9AN ECHOIC CHAM9ER CONTROL ROOM

Figure 1-1. Schematic diagram of recording equipment and its placement.

No. 3 was 1 meter from the talker’s lips at a 20°angle also on the right side; microp~one No. 4was placed as No. 1 but on the left side; andmicrophone lNo. 2 was placed as No. 3 buton the left. All microphones were at a 90° angleof incidence with respect to the talker’s lips (fig-s-u-es1-1 and 1-2).

Outside of the chamber, each microphonecomplement was electrically coupled to its re-spective power supply (Bruel & Kjaer, Type2801 ) and fed to a sepuate ampIifier and single-channel of two magnetic ‘A-inch tape recorders(Scully, Type 280). The outside channels of atriple-head tape assembly were employed tominimize the possibility of cross-talk (see figurel-l).

The talker visually monitored his 170cal out-put on arms (root-mean-square) audio-voltmeter(Bruel & Kjaer, Type 2410). A ta.lkback systempermitted communication between the chamber

-::’;’”,l.y’Y:p-

/’” /

0’ //’, // /J)Microphone 1

Examiner / ‘,’

/

//

/, “

o

Talker

Figure 1-2. Diagram of talker, examiner, and microphones in

racording chamlmr.

3

and recording technicians. An investigator waspresent in the chamber with the talker to aid inmonitoring and in timing.

The record gain of each channel was ad-justed to a -7 dB (decibel) re zero W (volumeunits) level to avoid saturation during peak vocaloutput. All recording was done on low-noise tape(Scotch Brand, Type 208) at 7?4 inches per sec-ond (ips). The signal-to-noise ratio exceeded 50dB .

Tape Assembly Procedures

A group of hearing and speech scientistsjudged the recording on channel 3 as “soundingthe most natural.” AU further procedures wereaccomplished with this recording. Half-inch tapecopies were made of the master recording. T&fist generation tape was then spliced for theappropriate stimuli, and leader tape was insertedbetween items to ensure equal intends of re-sponse time.

Tracings of the stimuli of each list weremade on a graphic level recorder (General Radio,Type 1521-B) at 20 ips writing speed in orderthat the average level of the peaks could be cal-culated. A 1000-Hertz (Hz) tone generated byan oscillator (Hewlett-Packard, Model 200 AB)was placed at the beginning of each list at theaverage peak level to be used as a calibrationtone in administration of tbe test lists. A secondgeneration %nch tape (Scotch Brand, Type208) was then made for experimental purposes.

Verification of Stimuli

To determine that each stimulus was the in-tended one, the 10 Iists were presented soundfield to 16 normal hearing college students in asemi-sound-treated classroom. The ambientnoise level was 60 dB Sound Pressure Level(SPL) as measured on the C-scale of a sormd-level meter (Bruel & Kjaer, Type 2205)- 15 feetin front of the speaker. The level of the 1000 Hzcalibration tone at the begirming of each list wasadjusted to obtain a 65 dB SPL reading on thesame meter. All lists were presented at this levelfor the initial session.

Listeners were seated in front of tbe speakerat a distance of 5 to 25 feet. They were in-structed to write the entire sentence on prepared

answer sheets. Ample time was provided aftereach sentence for the write-dowm task. Two dayslater, the same 16 listeners performed the sametask again, sitting in tke same position relative tothe speaker as they had for the initial session.Two changes in procedure were made. First, thelevel of the 1000 Hz calibration tone was ad-justed to 60 dB SPL on the sound-level meter,and second, the ordler of list presentation waschanged. Each session took 1 hour and 10 min-utes. Following the second session, the listenerswere asked to write their reactions to both thetask and stimulus material.

The results from both sessions were markedfor the number of correct key word responses.None of the listeners had more than one errorper list, nor were any of the errors consistentamong listeners. It was concluded that the listsall contained the correct stimuli and were notdtiferent from each other when perceived bynormal-hearing listeners at a comfortable listen-ing level.

The reactions of the listeners were directedtoward the repetitive and boring nature of thetask and fatigue fro m writing responses. Com-ments about the stimulus material were few, andthose listeners reported that the content did notaffect them one way (oranother.

Quality of Recording

The quality of a taped copy of this masterrecording of the RCID Sentences was analyzedby Mr. Edward D. Burnett, Physicist, Sound $ec-tion, National Bureau of Standards.

His report shows the broad-band noise to be47 dB below the reference tone (which is -9 dBre O VU) when the tapes are played with a fulltrack head. This is 4 dll better than predictedfor a fourth generation tape. The type 208 tapeis well suited for this application. The referencetone indicates the peak levels as measured at the20 inches/second w-ding speed of a graphic levelrecorder. The instantaneous peaks, as observedon an oscilloscope, am at least 15 dB higher. Thespeech waveform is highly asymmetrical as iscommon for male speech. The level of the signalon the tape is well chosen.

The reference tone does not show any appre-ciable second-order (distortion which indicatesthat the bias wavefonm was symmetrical and no

4

heads were magnetized. No indications of over-load distortion were found. The lower frequencylimit is 100 Hz which is a characteristic of thespeaker’s voice. Components up to 14,000 Hzwere detected, which indicates that heads werein good alignment.

PI LOT STUDY

Prior to the experimental study, 10 normal-hearing subjects with normal orological historieswere utilized to determine the method and sen-sation levels necessa~ to obtain performance-in-tensity functions for the RCID Sentences. Thecriterion for normal hearing was a pure-tonethreshold of 10 dB or less (ANSI, 1969) at250, 500, 750, 1000, 1300, 2000, 3000, 4000,md 6000 Hz.

The experimental task was performed withthe subject seated in the anechoic chamber de-scribed earlier. The test materkd, reproduced bya tape recorder (ScUUy, Type 280), was chan-neled through a speech audiometer (Grason-Stadler, Model 162) to one earphone of a set(Telephonic, TDH-39 in JIX--ll/AR cushions).The same earphone ~~zis used throughout thetests. The 1000 Hz calihr~tion tones on the RCIDtape were used to establish O dB sound pressurelevel (SPL) on the speech audiometer. Periodiccalibration checks were made of the equipmentthroughout the period of data collection.

Prior to the test session, subjects were toldby the examiner that they would hear Iistsaofsentences at different signal levels and were torepeat all or any of the ~entence. The give-backmethod of response ww chosen as a result of thestimuli verification study and the time factor.Subjects dso received ~he instructions, prr-rc-corded on t~pe, ~vhich c(lnsisted of:

You will hew ten sentmces, each preceded by anumber. Please listen carefully and repeat only thesentence.

A systematized randomization of the listswas employed so that c~ich list was presented anequal number of times at the same SPL acrosssubjects. The first five subjects in the pilot studyreceived two sentences at each level beginning at34 dB SPL and decreasing in 2 dB steps. All

subjects repeated the first two sentences cor-rectIy. The descending method of presentationproceeded in 2 dB steps to 16 dB SPL. At thislevel all subjects were unable to respond to anyportion of the sentences. The remaining eightlists were then presented at sound pressure Ievelsof 18, 20, 22, 24, 26, 28, 30, and 32 dB in anascending fashion. The examiner recorded all re-sponses, including substitutions, on an answersheet containing the appropriate list stimuli. Allfive subjects obtained O percent correct at 18 dBSPL and 100 percent correct at 32 dB SPL.

The next five subjects were trained by thesame descending method with two lists, but re-ceived the following eight Iists in a descendingmode starting at 32 dB SPL and ending at 18 dBSPL. The percentages of correct responses weresimilar to those obtained with the ascendingmethod; however, these subjects reported a gen-eral frustration with their inability to make cor-rect responses as the signal was attenuated. Dur-ing the performance of this task, four of the fivesubjects verbally responded to each stimulusthat they could not identify completely with aphrase similar to “1 can’t make it out. ”

Based on the results and subjective evalua-tions of the task, two lists were used for trainingthe subject and the remaining eight lists for ob-taining articulation-gain functions by an ascend-ing method.

STUDY AMONGNORMAL-HEARING SUBJECTS

Subjects and Method

One hundred college students–7 males and93 females between the ages of 18 and 25 years–\vere selected for participation in this experi-ment. Each was questioned about his orologicalhisto~ and screened audiometrically as in thepilot study. Only those u-it.h normal orologicalhistory and audiometric test results were se-lected for this part of the study. All testing wasdone monaurdl~”, znd an equal number of stu-dents ~vere tested in the right and in the left ear.I’erbal instructions ~vere given to each subject asin the pilot study, and the;” also rrcriled the re-corded instnsctions (at 50 dB SPL) prior to each

5

list. A systematized presentation of the lists wasemployed so that each list was presented an equalnumber of times as training lists and test lists atthe same signal level across subjects. Using theequipment described earlier, each subject re-ceived the stimuli of the first two lists to ac-custom them to the task. Four sentences werepresented at 34 dB SPL and two sentences eachat 32, 30, 28, 26, 24, 22, 20, and 18 dB SPL.Each of the remaining eight lists were presentedin full at each presentation level, beginning with18 dB SPL and increasing in 2 dB increments to32 dB SPL. All responses were made verballyand recorded by the examiner in the mannerdescribed earlier. The percentage of correct keywords was calculated for each list. The test ses-sion took 35 minutes.

Findings

Table 1 summarizes the data obtained with100 normal-hearing subjects combined acrosslists. Each list was heard by 10 subjects at eachpresentation level, The performance-intensityfunction derived from these data is shown infigure 2. This curve has the characteristic config-urate on of traditional performance-intensityfunctions derived from monosyllabic stimuli inthat it exhibits a linear function which reaches aplateau at higher sensation levels. The lower seg-ment of the cmwe is linear and has a slope of10-percent increase in correct responses per deci-bel increase in presentation leveI. This linear seg-ment terminates at about 24 dB SPL, where dis-crimination scores are about 76 percent. The

Table 1. Mean, standard deviation, and standard error of thamean obtained with 10 lists of RCID Sentences from 100norrml-hearing subjects at 8 presentation levels

Presentationlevel

(dB SPL)

18da b.... .. ... . .. ... .. ... .... .. ... .20dB ... .. .... .. .... .. ... ... . .. .... .22 dB ..... .... .. . .. ... .. .. ... .. .. ...24 da b... .. ... .. .. ... .. .. . .... .. ....26 da b...,...,,., . .. .. .. . . .... .. .. ..28 dB .. ... . .... .. .. .. ... .. .. .. ... ...30 dB ..... . .... .. .. . ..... ... ... .. .. .32 dB .... .. . .. .... .. . ..... .. .. .. ... .

Mean

16.6636.7B

58.1275.94

87.2494.6297.9098.BO

Standard

deviation

17.4624.6823.2720.24

14,857.363.902.07

Standard

error of

meen

1.742.472.332.021.490.740.390.21

da SOLIND PRESSURE LEVEL (FDR NORMAL. HEA181NGI

Dk~-2 D24EE1D 12141e

dB SENSATION LEVEL (FOR lMPAIflED-HEARING)

Figure 2. Performance-intensity functions derived from mean

discrimination scores on the RCID Sentence Lists by 100normal-hearing and 55 impaired-hearing subjects.

upper portion of the function progresses in acurvilinear marnmr until it approaches asymp-tosis, and almost perfect discriminatiom~ of 99percent is achieved at 32 dB SPL.

The measures of score variability, as indi-cated by the standard deviations in table 1, alsocompare favorably with the literature on moncs-

‘tisyllabic discrimination. Excluding the variabilityamong scores at 118 dB SPL, where the range ofscores is truncated at zero, the variability wasfound to be large within the linear portions ofthe function. Variability decreased progressivelyas the signal intensity increased to the pointwhere almost perfect responses were made. Thevariability present at the highest presentationlevel probably reflects the subjects’ random er-rors due to physic) logical factors or lack of atten-tion.

As an indication of subject consistency,Pearson correlaticm coefficients were calculatedto determine the relationship between indi-viduals’ scores at the different levels of presenta-tion. The coefficients shown in table 2 revealthat scores obtained under presentation levelsadjacent to one another show higher correlations

6

Table 2. Pearson correlation coefficient from 100 normal-hearing subjects’ =or- on 10 listsof RCI D Sentencesat 8 presentation levels

I

I Presentation level (dB SPL)Presentation level

(dB SPL)

18 dB .. ... ........................20 da b..............................22 dB ...............................24 dB ...............................26 dB ....... ......................28 dB ...............................30dB ...............................

20 22 24 26 28 30 32

.1.629 1.570 1.451 1.450 1.331 .157 .181

1.613 1.611 1.451 1.387 1.202 .141

1.656 1.520 1.456 .098 1.212

1.610 1.568 1,266 .1371.6% 1.301 .096

1.521 ~.185.235

ls,gnifjjant difference from zero at .05 lweL

than scores obtained under levels 4 to 12 dBapart. This trend would be expected in view ofthe influence of signal intensity differences.Each column or row follows a logical decrease inthe size of the correlation coefficient as the sig-nal intensity difference increases. The generaIprogression of decreasing coefficients as thepresentation levels increase probably reflects themuncatirtg effects when scores beyond 100 per-cent cannot be obtained. These data indicate thatsubjects maintained systematic relationshipsamong themselves at different sound pressureIevek.

In an effort to evaluate interlist equivalency,the scores of 10 subjects for each of the 10 listsA through J at each presentation level weredrawn from the data described above. The re-sults are shown in table 3. It is apparent fromthese data that at the lower levels, particularlythose identified in the linear portion of tiecurve for 100 subjects (through 24 dB SPL), themean and standard deviation values are quite dif-ferent. Generally, there is a range of 30 percentfrom the lowest to the highest mean at theselevels. Not until the scores reach 90 percent orbetter, do the mean and standard deviationvalues appear similar. Tests for significant differ-ences among means revealed that the lists weredifferent through 28 dB SI’L. The means of the10 lists were not significantly different (.05 levelof confidence) at 30 and 32 dB SPL.

STUDY AMONGSENSORINEURAL-HEARING-

IMPAIRED SUBJECTS

Subjects

Fifty-five male military veterans served assubjects for this part of the study. The age rangewas 22 to 63 years with a mean age of 45 yeas.It was the intent of this part of the study togather data on all patients seen at the Audiologyand Speech Pathology Section of the VeteransAdministration Hospital, Washington, D.C. whoshowed any degree of sensorineural hearing im-pairment. Consequently, no rigid criteria forsubject selection were required. Speech recep-tion thresholds (SRT) could be as low as 2 dBabove hearing threshold level (HTL) and pure-tone air conduction thresholds could be withinnormal range with a drop of 25 dB HTL at onlyone test frequency. The group mean SRT for thetest ears was 17 all!. Generally, the audiometricconfigurations of the test ears were character-ized by normal hearing in the lower frequencieswith precipitous losses at various octaves above1000 Hz. Thirty-one percent of the test popula-tion had a flat or gradually sloping configura-tion. The remaining subjects showed a drop of atleast 25 dB starting at the following frequencies:16 percent at 1500 Hz, 13 percent at 2000 Hz,35 percent at 3000 Hz, and 5 percent at 4000Hz.

7

Table 3. Mean, standard deviation, and standard error Of the mean Obtained On each Of 10 lists Of RCI D Sentences from 101 normal-hearing subjects at 8 presentation levels

Prs5entation level(dB Sp L)

18dBMean .. ... .. .. ... .. .... . .. . ... ..Standard deviation .. ... .Standard error .... .... .... .

20 dBMean .. ... .. ... ... ... ..... .. . ...Standard deviation .. .. ..Standard error ... . .. .... ..

22 dBMean ... ... ... .. ... .. ... ... . ... .Standard deviation .. .. . .Standard error . ... . .... . .. .

24dBMean .. ... .. .. .. .. .... .. .. .. ... .Standard deviation .... ..

Standard error .... .. .. ... . .

26dBMean ...... .. ..... .. .. .. .... ....Standard deviation .. .. ..Standard error ... ..... .. .. .

28 d8Mean ... .. ... .. .. .... . ... .. .... .Standard deviation .. . .. .Standard error ... ... .. .. .. .

30dBMean ... ... . .... ... . ... . ... .... .Standard deviation ..... .Standard error ... .. ... ... ..

32 dBMean ... .... . . .. .. ... . ..... .. . ..

Standard deviation .. ... .Standard error . ... . .... .. ..

A

5.206.552.07

39.4027,57

8.73

64.8022.37

7.08

75.6013.98

4.42

93.406.261.98

95.405.501.74

95.409.052.86

99.60

0.840.27

B

17.8018.17

5.75

44.6027.7B

8.79

72.6023.04

7.29

90.2010.81

3.42

93.008.182.59

99.002.160.68

99.80’0.630.20

99,00

2.160.68

c

26.0021.46

6.79

37.2031.00

9.81

67.8022.91

7.25

89.808.872.81

96.004.521.43

98.801.690.54

98.402.700.85

99.40

0.970.31

D

27.8021.18

6.70

48.4023.64

7.48

55.0025.94

8.21

76.40

18.525.86

94.203.050.97

97.203.551.12

98,201.750.55

9B.80

2.150.68

List of RCI D Sentences

E

8.008.592.72

37.2022.37

7.08

41.6020.08

6.35

62.6024.33

7.70

71.6015.85

5.02

84.006.932.19

88.402.460.7B

97.403.781.20

F

12.0015.06

4.77

18.0012.58

3 .9B

55.2024.93

7.89

59.4024.73

7.83

77.2024.66

7.80

90.009.613.04

98.002.9B0.94

98s30

2.670.84

G——

27.6020.39

6.45

44.2030.13

9.54

64.4016.97

5.37

88.009.573.03

88.608.172.59

93.20B.342.64

97,202.860.91

‘99.40

0.970.31

H

5.207.612,41

25,6019.97

6.32

43.0021.61

6.84

60.0023.87

7.55

85.8021.15

6.69

90.806.271.9B

97.203.681.17

98.60

1.650.52

I

19.2016.23

5.14

22.BO15,67

4.96

67.0023.84

7.59

77.6017.33

5.48

87.8011,72

3.71

86.209.262.93

97.803.711.17

98.60

0.840.27

J.—

17.8016.04

5.08

46.6020.16

6.38

49.8014,37

4.55

79.0013.57

4.29

84.8012.63

4.05

92.2010.52

3.33

98.603.7B1.20

98.20

2.200.70

Following an orological examination, bilater- the 55 male veterans selected as subjects foral pure-tone thresholds by air and bone conduc- this part of the study on air and bone conduc-tion were obtained for each subject. Only those tion hearing’ tests and on the W-22 Iists are sum-who had an air-bone gap of 10 dB or less were marized by age in the appendix. Among this studyretained for further testing. This consisted of a group, the degree of negative association be-determination of their SRT’S and discrimination tween their scores on the W-22 word lists andscores on recorded W-22 lists.4 Performance of their air-conduction hearing level at 10010 Hz

(-0.59) is similar to that with their speech recep-tion threshoId (-0.61), i.e., those with lower orbetter hearing levels attain higher W-22 scores orhear more words correctly.

Method

The em with the better SRT was thenchosen for the monaural RCID Sentence testing.If the SRT’S for both ears were the same, the earwith which the subject had obtained the highestW-22 speech discrimination score was selected.

The task was performed with the subjectseated in one room of an IAC testing suite. Thetest material, reproduced by a tape recorder(ScuUy, Type 280), was channeled through aspeech audiometer (Grason-Stadler, Model 162)to one earphone of a set of phones (Telephonic,TDH 49-1OZ in hlX-41/AR cushions). The sameearphone was used throughout. The 1000 Hzcalibration tones on the RCID tape were used toestablish O VU output level on the W meter ofthe speech audiometer. Periodic calibrationchecks were made of the equipment during theperiod of data collection.

Prior to the test session, subjects were toIdby the examiner that they would hear lists ofsentences at different signal levels and were torepeat all or any of the sentences. Subjects alsoreceived the prerecorded instructions on the testtape. A systematized presentation of the listswas employed so that each list was presented agiven number of times as training lists and testlists at the same sensation level across subjects.The reference level for presentation was the SRTof the subject’s test ear. For the training por-tion, two sentences were presented at each ofthe follo~ving sensation levels (SL’S): 16, 14,12, 10, 8, 6, 4-, 2, 0, -2 clB. The remainingeight lists were presented in full, begirming withO or +2 dB SL and increasing in 2 dB incrementsto 14 or 16 dB SL. After data on 20 subjectshad been obtained, some variation in the initiaIsensation level was instigated in order to explore

the performance intensity function below O dBSL and above 16 dB SL. AU responses weremade verbally and recorded by the examiner inthe manner described in the study among thenormal-heting group. The percentage of correctkeywords was calculated for each list. The testsession took 35 minutes.

Findings

Table 4 summarizes the data obtained from55 subjects with sensonneural losses. The num-ber of subjects for which the data at any sensa-tion level were obtained is indicated in the lastcolumn. The performance-intensi~ functionderived from these data is shown in figure 2 withthe performance-intensity function obttined

from 100 normal-hearing subjects. The curve forsensorineural subjects exhibits the linear func-tion ~vhich appears to terminate at about 8 dBSL, where subjects gave 74 percent correct re-sponses. The slope of this portion of the curverises at a rate of about 5 percent increase perdecibel increase in intensity. The upper portionof the function tends toward curvilinearity anddoes not approach asymptosis.

The variability of the discrimination scoresabout the mean values, as shown by the standarddeviations in table 4, indicates the large amountof heterogeneity among the impaired-hearingsubjects. It is also important to observe that thislarge variability is maintained from the lowersensation levels up to 14 dB SL. A gradual de-crease in subject variability is then observed buta pattern is not apparent. Standard errors of themean indicate that in other populations withsimilar sensorineuraI losses, one would expectabout two-thirds of the mean discriminationscores to be pIus or minus two key words.

Table 4. Mean, standard deviation, and standard error of the

mean obtained with 10 lists of RCID Sentences at 12 sen-sation levels for the 55 subjects with sensorineural hearingimpairments, and the number contributing to each mean.

Sensationlevel

4 dB ....... .....-2 da b.... ... ... .O da b............2dB .. ...........4dB .. . ... . ... ...6 da b.. . .. .. .. .. .8 dB .. .. ... .. . ...ICI da b.. .. ... .. .12 dB .. .. .. . ... .

14 da b..........16 da b..........18 dB ...........

Mean

14.3310.8930.5635.2749.2656.6573.6277.8285.4786.7793.5395.50

Standarddeviation

24.2110.3028.5928.1028.5832.1324.2023.7015.0920.15

8.624.50

StandardNumber of

error of

the meansubjects

9.88 63.43 95.72

5,13 :;4.83 355.51 343.98 374.13 332.76 303.95 262.09 171.59 8

9

INTERLISTEQUIVALENCY STUDY

In order to control the effect of variablelevels of stimuli presentation, another study wasconducted to evaluate interlist equivalency—assessing the same subject under different exper-imental conditions. There would be need inthe Health Examination Survey for sentencelists that could be used interchangeably, sincedifferent Iists of 10 sentences must be avail-able for use at each successive 10 dB inten-sity level at which the test needed to be pre-sented. The presentation was to be continueduntil the examinee was no longer able to hearenough of the speech test to pass.

Subjects

Subjects for this experiment were twen-ty male military veterans with normal or im-paired hearing. Their ages ranged from 23 to 69years, with a mean age of 45.7 years. The sub-

jects were divided into two groups of 10each. Group 1 had speech reception thresholds(SRT’S) in the test ear ranging from 2 to 24 dBwith a mean of 9.4 dB. Their lower pure toneaverages (PTA’s), usually two frequencies (500

and 1000 Hz), ranged from 2 to 30 dB with amean of 10.2 dB. Group 2 had SRT’S from O to34 dB and a mean of 15.2 dB. The Iower PTAfor these subjects went from O to 37 dB with a

mean of 16.4 dB.No restrictions with respect to degree, type,

or presence of hearing impairment were placedon subject selection. Five of the subjects couldbe classified as normal, two had mixed impair-ments, and the remainder had sensorineural im-pairments. All but three of the subjects hadW-22 discrimination scores of 90 percent or bet-ter in the test ear.

Following an orological examination, bilater-al pure-tone thresholds by air and bone conduc-tion were obtained for each subject as well asspeech reception thresholds and discriminationscores on recorded W-22 lists. The test ear waschosen on the basis of the lower PTA of thebetter ear for monaural testing with the RCIDsentence lists.

Method

The experimental task was performed w-ii.hthe subject seated in one room of an WC testingsuite. The test material, reproduced by a taperecorder (Scully, type 280), was chamieledthrough an audiometm- (Allison, Model 22) toone earphone of a set (Telephonic, TI)H39-1OZ in MX-41/AR cushions). The sanie ear-phone was used throughout the test. The 1~00HZ calibration tones on the RCID tape were

used to establish O (cm the VU (meter of theaudiometer.

Prior to the test session, subjects were toldby the examiner that they would hear lists ofsentences at a comfortable loudness level tidwere to repeat all or any of the sentences. Sub-jects also received file prerecorded instructionson the test tape. A systematized presentation ofthe lists was employed so that subjects in Group1 or 2 each heard a different order of the sen-tence lists. The reference, level for presen~ati@was the lower PTA of the subjects’ better ear.Group 1 heard all I.O lists at 10 dB -SL andGroup 2 heard all 10 lists at 20 dB SL. No train-ing lists were administered because all stiru&were presented at a comfortable loudness leveland any learning factor would be countedml-anced by the randomization of the lists. AU re-sponses were made verbally and recorded by the

e-xaminer in the manner described earlierj ‘l%epercentage of correct keywords was then czilcu-lated for each list. The test session took about35 minutes. ;.

Findings

Table 5 describes the performance of bothgroups of subjects separately and combined onthe 10 R(XD sentence lists. The mean conectresponses for Group 1 ranged from 88.8 percentfor list E to 98.2 percent for list I. The scores of,Group 1 were statistically evaluated by, anzikys(s-of variance. Statistical significance was ‘shownfor the main effect of Iists (F=2.67; df=9:,90;p<O.05). Subsequent comparisons by Duncm’sMultiple Range of the mean scores for 10 listsshowed that list E was significantly (p<Qi05)lower than all of the remaining lists w+ih the

10

exception of lists F and H. There were no signifi-cant differences among the means of the ninelists, except for list E.

The scores of Group 2 were also evaluatedby analysis of variance and no statistical signifi-

able 5. Mean, standard deviation, and standard error of themean obtained on RCID Sentence lists by 10 subjectsat 10 dB SL and 10 subjects at 20 dB SL, and by bothgroups combined.

RCID Sentenc-

Liti AMeml ... .. . ... .. . ..... .... ... .. ..Standerd deviation .... . . .Standard error .. .. ... . .... ..

List B

Mean .. . .. ... . .. .... . . .... .. . ... .Stantird deviation .. . .. ..Standrsrd error .. ... .... .. . . .

List cMean ... ... . ... .. .. .. .... .. . .... .Standati daviaton . . .. ....Stendard error .... .. ... .. . ..

List DMaan .. ... . .. ... .. .. .... .. ... .. ..Standard dawation .. .. .. .Standard arbor .. .. . .. .. .. .. .

Uat EMean .... . ..... . .. .... .. . ... .. . ..6tan&rd deviation .. .... .Strntird error .. ... .. ..... . .

List FMean ... .. . .. .. .. ... . . .... ... .. ..Standard deviation . ... . . .Standard error . .. .. ... .... . .

tist GMean .. . .. ... .. . .... .. ..... .. . ...Standerd deviation . ... . ..Standard error .. .. .. . .. .. .. .

List HMaan .. .. .... . ... ... .. . ... .. .. .. .Standard deviation .. .. ...Standard arbor.., . . .. ... .. . .

List IMaan .. .. .. .... . .. .. .. .... .. .. ...Standard daviation . .... ..Standard error, . . .. .. ... . .. .

List JMean ... .. .. . ... .... . . ... . . .. ... .Stsmdmd daviation . ... . ..Standard error . ... .. . ... .. . .

;roup 1atl O

dB SL

97.61.7440.552

97.22.8570.9M

96.2 ~3.8421.216

96.02.6830.849

88.89.4742.998

93.06.7682.142

96.25.0161.5s7

93.67.5262.382

9B.22.0880.661

97.24.4901.421

Sroup 2at 20

dB SL

98.62.3750.752

98.82.0400.645

98.22.2720.719

97.63.5551.125

98.4.1.2810.405

96.45.4891.740

97.8.2.6s300.823

96.65.1421.627

97.62.1540.682

95.85.1731.637

Groups 1and 2

98.12.1980.492

98.02.6750.588

97.23.3970.760

96.83.3340.746

94.18.8131.972

84,76.5621.468

97.04.1800.936

95.16.7891.519

97.92.1420.479

96.55.0211.123

cance was shown (F=O.94; df=9,90; p>O.05).The scores from the two groups were similar atthe 10 and 20 dB SL’S, and the F-test for homo-geneity of variance (HartIey’s) was performed onthe data to determine the feasibility of combin-ing the data into one group with an N of 20. Theresults of the F-test indicated that the variancedue to experimental error within each of thetreatment populations was homogeneous(F=2,22; df=2,9; p>O.05).

Finally, an analysis of variance performed onGroups 1 and 2 combined indicated no signifi-cant differences among the scores on the 10 sen-tence lists (F=l.66; df=9,190; p>O.05). Thelargest mean difference for Groups 1 and 2 com-bined is between Lists A and E and constitutesonly 4 percent or avo keywords.

DISCUSSION

Normal Hearing

Several factors probably contribute to theIarge differences among means and the wide vari-ability among scores at levels from 18 dB to 28dB SPL. One consideration would be the Iack ofhomogeneity among sentence items due to thelinguistic constraints imposed by possible wordsequences. c Quantification of sentence stimulifor the measurement of discrimination abilityhas not received much investigation, and someof those who have worked in this area feel thatsentence stimuli are “very complicated” and re-quire further study.]

Another factor which undoubtedly contri-buted to the observed differences among lists isrelated to the experimental method. It has beennoted that all subjects heard the lists at speeifiedlevels re 0.0002 microbar, and though they hadbeen screened for normal-hearing, it is generallyagreed among audiologists that there is markedvariability in sensitivity among persons meetingthe normal-hearing criteria. The level which mayhave been the threshoId of detectabdity or in-telligibility for one subject could easily havebeen displaced plus or minus several decibels foranother subject. This consideration is simplyanother way of viewing the marked variabilityamong subjects. However, as the data across alI

11

10 lists for the entire group of 100 subjects indi-cates, measures of a group of tests are morestable than measures for a single test.

The p e rf ormance-intensity function fornormal-hearing subjects is characteristic of func-tions obtained with monosyllabic stimuli in

~ that it contains a linear portion which reaches aplateau at higher signal intensities. The linearportion has a slope of 10 percent increase inscore per decibel increase in presentation level.This linear segment terminates at a level ofabout 24 dB SPL, where discrimination scoresapproximate 76 percent. Asymptosis occurs at alevel of 30 dB SPL.

For normal hearing subjects, these findingswould indicate that all 10 lists of the RCID Sen-tences may be used interchangeably for obtain-ing maxi mum discrimination scores fromnormal-hearing persons, provided the signal levelunder earphone presentation is at Ieast 30 dBSPL.

Hearing Impaired

Comparison of findings among the norrnal-hearing and impaired-hearing populations showmarked differences. First, the slope of the linearportion of the curve for the impaired-hearingsubjects was more gradurd-approximately 5 per-cent per decibel compared to 10 percent perdecibel for the normal-heting subjects. The lin-ear portion for both groups of subjects termina-ted where approximately 76 percent correct re-sponses were obtained. But this occurred at asensation level of 4 dB for normals and 8 dB forthe impaired listener.

The second difference between tie twogroups of subjects may be observed where thenonlinear portion of the performance-intensityfunction approached saturation. The normalgroup attained this level at 32 dB SPL and theimpaired group did not achieve a plateau at thesensation levels employed for this study.

The third difference between the normal andimpaired-hearing subjects is reflected in the sub-ject variability within each group. As indicatedby the standard deviations, the normal-hearinggroup had less variability throughout the sensa-tion levels examined than the impaired-hearinggroup. This would indicate that the impaired

group was more heterogeneous in the speech dis-crimination task than was the normal-hearinggroup, Also, the variability within the normal-heanng group decreased progressively as the sig-nal intensity increased to the point where almostperfect responses were made, but the same typeof pattern of intersubject variability was notmaintained by the sensorineur~ group.

Interlist Equivalency

The findings from this study indicate thatthe University of Maryland recordings of the 10RCID sentence lists could be used interchange-ably for the Health and Nutrition ExaminationSurvey as a measure of the extent of functionalhearing impairment for normal conversationallevels of speech in the adult.population. The dis-crimination scores obtained from subjelcts inGroup 1 at 10 dB above the PTA of the betterear and scores obtained from subjects in Group2 at 20 dB above the PTA of the better ear werestatistically evaluated. As a result of score simi-larity between the groups of subjects and a find-ing of homogeneous variances, the data fromGroups 1 and 2 were combined, No significantdifferences were found among any combinationof the 10 means. The standard errors of themeans indicate that with other populations, onewould expect about twe-thirds of the discrimi-nation scores to bc within plus or minus one ortwo keywords of the means obtained at 10) or 20dB sensation levels. Ay real differences amongthe lists would be apparent only if the sensationlevels of presentation were less than 10 dB andrepresentative of the linear portion of a subject’sperformance-intensity function for these mate-rials. ‘,

SUMMARY

These’ developmental studies demonstratethat the recordings of the RCID sentence listscan readily be used as a discrimination test toseparate listeners with normal hearing fromthose with senso]ineural losses. This new test,designed for use im the Health and Nutrition Ex-amination Survey of 1974-75, is capable of pro-viding a rapid a.messment of hearing discrimina-tion ability for speech sounds.

12

REFERENCES

/ lSecond Conference on Speech Discrimination, spon- rials for Speech Audiometry. ]. Speech Hear Disord.sored by the Office of Vocational Rehabilitation. 17:321-337, 1952.Houston, Texas, 1962. 5Davis, H., and Silverman, S. R.: Hearing and Deaf-

‘Personal communication from E. L. R. Corliss on ness, 3d ed. New York. HoIt, Rinehart and Winston, Inc.,Some Quantitative Consequences of Hesring Loss from 1970.the Standpoint of Communication Theory, 1975. 6Speaks, C., Jerger, J., and Jerger, S.: Performance-

3Egan, J. P.: Articulation Testing Methods. Laryngo - intensity ch aractenstics of Synthetic Sentences. J.

tsco e. 58:955-91, 1948. Speech Hear Res. 9:305-12, 1966.

Hirsh, I. J., Davis, H., Silverman, S. R., Reynolds, E.G., Eldert, E., and Benson, R.W.: Development of Mate-

13

APPENDIX

CONTENTS

Hearing Characteristics of the SeIISOrineuml Hearing-Impaired Subjects in the Study ..- . .. . ...-.. -------------- 16

APPENDIX TABLE

L Meao, standard deviation, and standard error of the mean on air- and bone-conduction hearing testsand W-22 word lists aud tbe correlation between akmnduction levels and W-22 sores for the 55 sub-jects with xnscmiuetual heruing impairments. ... . .. . . . . . . .. . .. .. .. .. .. .. .... . . . . . ... . . .... . . . .. .. .. .. .. ... . . . . 16

APPENDIX

HEARING CHARACTERISTICS OF THEHEARING-IMPAIRED SUBJECTS IN

SENSORINEURALTHE STUDY

The extent of hearing impairment for puretone sound in air- and bone-conduction, the sen-sation levels for speech reception, and thespeech discrimination level as determined withthe previously available W-22 word lists for the55 sensorineural-hearing-impaired subjects inthis developmental study and the degree of asso-ciation between the W-22 discrimination scores

and the pure tone air-conduction hearing levelsfor these 55 subjects are shown in Table I. Thesedata are included to give a general idea of theextent of hearing impairment for these subjects,although as previously stated no rigid criteria fortheir selection were used since all were patients inthe Audiology and Speech Section of the Veter-ans Administration Hc)spital in Washington, D,,C.

Table I. Mean, stendard deviation, and standard error of the mean on air- and bone-conduction hearing tests and the W-22 word lists andthe correlation between-air-conduction levels and W-22 scores for the 55 subjects with sensorilneural hearing impairments.

Air-conduction hearing levels (dB)

a

Bone-conduction hearing levels (dB) Speech testUumber

ofubjects 500 1000 2000 4000 500 1000 2CIO0 4000 w-22 SL re

Hertz Hertz Hertz Hertz Hertz Hertz Hertz Hartz scores SRT

—

55 55 55 55 55 15 15 21 26 55 55

Age

Number of subjects .. . ..

Mean values

All ages20-64 years .... . .. . .. 55 17.8 19.5 53.5 ] 29.7 I 29.7 40.931.3

24.425.025.040.230.6

36.7 53.3 91.8

20-29 years...,..., .. .. .. ....30-39 years .. .. .. .... ...._.40-49 years .. . .. .. .. .. .. .... .

50-54 years .. .. .. .... . ... ... .55-64 years .... .. .. .. .. . ... ..

87

1220

8

11.712.919.222.514.4

12.513.617.926.217.5

51.957.152.9

+63.565.6

42.557.545.0

+55.6+63.8

90.594.694.888.091.8

41.240.939.542.339.5

8.13

15.041.7

+32.9

20.0

12.5 36.232.5

+42.2

28.8

3B.3+35.0

20.0

Standard deviation

Al I ages

20-64 years .. .. .. ... . . . . I 14.03 I 17.29 1 21.33,1 33.921 17.60 I 20.07 16.30I

12.41 9.09—1 1 I I 1 1

2&29 years ... ... ... .. . .. ... .30-39 years .. .. . ... ... ... .. ..4049 years .. .. .. . ... .. .. ....60-54 years .. . ... .... .. .. .. ..55-64 years ... .... .. .. . ......

. . . 9.87

. . . 6.64

. . . 16.90

. . . 16.47

. . . 6.29

12.99 16.81 19.926.86 16.69 28.36 - 2.50

18.66- 22.63 17.42 12.35 19.3520.49 22.95 24.35 20.43 21 .3B

8.26 11.37 20.76 4.08 12.25

12.57II 8.87115.53114.19

14.365.60

11.658.35

10.52

16.316.323.88

13.184.04

6.240.954.55

12.23

2.18

‘1.08

Standard error of mean

All ages2064 Veals ...... .. .. . . . 1.89 2.33 2.87 4.57 4.55 5.19

. . . 3.48 4.58 5.94 7.04 - -

. . . 2.58 2.58 6.29 10.74 1.77

. . . 4.88 5.39 6.54 5.04 7.14 11.18

. . . 3.68 4.58 5.13 5.45 7.71 8.07

. . . 2.22 2.92 4.02 7.33 2.35 7.08

3.56 2.43 1.23—

20-29 Veals ... .. .. .. .. .. .. .. .

30-39 Veals..................4049 years ... .. .. .. .... .. ...

50-54 Veals . ... ... ..... . . ... .6564 Veals . ... ... ... ... . ....

6.289.44

5.187.10

10.18

2.804.39

2.785.26

5.76

2,391.12

2.951.42

:2.200.35‘1.32

}2,740.77

16 *U.S. GOVERNMENT PRINTING OFFICE: 1977— 260437:26

VITAL AND HEALTH STATISTICS PUBLICATIONS SERIES

Formerly Public Health Service Publication No. 1000

Series 1. Programs and Collection Rocedutes.-Reports which describe the general programs of the National Center for Health Statistics and its offices and divisions, data collection methods used, definitions, and other material necessary for understanding the data.

Series 2. Data Evaluation and Methods Research.-Studies of new statistical methodology including experimental tests of new survey methods, studies of vital statistics collection methods, new analytical techniques, objective evaluations of reliability of collected data, contributions to statistical theory.

Series 3. Analytical Studies. -Reports presenting analytical or interpretive studies based on vital and health statistics, carrying the analysis further than the expository types of reports in the other series.

Series 4. Documents and Committee Reports.-Final reports of major committees concerned with vital and health statistics, and documents such as recommended model vital registration laws and revised birth and death certificates.

Series 10. Data from the Health Interview Survey.-Statistics on illness; accidental injuries; disability; use of hospital, medical, dental, and other services; and other health-related topics, based on data collected in a continuing national household interview survey.

Series Il. Data from the Health Examination Survey.-Data from direct examination, testing, and measurement of national samples of the civilian, noninstitutionalized population provide the basis for two types of reports: (1) estimates of the medically defined prevalence of specific diseases in the United States and the distributions of the population with respect to physical, physiological, and psychological charac- teristics; and (2) analysis of relationships among the various measurements without reference to an explicit finite universe of persons.

Series 12. Data from the Institutionalized Population Surveys.-Discontinued effective 1975. Future reports from these surveys will be in Series 13.

Series 13. Data on Health Resources Utilization.-Statistics on the utilization of health manpower and facilities providing long-term care, ambulatory care, hospital care, and family planning setices.

Series 14. Data on Health Resources: Manpower and Facilities.-Statistics on the numbers, geographic distrib- ution, and characteristics of health resources including physicians, dentists, nurses, other health occu- pations, hospitals, nursing homes, and outpatient facilities.

Series 20. Data on Mortality.-Various statistics on mortality other than as included in regular annual or monthly reports. Special analyses by cause of death, age, and other demographic variables; geographic and time series analyses; and statistics on characteristics of deaths not available from the vital records, based on sample surveys of those records.

Series 21. Data on Natality, Marriage, and Divorce.-Various statistics on natality, marriage, and divorce other than as included in regular annual or monthly reports. Special analyses by demographic variables; geographic and time series analyses; studies of fertility; and statistics on characteristics of births not available from the vital records, based on sample surveys of those records.

Series 22. Data from the National Mortality and Natality Surveys.-Discontinued effective 1975. Future reports from these sample surveys based on vital records will be included in Series 20 and 21, respectively.

Series 23. Data from the National Survey of Family Growth.-Statistics on fertility, family formation and disso- lution, family planning, and related maternal and infant health topics derived from a biennial survey of a nationwide probability sample of ever-married women 1544 years of age.

For a list of titles of reports published in these series, write to: Scientific and Technical Information Branch National Center for Health Statistics Public Health Service, HRA Hyattsville, Md. 20782