YLIBRARY ......,, 11177717711111111111 THESIS ‘. -~( nun...” R 1“ Hill : a» . V ‘ 1 a... 7 105710713 “3"“ - c3..-“ : I ““6?- 4“;ng 1 '5““" :9- '9‘: 'whfl 2&3 I"? ? *gr'». This is to certify that the thesis entitled The Effects of Experimenter versus Subject Fit and Subject Training on Hearing Protector Attenuation presented by Kimberly A. Payne has been accepted towards fulfillment of the requirements for ._M-_A-__.degree in Audiology — & Speech Sciences Major professor Michael R. Chia‘i Date 6%0 /K§ 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution Msu‘ LIBRARIES .— RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. -“n at.» ‘15. THE EFFECTS OF EXPERIMENTER VERSUS SUBJECT FIT AND SUBJECT TRAINING ON HEARING PROTECTOR ATTENUATION BY Kimberly A. Payne A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Audiology and Speech Sciences 1983 COpyright by Kimberly Ann Payne 1983 ’- /%2-.3//5 ABSTRACT THE EFFECTS OF EXPERIMENTER VERSUS SUBJECT FIT AND SUBJECT TRAINING ON HEARING PROTECTOR ATTENUATION BY Kimberly A. Payne In 1979, the Environmental Protection Agency ruled that all domestic hearing protection devices (HPDs) must bear a single number index of effect designated Noise Reduction Rating (NRR). The literature indicates the NRR is of questionable reliability and validity. This study investigated the effects of subject training and HPD fitting method upon NRR and real- ear attenuation at threshold. Subjects were forty bilaterally normal-hearing listeners. Pre- and post-tests assessed the effects of a training program and attenuation measurements were made in general accord with ANSI 53.19-1974 at each of nine noise bands. NRRs were computed per EPA (1979). Results showed that: (l) the training program provided significant information gain: (2) experimenter fit produced slightly (but not significantly) greater attenuation than subject fit; (3) training had no impact on mean attenuation; and (4) measured NRRs were considerably smaller than manufacturer's data. To accomplish great things, we must not only act, but also dream, not only plan, but also believe. '-Anatole France ACKNOWLEDGEMENTS There are so many to acknowledge, and space for so few. A most sincere thank you to Michael R. Chial, Ph.D. for serving as my committee chair and for being a true friend. I am grateful for his guidance, his expertise, his opinions, and his praise. Above all, I am grateful for his committment to this project and for teaching me how to work until it hurt. Thanks to the members of my committee. To Linda Lou Smith, Ph.D. for her support and her friendship, and for being there for all of us. To Paul A. Cooke, Ph.D. for his guidance and willingness to serve as a committee member. A thank you to the very best teacher ever - Patricia E. Connelly, Ph.D. Her friendship and support shall never be forgotten. Thanks to Nick Hinkle for his engineering and industrial expertise in building the test capsule for this study. His time and willingness to help is greatly appreciated. To Rod, Mary, Gail, Claire, Judi, and Fred. You are all the greatest. The "incredible seven" must always endure. A very special thank you and a hug to my parents whose love and support has only grown in magnitude. Thank you for giving so that I could receive. I love you both. And finally, once again, to JFN, who may now know. TABLE OF CONTENTS BEES LIST OF TABLES ................... .... ......... ... Viii LIST OF FIGURES ............. . ..... . .............. xi LIST OF APPENDICES ..... . ......................... xii CHAPTER I--INTRODUCTION BACKGROUND..;... ..... ...; .......... ' ..... . ........ 1 REVIEW OF THE LITERATURE..... .............. . ..... 4 Federal Regulation of Noise ......... . .......... 4 EPA, 1979. ................... .... ..... ......... 5 OSHA, 1981000..........OQOOOOOOO......OIOOOOOOO 8 Interlaboratory Differences in Measured Atten- uation ................... ......O.............. 9 Laboratory versus Field Data................... ll Sources of Variation in Laboratory and Field Data... ....................................... 22 STATEMENT OF THE PROBLEM ......................... 28 CHAPTER II--METHODS INTRODUCTION ..................................... 30 SUBJECTS .............................. . .......... 30 STIMULUS MATERIALS... .......................... .. 32 Hearing Protection Devices ..................... 32 Training Program ............................... 32 Signal Generation .............................. 33 iv EXPERIMENTAL APPARATUS............... ...... . ..... Environment............. .............. . ........ Signal Presentation System ..................... Logic System................................... Response Acquisition System..... ..... .......... System Calibration.... ..... ........... ..... .... Temporal Parameters....... ......... ......... Attenuator Linearity ...... ..... ....... ...... Harmonic Distortion ......................... Signal Level.. .............................. Reverberation Time........ ....... . .......... Spatial Uniformity of Acoustic Field ........ Ambient Noise. .............................. EXPERIMENTAL PROCEDURES .......................... Subject Screening. ............................. Running Calibration..... ....................... Data Collection Procedures ..... . .............. . DATA REDUCTION ................................... Form and Volume of Subject Data....... ......... Pre- and Post-Test Scores... ....... .... ........ Attenuation Data ......... . ..................... Noise Reduction Ratings (NRRs) ................. CHAPTER III--RESULTS INTRODUCTION ..................................... DATA REDUCTION AND ANALYSIS ...................... Statistical Procedures ......................... Information Gain. .............................. 37 41 43 43 43 44 44 44 48 48 so 51 51 51 52 54 54 54 54 55 56 57 57 58 Page Description of Outcomes..... ...... .......... 59 Analysis of Outcomes........................ 63 Real-Ear Attenuation Data..... ..... . ........... 63 Reliability of Data Reduction ............... 67 Description of Outcomes............ ......... 67 Reliability of Subject and Group Data....... 74 Relation Between Information Gain and Real- Ear Attenuation ............. .... ........... 90 Analysis of Outcomes........... ....... ...... 90 Summary... .................................. 103 Noise Reduction Ratings (NRRS).. ............... 104 CHAPTER IV--DISCUSSION INTRODUCTION ...... .. ............................. 107 INFOMTION GAIN O O O O O O O O O O O C O C O O O O O O C O C O O O O O O C O O O l 0 8 REAL-EAR ATTENUATION AND NOISE REDUCTION RATINGS (NRRS)................... .......... ..... ........ 109 RELATION OF OUTCOMES TO PRIOR RESEARCH ........... 114 FINDINGS ......................................... 124 IMPLICATIONS AND SUGGESTIONS FOR FURTHER RESEARCH............ ..... ............. .......... 126 CONCLUSIONS ...................................... 128 CHAPTER V--SUMMARY AND CONCLUSIONS INTRODUCTION ..................................... 130 METHODS .......................................... 131 RESULTS... ....................................... I33 CONCLUSIONS ...................................... 134 vi Page APPENDICESOOOOOOOOOOO000.00.000.00...00.00.00.000 137 REFERENCES ........................ . .............. 209 vii Table 1.1. 1.2. 1.3. 1.4. 2.1. 2.2. 2.3. 3.1. 3.2. LIST OF TABLES Sample calculation of the noise reduction rating (Michael and Bienvenue, 1980, p. 545)oooooooooooocoo-coo...ooooooooooooooooo Mean and standard deviation values of attenuation of earplugs, in decibels....... Summary information from l0 real-world studieSOOOOOOOO......OOOOOOOOOOOIOOOO...... Summary of labeled vs. real-world perfor- mance.................. ........... . ........ Values of noise band spectra analysis...... Acoustical attenuator linearity measurement resultSOOOOOOOOO0.0.00.00...00.00.00.000... Signal level measurement results........... Means, standard deviations and standard errors of the mean for pre-test, post-test, and information gain scores (percent correct) across groups..................... Correlation coefficients relating pre-test, post-test and information gain scores for four groups of subjects. Groups A and 8 received training; Groups C and D did not.. Results of ANOVA of pre-test scores........ Results of two-way ANOVA of information gain scores.... ..... ................ ....... Attenuation data across subjects (10 sub- jects per group)........................... Attenuation data across subjects and trials (10 subjects per group; 3 trials) .......... Attenuation data across subjects ........... viii Page 18 19 21 36 45 49 60 64 65 66 68 73 75 Table Page 3.8. Reliability of individual subject data -- Group A (experimenter fit and training).... 77 3.9. Reliability of individual subject data -- Group B (subject fit and training)......... 78 3.10. Reliability of individual subject data -- Group C (experimenter fit and no training). 79 3.11. Reliability of individual subject data -- Group D (subject fit and no training)...... 80 3.12. Reliability across trials and noise bands for Group A (experimenter fit and training).................................. 81 3.13. Reliability across trials and noise bands for Group B (subject fit and training)..... 82 3.14. Reliability across trials and noise bands for Group C (experimenter fit and no training)....... .......... ...... ....... .... 83 3.15. Reliability across trials and noise bands for Group D (subject fit and no training).. 84 3.16. Reliability across trials and noise bands for Groups A and C (experimenter fit)...... 86 3.17. Reliability across trials and noise bands for Groups B and D (subject fit)........... 87 3.18. Reliability across trials and noise bands for Groups A and B (training) ........... ... 88 3.19. Reliability across trials and noise bands for Groups C and D (no training)...... ..... 89 3.20. Relation between information gain and real- ear attenuation for all groups ............. 91 3.21. Results of three-way ANOVA of attenuation data ....................................... 92 3.22. Results of two-way ANOVA of attenuation data - 125 Hz .............................. 94 3.23. Results of two-way ANOVA of attenuation data - 250 Hz .............................. 95 3.24. Results of two-way ANOVA of attenuation data - 500 Hz .............................. 96 ix Table Page 3.25. Results of two-way ANOVA of attenuation data-1000 H2000... 00000 00.000000000000000 97 3.26. Results of two-way ANOVA of attenuation data-2000112. ooooooo ooooooooooooooooooooo 98 3.27. Results of two-way ANOVA of attenuation data - 3150 Hz. ...... ...................... 99 3.28. Results of two-way ANOVA of attenuation data - 4000 Hz .......... . ............. ..... 100 3.29. Results of two-way ANOVA of attenuation data-6300 Hz 00000 000000000000000000000000 101 3.30. Results of two way ANOVA of attenuation data - 8000 Hz ........ ... ..... ........ ..... 102 3.31. NRR results ..... . ..................... ..... 105 4.1. Attenuation data from this study and from MiChael's data0000 00000 0000 00000 00000000000 111 4.2. Attenuation results for the EAR hearing protection device ...... .................... 113 4.3. Converted ambient noise levels ............. 120 4.4 Ad hoc analysis of Trial 3 attenuation data ...... . ................................ 123 LIST OF FIGURES Figure 2.1. Schematic view of the test capsule......... 2.2. Block diagram of stimulus presentation system0000000000000000000000000000000000000 2.3. Block diagram of logic system.............. 2.4. Block diagram of instrumentation for signal level measurements0000000000000000000000000 3.1. Pre-test and post-test means for all groups (Group A=experimenter fit and training; Group B=subject fit and training; Group C= experimenter fit and no training; Group D: subject fit and no training)............... 3.2. Information gain means for all groups (Group A=experimenter fit and training; Group B=subject fit and training; Group C= experimenter fit and no training; Group D= subject fit and no training)............... 3.3. Means and standard deviations of attenuation data across ten subjects -- Group A (experimenter fit and training) ..... . ...... 3.4. Means and standard deviations of attenuation data across ten subjects -- Group B (subject fit and training) ........ ......... 3.5. Means and standard deviations of attenuation data across ten subjects -- Group C (experimenter fit and no training)......... 3.6. Means and standard deviations of attenuation data across ten subjects -- Group D (subject fit and no training) .............. xi 61 62 69 70 71 72 Appendix til 0 0CD LIST OF APPENDICES Subject Screening Forms ...... . ...... .... Run Protocol.. ...... .. ..... ............. Training Tests.......................... Speaker Wiring Diagrams................. Instrumentation, Methods and Results of Reverberation Time Measurements......... Instrumentation, Methods and Results of Spatial Uniformity of Acoustic Field Measurement80000000000000000000000 000000 Computer Program for Computation of NRR. Pre—test, Post-test and Information Gain scores (raw data)....................... Descriptive Statistics for Pre-test, Post-test, and Information Gain Scores.. Raw Attenuation Data...... .............. Results of Reliability Check on Exper- imenter Determined Thresholds ........... Confidence Intervals of Mean Attenuation for all Subjects... ..... ................ xii 183 187 192 195 197 199 204 205 CHAPTER I INTRODUCTION BACKGROUND It is well known that excessive noise can damage the human auditory system. Such damage can occur in any of three ways (Melnick, 1978). A noise-induced temporary threshold shift (NITTS) is an observed change in hearing threshold level (HTL) which is reversible and recovers after a certain period of time following exposure. Recovery depends upon a variety of factors including the characteristics of the noise, the frequency of the measured NITTS and the amount of time between the termination of exposure and when threshold is measured. Noise-induced permanent threshold shift (NIPTS) occurs when excessive noise exposure is chronic over a period of years. The associated hearing loss does not reverse. The relation between NITTS and NIPTS varies greatly among individuals; consequently, there is no accurate way to predict who will be affected by intense noise or how much damage will occur. Acoustic trauma, a third type of auditory damage, is a loss of sensitivity following a single exposure to extremely intense sounds such as an explosion. The traumatic event produces destruction of hair cells in the organ of Corti and generally some permanent hearing loss results. Damaging noise is prevalent in industrial environments. Protection from the harmful effects associated with intense noise ideally should focus on prevention. When appropriate administrative and engineering controls cannot sufficiently decrease noise to acceptable levels, however, personal hearing protection devices should be employed. Hearing protection devices can be classified into four general categories: helmets, ear canal caps, earmuffs and insert earplugs. There are over 200 different brands of devices commercially available, and it is reported that insert earplugs are the most popular (Smith and Borton, 1981). Hearing protection devices function to block the external auditory canal, thereby decreasing the sound pressure level (SPL) reaching the inner ear. Several different methods for measuring hearing pro- tection devices have been proposed and utilized (Nixon, 1982). These include: (1) Real Ear Attenuation at Threshold; (2) Loudness Balance; (3) Temporary Threshold Shift; (4) Aural Reflex; (5) Subjective Comparison; (6) Miniature Microphone: (7) Masked Threshold; (8) Hearing Loss for Speech; and (9) Cadaver Measurements. Each of these methods presents advantages and disadvantages. Prior to 1979, the effectiveness of hearing protection devices was evaluated behaviorally by an absolute threshold shift procedure (Berger, 1980a) whereby unoccluded thresholds were subtracted from occluded thresholds. These attenuation values were utilized to numerically describe the amount of protection from noise the wearer could expect to receive. In September, 1979, the Environmental Protection Agency (EPA, 1979) specified that all domestic hearing protection devices must bear a label containing a single-number estimate of effectiveness. This estimate is designated "Noise Reduction Rating" (NRR). The purpose of the NRR is to provide a simple basis for predicting protection in noisy environments and to allow comparison of the effectiveness of different protective devices. The NRR indicates the noise attenuation capability of a hearing protection device, weighted by an assumed noise spectrum and the statistical variations in band attenuation data obtained from a group of trained listeners (Juneau, 1982). For example, if a hearing protection device has an NRR of 25 dB, the average worker wearing the device should be able to expect that the SPL reaching the hearing mechanism is to be reduced by 25 dBA. Enforcement of the EPA (1979) regulation specifically involves: (1) label verification testing and reporting for each protection device in a manufacturer's product line; (2) the monitoring of products by random selection for testing; (3) audit testing of products by manufacturer's to insure that products comply with labeled values; and (4) remedial orders if noncompliance occurs. REVIEW OF THE LITERATURE Federal Regulation of Noise Federal laws governing occupational noise exposure and control began with the regulations issued under the authority of the Walsh-Healey Public Contracts Act Amendment of 1969. The Walsh-Healey Act specified that industrial noise must be controlled "to minimize fatigue and industrial accidents," provided a table of permissable noise exposure levels (90 dBA limit for 8-hours exposure), and specified the use of engineering and administrative controls to reduce hazardous noise levels. The Act was applicable to all industries with governmental contracts exceeding $10,000. The Occupational Safety and Health Act was passed in 1970.. This Act established the Occupational Safety and Health Administration (OSHA) and extended federal authority for industrial noise control to all industries involved in interstate commerce. The Act set standards for appropriate hearing conservation programs for employees when noise levels exceeded the permissable levels. An exposure limit was established at 90 dBA (Slow) of steady state noise for 8— hours duration. The Act also created the National Institute for Occupational Safety and Health (NIOSH) which was authorized to "develop and establish recommended safety and health standards." In 1972, NIOSH published a criteria document which recommended a permissable noise exposure time-weighted average limit of 85 dBA for 8-hours of exposure. EPA, 1979 As previously stated, the Environmental Protection Agency (EPA, 1979) specified that labels of all domestic hearing protection devices must bear an NRR. The behavioral test methods underlying the EPA'S (1979) NRR are described in American National Standards Institute (ANSI) 83.19-1974 "Measurement of Real-Ear Protection of Hearing Protectors and Physical Attenuation of Earmuffs." ANSI 83.19-1974 specifies that for each device which is evaluated, ten trained, normal-hearing subjects shall be tested three times each with an unoccluded ear and with the hearing protection device fit by the experimenter or the subject. Although ANSI allows either experimenter or subject fit, EPA (1979) requires experimenter fit for determination of a devices NRR. Measurements are to be obtained in a laboratory setting in a diffuse (reverberant) sound field and test signals are to be third-octave bands of noise with center frequencies ranging from 125 Hz to 8000 Hz. Martin (1982) explains that 1/3-octave bands of noise ...represent a reasonable compromise between the need for frequency-specific attenuation data and the practical noise environment. The diffuse soundfield...ensures that sound is incident upon the protectors under test from all directions, as is usually the case in industry (p. 277). Attenuation is determined by subtracting unoccluded audibility thresholds from occluded audibility thresholds. The NRR is then calculated from mean and standard deviation information generated from these values. Table 1.1 Shows a sample calculation of NRR (Michael and Bienvenue, 1980). Line 1 of the sample calculation is the assumed pink noise (equal SPL per octave) exposure level for octave band center frequencies 125-8000 Hz. Line 2 gives the adjustments for the "C"-weighted levels by which the pink noise octave band levels must be modified to yield a corresponding wide band level. The C-weighting network is a band-pass filter roughly equal to the inverse of the mean equal-loudness contour for normal ears at 70 phon. Line 3 is the algebraic difference between lines 1 and 2. These differences are summed logarithmically across fre- quency to yield a C-weighted, wide-band level. Line 4 gives the "A"-weighting adjustment values. The A-weighting network is a band-pass filter approximating the increase of the mean equal-loudness contour for normal ears at 40 phon. Line 5 gives the unprotected ear "A"-weighted levels which is calculated by subtracting the "A"-weighting adjust- ments from the assumed pink noise levels. Lines 6 and 7 give the mean and standard deviation values for a particular hearing protector as generated by the ANSI 83.19-1974 methodology. Line 8 gives the protected ear weighted levels. These values are obtained by subtracting the mean attenuation values from the unprotected ear "A"-weighted levels, then adding the doubled standard deviations. Thus, standard mEsm mo mQHEumuoo ou menufiummoH um: mmsHm> um coom paw ooow wo ommuw>¢ o monam> um ooov one ooom mo omoum> m mafia N 1 mmsHm> m mafia N n mmz .m H.om $.55 o.mm ~.mn o.mw m.mh v.om m.mn mo ca 5 mafia + o mafia 1 m mafia meo>oH pounmflm3 How oouoououm .m v v ca oa m w v m moaflu mo ca mcofluofl>oo oumocmum .5 mm N4 4m mm mm AH an socmsomom pm me EH coHumscouuo com: .o m.mm HQH ~.Hoa ooa m.om v.H¢ m.mm mo :a v used 1 a mafia undo>oa ouunmflm3 1:4: new omuomuoumca .m H.H1 o.H+ ~.H+ o m.mu m.m1 H.oau me am mucMEumsflea C H 3|: : . mo m.hoa an: .m 4 v c.5m ~.mm m.mm ooa cod ooa m.mm mo :H N mafia 1 H mafia umHm>uH ouunmwmz I:U: HMO QGUUUUOHQCD .m o.m1 m.o1 ~.o1 o o o ~.o mo ca mucmaamsfloa mCflugmHm3l..U: .N ooa cod ooa ooa ooa ooa ooa mo ca mao>oa omfioz Roam omasmmd .H nooom mooov ooom coca oom omm mma um cw Noncommum Houcmu comm m>ouoo ..mvm .m .omma .os:u>:owm poo Homnowzv mcflumn :ofiuosoun omfioc onu mo cofiumasoamo onEom .H.H canoe deviations of measured group attenuation "derate" mean attenuation values, resulting in a nominally more conservative index of effect. The resultant protected ear weighted levels are then summated logarithmically. The NRR is cal- culated as the decibel summation of the protected ear weighted levels subtracted from the decibel summation of the unprotected "C"-weighted levels. Three decibels (for spectral uncertainty) is subtracted from this result, and the NRR for a given hearing protection device is obtained. OSHA, 1981 In January, 1981, the Occupational Safety and Health Administration (OSHA) issued for public comment a hearing conservation amendment to its original occupational noise standard. The purpose of this amendment was "to prevent occupationally related cases of hearing impairment" (OSHA, 1981a, p. 4105). The regulation mandated a hearing con- servation program for all employees exposed to a time-weighted average noise level of 85 dBA in an 8-hour duration. "Hearing conservation included noise exposure monitoring, audiometric testing, the use of hearing protective devices where necessary, and employee education" (OSHA, 1981a, p. 4079). OSHA (1981a) mandated the use of hearing protection devices for all employees exposed to 90 dBA or more of steady state noise during 8-hours of exposure where appropriate administrative and engineering controls are not able to reduce noise to acceptable levels. Further, hearing protec- tion devices must be provided to all employees exposed to time-weighted average noise levels of 85 - 90 dBA, but only those employees exhibiting a significant threshold shift are reguired to wear them. OSHA (1981a) specified that training programs be implemented for all employees exposed to 85 dBA or more of noise. Training programs are to be repeated annually and must address the following issues: (1) The contents of the noise standard including the hearing conservation program; (2) the effects of noise on hearing; (3) specific machinery at the jobsite that could produce hazardous noise exposures; (4) the role of engineering and administrative controls in the reduction of noise exposure; (5) the contents of any noise control compliance plan in effect; (6) the purpose of hearing protectors, the advantages, disadvantages, and attenuation of various types, and instructions on selection, fitting, use and care; and (7)‘ the purpose of audiometric testing, and an explanation of test procedures (p. 4164). Finally, OSHA (1981a) mandated the use of the EPA (1979) regulation. Although several provisions of the or- iginal amendment (OSHA, 1981a) were not implemented, the bulk of the hearing conservation amendment was put into effect in August, 1981 (OSHA, 1981b). A final rule was issued by OSHA in March, 1983 (OSHA, 1983). Interlaboratory Differences in Measured Attenuation Berger, Kerivan and Mintz (1982) reported results found in an EPA funded inter—laboratory comparison. Eight 10 U.S. test laboratories were required to obtain attenuation data (re: EPA, 1979 and ANSI 53.19-1974) on four types of hearing protection devices. The NRR values were also calculated. The results indicated differences in mean and standard deviations among laboratories, probably attributable to differences in hearing protector fitting, subject selec- tion, subject training and data reduction techniques. Although not stated in this report, several other explanations for the inter-laboratory differences are poss- ible. These include: (1) departures from specified acoustical properties (Spatial, spectral, and temporal) of signal sources and test environments; (2) variations within acoustical specifications of signal sources and test environments; (3) differences in the accuracy and precision of signal sources and test environments; (4) variations in acoustical factors not currently covered by the EPA and ANSI documents; (5) variations in subjects' threshold criteria and training; and/or (6) variations in psychophysical measurement methods affecting the precision of attenuation measurements (e.g., step size). Berger, Kerivan and Mintz concluded that: The results...do not cast aspersions on the NRR per se, but rather on the data from which the NRR is computed. The results do Show, however, that application of any one set of laboratory data to a real world environment for the purposes of predicting an estimated protected noise exposure is a tenuous proposition at best... optimal attenuation values are of little use to designers, purchasers, or users who need some indication of the protection that hearing pro- tection devices can normally be expected to provide (p. 18). 11 Forshaw (1982) states several explanations for why variations occur from laboratory to laboratory. First, the standards (ANSI 53.19-1974 and EPA, 1979) are not sufficiently explicit on the selection of subjects and on the fitting procedures of the hearing protection devices for testing. Second, it is stated that there are sources of variance inherent in the real-ear threshold method of measurement. Third, the attenuation of protectors depends on factors such as size and shape of the external auditory canal and of the contours of the circumaural region of the head. Fourth, inter-subject differences may be significant from laboratory to laboratory when only ten subjects are used. Finally, Shifts in a subject's attention span and his/her signal detection criterion may be a source of error because of the difficult listening task employed. Laboratory versus Field Data The validity of the NRR is open to question. Although the NRR is intended to describe the overall attenuation characteristics of a protection device, measurements are obtained under optimal conditions. It is obvious that industrial field environments differ significantly from laboratory environments. Therefore, laboratory measurements of hearing protection devices may not accurately reflect effectiveness of the device in the industrial field setting. Several studies have supported this contention. Padilla (1976) compared the attenuation characteristics of earplugs in controlled laboratory environments and in 12 uncontrolled industrial environments. Subjects for this study were industrial workers who routinely performed noisy tasks and who were required to wear hearing protection devices. The subjects were divided into two groups: group A subjects were brought into the laboratory for testing; group B subjects were tested in the field. Results indicated that (1) some individuals were adequately protected but many were not; (2) overall mean attenuation for the field testing was only 12 dB at 500 Hz (pure tone signal simultaneously directed to both ears); (3) individually fitted earplugs were more effective than pre-sized earplugs; and (4) the degree of protection is influenced by the fitting technique. Padilla stated: ...when the subjects know that their earplugs are going to be tested and that the test is only going to take a few minutes, they are apparently further motivated and a better effort is made to obtain proper earplug placement. This may indicate that perhaps other types of motivation, enforce- ment, etc. need to be investigated (p. 35). Padilla concluded that laboratory data do not accurately represent field performance and that a Significant number of employees were unprotected from hazardous noise in in- dustrial settings. Regan (1977) obtained attenuation data for earplugs on 32 subjects employed by a steel stamping company. This study sought to determine whether manufacturers' atten- uation results accurately reflected the actual attenuation provided by hearing protection devices in industrial work environments. The study also compared attenuation values 13 between various types of protection devices. All subjects routinely used hearing protection devices. Four types of devices were used: (1) malleable, soft sponge inserts, (2) non-malleable rubber inserts, (3) custom-fitted earplugs, and (4) earmuffs. Data were collected by escorting subjects directly from their work stations to an audiometric test van near the plant. Escorts were provided to insure that no manipulation of the hearing protection device took place. Subjects were tested four times each during a two-week period utilizing the test method described in an earlier version of ANSI 53.19-1974 (ANSI 224-22—1957). Results indicated that the attenuation provided to these industrial workers was significantly less than that specified by the manufacturers. Results further indicated. that custom-fitted earplugs offered the least amount of attenuation and malleable sponge inserts offered the greatest attenuation. Regan concluded that these hearing protection devices provided an inefficient means of protecting employees from hazardous noise. Michael, §E_§l. (1976, as cited in Edwards, gE_§l., 1978 and in Michael and Bienvenue, 1980) developed a field test method and a special headphone system for measuring attenuation characteristics of hearing protection devices. The field test method was designed to document the variability of hearing protector performance in the industrial environ- ment; it was not designed to replace the ANSI 53.19-1974 method. The Study provided a means of correcting attenuation 14 measurements made in the field to values that would have been obtained in the laboratory. Edwards, et_gl. (1978) studied the attenuation char- acteristics of three types of earplugs on 168 workers from six industrial sites. Subjects were tested five times each over a period of five days. Subjects were randomly selected for testing from their work sites and escorted to an audiometric test van near the plant. This study utilized the field test method described by Michael, gE_§l. (1976) and subject fit of the hearing protection deviCes. It was found that on the average, industrial employees received only 33 - 54% of the maximum protection afforded by the devices. Results further indicated that when the attenuation corrections suggested by Michael, gt_gl. (1976) were employ- ed, field results more closely compared to laboratory results. These researchers recommended that: (1) additional testing of industrial employees using other types of hearing protection devices is needed; (2) additional research is needed to determine why hearing protection devices are not correctly attenuating noise; and (3) attenuation values provided by manufacturer's should be decreased by 60% to reflect realistic values for protection in industrial work settings. Alberti, gE_gl. (1979) evaluated the attenuation charac- teristics of three types of earplugs and a group of assorted earmuffs on 88 industrial workers. These subjects were required to fit the protector to themselves. Open and 15 occluded thresholds were obtained using l/3-octave bands of noise at center frequencies ranging from 125 - 6000 Hz. Testing was conducted in a free-field sound proof booth utilizing a psychophysical method of limits. Mean atten- uation for each group of hearing protectors was computed by subtracting unoccluded thresholds from occluded thresholds. Results indicated that attenuation increased with frequency for each type of device through 3000 Hz. Above 3000 Hz, some dropoff in attenuation was noted. At 125 and 250 Hz, the earplugs provided greater attenuation, however, above 250 Hz, the earmuffs provided significantly more atten- uation than the earplugs. Custom molded earplugs provided considerably less attenuation across all frequencies. Large standard deviations were also found. Attenuation values were found to be less than values given by the manufacturers of the protection devices. These results are consistent with those of similar studies indicating that attenuation of hearing protection devices in industry is considerably lower than manufacturer's- laboratory data would Suggest. The reasons cited for these differences relate to fit of the devices. Berger (1982) reports results of a similar study. Sixty-five randomly selected, untrained persons served as subjects. Subjects were screened otoscopically and administered a battery of four tests: These were (1) pure tone unoccluded thresholds under earphones; 16 (2) 1/3-octave band unoccluded thresholds in a diffuse soundfield; (3) l/3-octave band occluded (subject fit) thresholds in a diffuse soundfield; and (4) l/3-octave band occluded (experimenter fit) threshold measures in a diffuse soundfield. Test conditions (2), (3), and (4) were conducted in accordance with ANSI 83.19-1974 with the exception that subjects were tested only once. Testing was conducted in the order outlined above. An insert type protector was used. Results indicated significantly poorer attenuation below 2000 Hz of the subject fit group. Comparison was made of the untrained experimenter fit group and ten trained subjects with experimenter fit of the device. Attenuation results were similar. Berger also compared two previous field studies, manufacturers' data, and data from the subject fit group of this study. Subject fit data were very similar to the data from the two in-field studies. Manufacturers' label attenuation results were significantly better than these three groups. It can be concluded from these results that laboratory test methods utilizing subject fit of hearing protection devices may yield attenuation values which are in better agreement with attenuation values obtained in the field. Abel, Alberti and Riko (1982) studied the attenuation of six types of earplugs and four types of earmuffs on 347 industrial employees. Subjects were required to fit their I? own protectors without instruction. One-third octave bands of noise were employed and testing was conducted in a sound treated booth. The psychophysical method of limits was used for all threshold estimates. Results indicated a wide variation in attenuation values within and between hearing protector types. These atten- uation values were significantly less than the manufacturers specified values. The primary reason cited for these differences relates to fit of the protector. In the industrial field environment, hearing protection devices are often improperly worn, and therefore may not provide sufficient protection from harmful noise. Martin (1982) compared attenuation data from experimenter fit and subject fit conditions for a pre-molded plastic earplug. The number of subjects and exact test conditions were not specified; it was stated that a real-ear threshold method similar to the ANSI standard was employed. Table 1.2 shows the mean and standard deviation values for these test conditions. Martin found that the experimenter fit situation results in significantly higher mean attenuation values and significantly lower standard deviation values then the subject fit group. He concluded that subject fit produces atten- uation measurements which more closely approximate real world performance than does experimenter fit. Berger (1983) reviewed data from ten studies published since 1975. Table 1.3 summarizes his findings for 1551 18 Table 1.2. Mean and standard deviation values of attenuation of earplugs, in decibels. From Martin (1982). Test Frequency in Hertz 250 500 1k 2k 3.1k 4k 6.3k 8k Experimenter i 25.1 25.8 29.1 34.1 38.6 34.7 32.3 30.9 Fit SD 4.3 4.9 3.7 4.3 5.6 5.7 6.0 5.8 Subject 2 16.9 16.4 18.8 24.0 30.0 28.4 28.1 29.7 Fit SD 9.3 12.1 8.4 7.8 9.9 8.1 11.3 10.5 19 Table 1.3. Summary information from 10 real-world studies. From Berger (1983). . Testa No. of Device Type Subjects NRR84 Foam Plug l 58 19 (EAR and Decidamp) 3 24 9 3 55 9 l 56 5 l 56 12 l 31 9 Custom Molded l 7 7 l 6 4b l 230 8 3 48 3 l 56 8 1 44 4 Willson EPlOO l 22 0 l 28 -2 3 45 10b V-SlR '1 9 7b 1 183 -l l 84 1 3 20 2b MSA Accu-Fit l 13 2 Norton Com-Fit 3 18 7 Bilsom Fiberglass (down) 1 56 3 (down) 1 28 4 (POP) l 28 4 (soft) l 36 1 David Clark 3 17 15 Safety Supply # 258 3 15 12 Hellberg MK-IV 3 58 ll MSA MK IV 3 47 11 2 15 4 Welsh 4530 l 5 20b EarmuffsC 2 101 14 A0 1720 2 ll 7 Glendale 900 2 10 10 Bilsom UF-l l 31 13 Total Subjects = 1551 aTest types are: l. Real-ear attenuation at threshold with employees pulled from work stations; 2. Dosimeter mics inside and outside earmuff on employees in work place; 3. Real-ear attenuation at threshold with employees reporting to outside test clinic and using their HPDS as normally worn. bNRRS estimated from measured dBA noise reduction or from attenuation data at only 500 or 1000 Hz. CNo model was specified since many different models were used. 20 total subjects, 50 different industrial sites, and several types of hearing protection devices (Berger, 1983, p. 13). The designation NRR84 indicates NRR calculations based upon a single standard deviation correction factor, which suggests that 84% of a normal population would be expected to produce NRR values equal to or larger than the data Shown. Berger contends that this is a more valid estimate than the EPA procedure which employs a correction factor of two standard deviations to describe expectations for 98% of the normal population. As is evident in Table 1.3, the NRR84 values range from 0-20 dB indicating a very wide range of pro- tection in the industrial environment. Table 1.4 (Berger, 1983, p. 14) compares the field NRR84 to the manufacturer listed NRR'98 for the same 1551 subjects. These data clearly indicate that employees are receiving significantly less protection in the industrial field environment than manufacturers' (laboratory) data would suggest. Berger concluded that the NRR can be a practical and suitable estimate of noise reduction if some changes are made. He suggests that 10 dB be subtracted from labeled NRRS before being subtracted from C-weighted sound levels and that improved motivation, training and supervision take place in the field to insure proper use of hearing protection. The above research indicates that laboratory test results relate poorly to field test results. Berger (1980b) cited several observations relatedtxn the issue of laboratory 21 Table 1.4. Summary of labeled vs. real-world performance. From Berger (1983). No. of Dev1ce Subjects NRR98 NRR84 A Foam Plug 280 29 ll 18 Custom Molded 391 14 6 8 Willson EPlOO 95 15 3 12 V-SlR 296 23 2 21 MSA Accu-Fit 13 14 2 12 Norton Com-Fit 18 26 7 l9 Bilsom Fiberglass 148 22 3 l9 EARPLUGS (AVERAGE) 1241 20 5 15 David Clark 17 23 15 8 Safety Supply #258 15 22 12 10 Hellbert MK-IV 58 23 ll 12 MSA MK IV 62 23 8 15 Welsh 4530 5 25 20 5 Genrl. Muffs 101 22 14 8 A0 1720 ll 25 7 18 Gelndale 900 10 22 10 12 Bilsom UF-l 31 22 13 9 EARMUFFS (AVERAGE) 310 23 12 ll GRAND AVERAGE 1551 22 9 13 22 versus real-world performance of hearing protection devices: (1) Manufacturers' laboratory data overrate the real world performance of hearing protection devices. For a comfortable protector, this data can indicate the protection that conscientious, well-trained users will receive. For an uncom- fortable device it is virtually meaningless. (2) Manufacturers' laboratory data are useful for research and development and may yield an in- dication of the rank ordering of various hearing protection devices. (3) Laboratory experiments...which are designed to simulate real world performance can provide useful indications of the actual attenuation typically provided by hearing protection devices (p. 3). Sources of Variation in Laboratory and Field Data It has been established that laboratory performance of hearing protection devices differs significantly from real- world performance.‘ Several explanations for this difference are possible. All relate to the required NRR measurement technique. First, the measurement method (EPA, 1979; ANSI $3.19- 1974) specifies that all measurements be obtained in a diffuse (reverberant) sound field. Such a controlled acoustic environment produces optimally stable results. Industrial environments almost certainly do not conform to such a field; thus there may be interactions among acoustic fields, bodies, and hearing protection devices themselves. This is a source of variance which is expected to have an impact on the effectiveness of hearing protection devices in industrial environments. Related to this issue is the mobility of the subject. NRR test subjects must be seated quietly 23 and are immobile for relatively long periods of time; in- dustrial workers generally move about during a work day. In the field, workers will perspire and engage in various jaw motions (i.e., talking, chewing, etc.) which may in- fluence the fit of hearing protectors. Consequently, industrial workers may, without realizing it, displace or inappropriately adjust the hearing protection device, thereby decreasing its attenuation characteristics. Related to this issue are possible changes in the threshold criteria of workers tested in field settings. Industrial workers are not trained listeners and it is expected that trained laboratory subjects would exhibit more stable results during testing. Published reports of field tests tend not to provide the control condition data necessary to allow assessment of this effect. A second explanation relates to psychophysical method of measurement. ANSI 83.19-1974 does not specify the psycho- physical method to be used for testing hearing protection devices. Different psychophysical measurement methods produce different variances (Gescheider, 1976). For example, the method of adjustment generally produces smaller standard errors of measurement than other psychophysical methods. Because a related index of dispersion is used in the computation of the NRR, different psychophysical methods of measurement would be expected to produce different estimates of NRR. The laboratory that provides the majority of NRR 24 testing (Pennsylvania State University) employs a method of adjustment; published reports of field tests most often indicate use of a modified method of limits. Other psycho- physical method variables influencing outcomes (both in terms of precision and variance) include attenuation step size, the direction of signal level change (increasing or decreasing level), and the number of threshold crossings used to estimate threshold for a given frequency or noise band. Humes (1983) obtained attenuation values of ten hearing protection devices (5 earmuffs, 5 earplugs) utilizing four psychophysical measurement methods. Ten normal hearing young adults served as subjects. Each subject was tested for all ten protection devices and each of the four psycho- physical procedures. One-third octave bands of noise were presented in a diffuse soundfield in accordance with ANSI 83.19- 1974. The four psychophysical procedures utilized were: (1) the real-ear threshold procedure described in ANSI 83.19-1974 utilizing a transformed up- down method (Levitt, 1971); (2) the magnitude-estimation procedure for loudness (Stevens, 1975) which produced unprotected and protected loudness growth functions; (3) the reaction-time paradigm whereby protected and unprotected reaction-time intensity functions were compared; and (4) the masked bone-conduction threshold procedure whereby the mean of the unprotected masked thresholds were subtracted from the mean of the protected masked threshold. Procedures (2), (3), and (4) used 1/3 octave bands of noise having intensity levels ranging from 50 - 90 dB SPL to 25 evaluate the protection devices. This allowed for the assess- ment of attenuation linearity of hearing protection devices. Attenuation results were computed for each subject and for each device across each psychophysical procedure. Results for noise levels ranging from 50 - 90 dB SPL showed that attenuation was linear over this range. Linearity of hearing protection devices could not be assessed at levels above 90 dB SPL, and it was suggested that a method be devised to do this as industrial noise levels often exceed this level and protection at these levels need to be deter- mined. Humes stated that the preferred method of determining attenuation characteristics of hearing protection devices is the real-ear method. The magnitude estimation and reaction- time procedures tended to underestimate attenuation when compared to other procedures. The reaction-time paradigm was found to be the most difficult to implement and the most time consuming and therefore not recommended. It was recommended that the masked bone conduction technique be incorporated into the ANSI standard to assess attenuation at high intensity levels. Attenuation results from this study were compared to data from other studies. It was found that manufacturers' attenuation data were considerably higher than attenuation data obtained from this study, regardless of the psycho- physical procedure used. Humes concluded that "...manu- facturer's specifications of attenuation characteristics 26 are often optimal as opposed to typical characteristics" (p. 310). The EPA (1979) specifies that NRR measurements are to be made with the device fit by the experimenter according to the manufacturers specifications. It has been suggested that experimenter fit relates poorly to subject fit (Juneau, 1982; Smith and Borton, 1981; Berger, 1980c). This may be the single most important reason why attenuation data obtained in the laboratory do not compare to attenuation data obtained in industrial field environments. Smith, gE_gl. (1980) as cited in Smith and Borton, 1981) had 100 adult subjects choose the 'best fitting' earplugs for both ears. Results indicated that 68% of the subjects chose earplugs which were too small. Smith, gt_gl. concluded that indus- trial employees may have the same problem when fitting them- selves with ear protection. Smith and Borton (1981) state that little research has accurately addressed this issue. The laboratory test situation requires careful fit and adjustment of the hearing protection device and for this reason, optimal attenuation is obtained. This may not be the case in industrial work environments. Because of poor comfort, motivation or training, hearing protection devices may be inappropriately fit, thus decreasing their effective- ness (Berger, 1980c). Another cause of poor fit may be due to readjustment of the protection device throughout the duration of the work day. As stated, jaw movement can displace the device causing poor fit and protection. 27 A final issue with regard to experimenter versus subject fit is interaction with the type of protector. Because of the ability to visually observe placement, experimenters would be expected to accomplish more consistent positioning than subjects of earmuffs. Because of the ability to tactually observe placement, subjects (especially trained or experienced ones) would be expected to accomplish more consistent positioning of insert protectors. Closely associated with the issue of experimenter ver- sus subject fit is the issue of the effect of training of subjects when obtaining NRR data. Optimal laboratory data are obtained through the training of subjects. This is done to minimize response errors and is accomplished through the use of a response consistency criterion. Additionally, the NRR measurement technique utilizes trained and motivated subjects who are knowledgeable about the purpose and function of hearing protection devices. The issue is between the goals of stable (reliable) measurements and accurate (predictively valid) measurements. Laboratory methods, with their attempts to minimize variations from sources other than the hearing protector, approximate the former through procedural control. In so doing, they may accomplish reliability at the expense of validity. Tobias (1982) summarized many of the issues raised above: Manufacturers, who are paying laboratories to run these tests for them, clearly should prefer measurers who come out with the best results—- 28 the better the measured attenuation two standard deviations below the mean, the better their hearing protector's Noise Reduction Rating. Anything that improves the mean such as selecting only the best subjects and making the best possible fittings of the protectors, and anything that decreases the size of the standard deviation, such as homogenizing the population of test subjects, will lead to good scores. Manufacturers should be pleased. But the ultimate users probably should not. Those scores no longer serve the purpose that the Environmental Protection Agency must have intended. When variability is artificially decreased, one no longer has a reasonable basis for judging how well a given device will work on the person at the second percentile among wearers of that device. These labs are doing everything strictly according to the standard. They are not cheating. They are not changing the rules. Yet when they publish data, they Show ratings at the mean that are often very close to the ratings one or two standard deviations below the mean. By compressing the range of normal variations, they give attenuation values that say nearly nothing about real-world variability. As a result, one begins to believe that their data are no_more informative than if they had been collected on a single subject...Measurers of hearing protectors need to continue to evaluate and re-evaluate test procedures, to modify them, to interpret them, and to ignore them at the proper times. The influence of these procedures and their variations on economics, on safety, and on health are potentially enormous (p. 172-173). STATEMENT OF THE PROBLEM Several studies indicate substantial differences be- tween laboratory attenuation measurement technique results and the actual field tested attenuation results. Such differences, in turn will be reflected in NRR results. These studies indicate that hearing protection devices utilized in indus- trial work environments do not always afford the maximum protection indicated by the manufacturers of these devices. Berger (1980b) stated that existing test methods for hearing 29 protection device performance can be utilized with modifications related to selection, fitting, and training of subjects. This study sought to determine whether the variables of hearing protector fit and subject training affect atten- uation and NRR data obtained in the laboratory testing situation. The following questions were asked: (1) Is information gain significantly affected by the presence of a training program? (2) Do real-ear attenuation values differ significantly as a function of experimenter fitting versus subject fitting of hearing protection devices? (3) Do real-ear attenuation values differ significantly as a function of trained versus untrained listeners? (4) Do real-ear attenuation values differ significantly as a function of the interaction between fitting method and subject training? (5) Do real-ear attenuation values differ significantly as a function of test band? (6) What is the correlation between information gain and real-ear attenuation as a function of fitting method and subject training? (7) Do NRR estimates differ as a function of experimenter fitting versus subject fitting of hearing protection devices? (8) Do NRR estimates differ as a function of trained versus untrained listeners? CHAPTER II METHODS INTRODUCTION In September, 1979, the Environmental Protection Agency ruled that all domestic hearing protection devices must bear a label containing the devices Noise Reduction Rating (NRR). The NRR indicates the noise attenuation capability of a hearing protection device, weighted by an assumed noise spectrum and the statistical variations in band attenuation data obtained from a group of trained listeners. Several studies (Padilla, 1976; Regan, 1977; Edwards, gt_gl. 1978; Alberti, gt_§1. 1979; Abel, Alberti, and Riko, 1982; Berger, 1983) indicate that attenuation data (which were used to compute the NRR) generated in controlled laboratory settings do not accurately reflect effectiveness of protection devices in the industrial field setting. It has been suggested that the NRR underestimates the actual protection provided in industrial environments. This study sought to determine whether the variables of hearing protector fit and subject training affect attenuation and NRR data obtained in the laboratory testing Situation. SUBJECTS Subjects were forty adult listeners (20 females; 20 30 31 males) with normal hearing bilaterally. Subjects were initially naive about hearing protection devices (fit, types, usage). All subjects completed a case history form (Appendix A) and underwent otoscopic, audiometric and impedance testing. Subjects reported no history of otologic surgery, familial history of hearing loss, current upper respiratory infections, vertigo, tinnitus or hearing loss and were free of excess cerumen. Pure tone air and bone conduction hearing threshold levels (HTLs) were no poorer than 10 dB at test frequencies between 250 and 4000 Hz and no poorer than 15 dB at 8000 Hz. Subjects exhibited Type A tympanograms (Jerger, 1970), acoustic reflex thresholds within a normal sound pressure level (SPL) range of 70-100 dB HTL and 60-90 dB SL at 500 Hz, 1000 Hz and 2000 Hz, and absence of acoustic reflex decay at 500 Hz and 1000 Hz bilaterally (i.e., not more than 50% reduction in response amplitude during a lO-second period). Audiological and oto- logical normalcy was confirmed within three days of experi— mental testing and recorded on the screening form presented in Appendix A. The audiometer and impedance bridge met relevant requirements of ANSI 83.6-1969 "American National Standard Specifications for Audiometers." Subjects were randomly assigned to one of four groups. Group A consisted of ten subjects with hearing protection devices fit by the experimenter (re: ANSI 83.19-1974 "Method for the Measurement of Real-Ear Protection of Hearing Protectors and Physical Attenuation of Earmuffs"). 32 Group A subjects also participated in a training program (see below). Group B consisted of ten subjects with hearing protection devices fit by the subjects. These subjects also participated in the training program. Group C con- sisted of ten untrained listeners with hearing protectors fit by the experimenter. Group D consisted of ten untrained subjects with hearing protectors fit by the subjects. Groups C and D did not receive the training program. STIMULUS MATERIALS Hearing Protection Devices The hearing protection devices used for this study were the Bilsom "Propp—o-Plast" disposable insert-type plugs. These devices are composed of a cotton-like material covered by a polyethylene film wrapper. This earplug is given an NRR of 20 decibels. TrainingiProgram Twenty subjects (Groups A and B) participated in a training program. This consisted of a commercially available (Bilsom International) multi-media education program. A videotape program was presented emphasizing the following information and affective topics: (1) the hearing mechanism and how it operates; (2) the effects of noise on the hearing mechanism; (3) the consequences of noise-induced hearing loss; and (4) hearing protection devices - how and why they work. 33 Subjects were given the opportunity to ask questions regarding the information presented and given practice in fitting the hearing protection devices. Subjects partici- pated in this training program on the day of data collection. All forty subjects were given a 40 item, multiple choice (5 item) pre-test and a 20 item, multiple choice post-test. The pre-test consisted of 20 content items designed to assess the informational and affective effects of the training program. It also consisted of 20 distractor items (anatomy, physiology and pathology of the eye). The post-test con- sisted of 20 content items. Alternate forms of the test were administered at the time of audiometric screening and again directly following data acquisition. Copies of the pre- and post-tests and the correct answers may be found in Appendix C. Signal Generation The signals used in this study were narrow bands of noise with the following center frequencies: 125, 250, 500, 1000, 2000, 3150, 4000, 6300, and 8000 Hz. These noise bands were numbered from 1 - 9. The original signal source (Bruel & Kjaer Model 1024) was operated in a sine-band mode, then band-pass filtered (Krohn-Hite Model 3550) at a rejection rate of 24 dB per octave. Band 1 was generated using a 30 Hz wide band; band 2 was generated using a 100 Hz wide band and bands 3 - 9 were generated using a 300 Hz wide band. Upper and lower cut-off frequencies were calculated from equations given 34 in ANSI 81.11-1966 "Specification for Octave, Half-Octave, and Third-Octave Band Filter Sets." The equations are as follows: f 0.8909 x f L C 1.1225 x fc fH where fL = low cut-off frequency (Hz) fH fC The output of the filter was routed through a selector high cut-off frequency (Hz) center frequency of band (Hz) switch to channel 2 of a four-track, reel-to-reel tape recorder (Teac Model A—2340 SX) operated at 7.5 inches per second and using a 1.5 mil polyester tape (Ampex Model 406). The other input to the selector switch was a 1000 Hz sinu- soidal level calibration tone produced by a function gener- ator (Wavetek Model 185). Channel 4 of the tape recorder was driven by a microphone (Audio-Technica Model ATM41). Level calibration tones were recorded for 60-seconds on channels 2 and 4. For each of the nine noise bands, a lO-second level calibration tone was recorded on channel 2, followed by 45-seconds of stimuli. Appropriate voice labels announcing each band were placed on recorder channel 4 just prior to the beginning of the band on channel 2. Similarly, 60-second bands of white noise were generated with a sine-random generator (Bruel & Kjaer Model 1024) and recorded on tape channel 2. These bands were recorded without the band-pass filter. Stimuli were then replayed and VU level differences between level calibration tones and noise bands were noted 35 for each band. The output of recorder channel 2 was routed to the input of channel 1. Stimuli were re-recorded with input gain settings selected to minimize VU level differences between level calibration tones and noise bands. Final level differences between noise bands and calibration tones were less than 1 VU. Thus the final stimulus tape consisted of (l) a 60- second, 1000 Hz level calibration tone on channels 1 and 4; (2) a 60-second band of white noise on channel 1; (3-12) a lO-second calibration tone and a 45-second noise band on channel 1, preceeded by a voice label on channel 4; and (13) an additional 60-second band of white noise on channel 1. Leader tape separated successive stimuli. The spectra of the noise bands were verified as follows. The output of the tape recorder was routed to a narrOWbband analyzer (Bruel & Kjaer Model 2107), then to a graphic level recorder (Bruel & Kjaer Model 2305). Each stimulus band was analyzed in terms of center frequency, high and low cut-off frequency and amplitude at 1 octave above and below the center frequency. Results are given in Table 2.1. Measured bandwidths were somewhat narrower than those Specified by ANSI 81.11-1966 for noise bands centered at 3150, 4000, 6300, and 8000 Hz. EXPERIMENTAL APPARATUS Environment The test chamber used in this study was an IAC reverberation 36 Table 2.1. Values of noise band spectra analysis. Attenuation”Attenuation l octave l octave Band fL* fc* fH* BW* above fC below fC l 115 125 140 30 58.5 dB 59.0 dB 2 230 250 280 100 53.0 dB 56.0 dB 3 445 500 560 300 55.0 dB 45.0 dB 4 900 1000 1100 300 45.0 dB 58.0 dB 5 1800 2000 2200 300 65.0 dB 45.0 dB 6 2800 3150 3500 300 52.0 dB 57.5 dB 7 3600 4000 4400 300 60.0 dB 60.0 dB 8 5600 6300 7000 300 59.0 dB 55.0 dB 9 7000 8000 9000 300 55.0 dB 58.5 dB *in Hertz 37 chamber equiped with a smaller test capsule customized to accomodate the acoustical requirements of ANSI $3.19- 1974. The test capsule was a hexagonal shaped room with a steel floor and masonite-covered walls and ceiling. Figure 2.1 presents a diagramatic view of the test capsule and its contents. Signal Presentation System Figure 2.2 presents a block diagram of the stimulus presentation system used in this experiment. The noise bands and voice labels were reproduced by a tape recorder (Teac Model A-2340 SX) on channels 1 and 4, respectively. The voice label channel was routed to an amplifier-speaker combination (Ampex Model AA620) to allow monitoring of stimuli by the experimenter. Test signals were reproduced, split, and then routed to an electronic switch (Coulbourn Model 584-04), and to a contour-following integrator (Coulbourn Model 576-01) and a bipolar comparator (Coulbourn 821-06). The integrator—comparator subsystem controlled the logic system discussed below. The output of the electronic switch was routed to an amplifier (Coulbourn Model 882-24). The electronic switch was activated by a pair of timers (controlled by the logic system) which gated the signal on for 500 msec and off for 500 msec with exponential rise and fall times of 50 msec. The signal was then routed to the external signal input of a Bekesy audiometer (Grason-Stadler Model E800). Internally, the Bekesy audiometer pre-amplified the signal, then passed it through a recording attenuator 38 61 cm - Subject Reflecton Status Lamps H H O\ o 3 l / 1- — d. _ - ~ / l m p—c ' am I V nuoxom \ [XX\\\\ \\\\\\\ \\\\\\\\\\\L\\ \X] .N.m MXZmem u m .. M Q L m m ..ouofi “.15 .l \ x m \ A & N .3330“. ll “um i X\\ omcnucas «A \ \- LU x\L “H” l , :0u«3womumoanom 1\\ \\\ r— u n m \ \ u x f .onEo uomwmam acncvm mam use: nocuoomz once doom -o»-~oo¢ nuuuzm #«cOuuou auunam _ Damon _ 40 controlled by a subject-operated switch, a step attenuator controlled by the experimenter, and an output amplifier. The output of the audiometer was split, then routed to the right and left channels of an equalizer (Radio Shack Model 31-2000). The right channel of the equalizer was low-pass filtered at 1.0 kHz and routed to the right input channel of a power amplifier (McIntosh Model MC 2205). The left equalizer channel was high-pass filtered at 1.0 kHz and routed to the left input channel of the power amplifier. From this point, the left channel was routed to Speaker array # l and the right channel to speaker array # 2. Secondary outputs from the two channels of the power amplifier allowed measurement of the voltage driving each loudspeaker array (Bruel & Kjaer Model 2409). Speaker array # 1 consisted of twelve, ll-cm midrange speakers (Radio Shack Model 40-12828, 50 watts, 8 ohms). Speaker array # 2 consisted of three, 38-cm woofers (Radio Shack Model 40-1315A, 100 watts, 8 ohms). These speaker arrays were organized in three panels, 91-cm on a side. Each panel consisted of one woofer and four midrange speakers. One panel was suspended above the subject at an elevation of 198 cm above the floor. The other two panels were mounted on legs which elevated the lower edge of the panel 57 cm from the floor. These panels were oriented at approximately 450 azimuth relative to the subject. The wiring diagram of the speaker arrays is found in Appendix D. 41 Logic System' Figure 2.3 illustrates the logic system used in this study. To semi-automate signal control, communication with the subject and response acquisition, one output from record- er channel 1 activated a contour-following integrator subsystem (Coulbourn Model 876—01). The integrator summated test signal energy over 20 msec and produced a dc output proportional to the summated input. This signal was routed to a bipolar comparator (Coulbourn Model 521-06) which functioned as a one-bit analog-to-digital converter. The complimentary output of the comparator was true if the input to the device was below a threshold voltage selected to index the presence of a test signal on the tape. .‘When the signal was absent or below threshold, the signal "off" timers forced the electronic switch open (off). Other components of the logic system made it possible to hold the test signal off when the subject was out of position or when the system was turned off by means of an experimenter control switch. Subject head position was sensed by a pair of infrared photo cell alarm devices (Radio Shack Model 49-201) positioned inside the test capsule. One alarm sensed subject head position in an anterior-posterior plane, the other in a sagittal plane. This subsystem allowed no more than 18 cm (:’9 cm) of head movement from side-to-side and no more than 20 cm (i 10 cm) of movement from front-to-back. The logic system also controlled a set of lamps located 42 .Emumxm owmoH mo EmHmMflozxoon .m.~ musmsm U~m no #30 oxuhmuk #02 nos 50 P20 r0<fl¢ :t..)m , song; #5 .... ... 7 coat 6 5“. R— A s so: , ...:- aoZ 2.9: 3: 5316 n a .x \ ‘ ‘ go az< - .4_ q u. _ Jrrur no.5 (a “was nan-nausea nounuoa.:_ mas. «o A. uasoazo . . Ir ‘1] Luu_)h atnhmuh IO D—m and? A casts cannon uuonaam o as; vaaoun noucclsuoauu a. ._ ..~¢¢< . o. «no Baua" O‘cdhuUOaB .nououoc 0519 too. .cc9.m . ‘Zuurf... l £01nnm o snazqr L 43 inside the test capsule. These lamps were used to tell the subject to 'get ready;' 'wait;' or 'track' a test signal. A fourth lamp alerted the subject if he/she was 'out of position.‘ A similar set of lamps located outside the test chamber informed the experimenter of the status of the system. Response Acquisition System The response acquisition system consisted of the status lamps noted above and a Bekesy recording audiometer con- trolled by a subject response switch. When the switch was depressed, the recording attenuator reduced the signal level at a rate of 5 dB per second; when the switch was released, the signal level increased at the same rate. For all noise band threshold measurements, the audio- meter was configured with the calibration cam set in the SPL channel, then disengaged from the plotter drive at the 1000 Hz position of the cam. System Calibration Temporal Parameters. The temporal parameters of the test signals (on-time, inter-stimulus interval, rise-fall time) were verified through measurements with a storage oscilloscope (Tektronix Model T912) of the electrical signal (presented to the Bekesy audiometer. Measured signal duration was 500 msec with a rise-fall time of 50 msec. Inter- stimulus interval was 500 msec. It was assumed that other components of the signal presentation system had negligible 44 effects upon the temporal features of the signal. Attenuator Linearity. Attenuator linearity was assessed acoustically. A free-field microphone (Bruel & Kjaer Model 4145) was placed on a tripod and placed inside the test capsule at a position comparable to the center of the head of a subject. The micrOphone signal was pre- amplified (Bruel & Kjaer Model 2619 and 2804), then routed to a measurement amplifier (Bruel & Kjaer Model 2607). Microphone sensitivity was calibrated with a level calibrator (GenRad Model 1986). Results (see Table 2.2) were within the specifications of ANSI 83.6-1969. Harmonic Distortion. ANSI 83.19-1974 specifies that the entire system produce less than 5% total harmonic distortion measured at the position of the subjects head. Harmonic distortion could not be measured because of laboratory equipment constraints. Signal Level. Each of the test stimuli was reproduced in a steady-state mode (the electronic switch was held on) for signal level calibration. The recording attenuator of the Bekesy audiometer was set at the lOO-dB position. The +20-dB pad was engaged at various frequencies. Figure 2.4 presents the block diagram of the instrumentation used for these measurements. A pressure microphone (Bruel & Kjaer Model 4144) was placed in the test capsule at the center head position. The microphone was routed to a pre-amplifier 45 Table 2.2. Acoustical attenuator linearity measurement results. Signal Attenuator Position Measured Value White Noise 100 dB 88.0 dB 90 dB 78.0 dB 80 dB 68.0 dB 70 dB 58.0 dB 60 dB 49.5 dB 125 Hz 100 dB 90.0 dB 90 dB 80.0 dB 80 dB 70.0 dB 70 dB 61.0 dB 60 dB 51.0 dB 250 Hz 100 dB 100.0 dB 90 dB 91.0 dB 80 dB 81.0 dB 70 dB 71.0 dB 60 dB A 61.0 dB 50 dB 51.0 dB 500 Hz 100 dB 96.0 dB 90 dB 86.5 dB 80 dB 76.0 dB 70 dB 66.0 dB 60 dB 56.5 dB 1000 Hz 100 dB 85.5 dB 90 dB 75.5 dB 80 dB 66.5 dB 70 dB 56.0 dB 60 dB 48.0 dB 2000 Hz 100 dB 87.5 dB 90 dB 78.0 dB 80 dB 68.0 dB 70 dB 58.0 dB 60 dB 49.0 dB 3150 Hz 100 dB 83.0 dB 90 dB 73.0 dB 80 dB 63.0 dB 70 dB 54.0 dB 46 Table 2.2 Continued. Signal Attenuator Position Measured Value 4000 Hz 100 dB 83.0 dB 90 dB 73.0 dB 80 dB 64.0 dB 70 dB 54.0 dB 6300 Hz 100 dB 74.0 dB 90 dB 65.0 dB 80 dB 55.5 dB 8000 Hz 100 dB 75.0 dB 90 dB 65.0 dB 80 dB 56.0 dB 70 dB 48.0 dB 47 .mUCmEGHDWMQE Hm>mH Hmcmflm HON sewumucmesuumcfl mo Emwmofio xoon flmHMHHdE< acme Imusmmmz wammsm umzom Caz .QEMImMm .v.m musmflm la 48 (Bruel & Kjaer Model 2619) and powered by a microphone power supply (Bruel & Kjaer Model 2804). The output was routed to a measurement amplifier (Bruel & Kjaer Model 2607). The microphone was calibrated with a level calibrator (Gen- Rad Model 1986). Additionally, speaker voltages were measured with an electronic voltmeter (Bruel & Kjaer Model 2409) for the right and left channels for each of the test bands. Levels were measured for each of the nine test bands and overall levels were measured for the white noise signal. Voltages were also recorded. Table 2.3 presents the results of these measurements. Reverberation Time. Reverberation time of the test capsule was also measured. Reverberation time, T60, is defined as "the time that would be required for the mean square sound pressure level, originally in a steady-state, to fall 60 dB after the source has stopped (ANSI 83.19-1974, p. 2). ANSI 83.19-1974 specifies that the reverberation time in the capsule (without the subject present) shall be between 500 and 1600 msec for each test band. Appendix E describes the instrumentation, methods and results of reverberation time measurements. The chamber was marginally below specifications (between 300 and 460 msec) of ANSI 83.19-1974 for this para- meter at all frequencies. It was assumed that this would not affect the results of this study. Spatial Uniformity of Acoustic Field. The test chamber was also calibrated in terms of the spatial uniformity of the acoustic field in relation to the subjects head. ANSI 49 Table 2.3 Signal level measurement results. Octave Band Levels 'VOltage: Right Channel Voltage: Signal Left Channel White Noise 125 Hz 250 Hz 500 Hz 1000 Hz 2000 Hz 3150 Hz* 4000 Hz* 6300 Hz* 8000 Hz* White Noise 500 1600 2500 1600 400 85 300 650 760 540 540 mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts 500 30 30 50 140 450 10,500 15,000 12,500 15,000 1,000 mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts mvolts 88.0 90.0 100.0 97.0 86.0 81.0 103.0 102.0 97.0 94.0 89.0 dB dB dB 8‘8 dB dB dB dB *20 dB pad engaged. 50 83.19-1974 specifies that the SPL measured at six positions ' relative to the center of a subjects head (i 10 cm in the front-back dimension and i 15 cm in the up-down and right- left dimensions) shall remain within a range of 6 dB for all test bands. Further, the difference in SPL between the right-left positions shall not exceed 2 dB. Appendix F describes the instrumentation, methods and results of spatial uniformity measurements. The chamber and signal presentation system conformed to the specifications of ANSI 33.19-1974 for this parameter. Ambient Noise. ANSI S3.19-1974 specifies that the ambient noise in the test capsule (with test instrumentation on and no test signal present) shall not exceed specified levels. Because of laboratory equipment constraints, ambient noise levels could not be measured for individual octave bands. The overall ambient noise was measured in the test capsule using the instrumentation described in the spatial uniformity of the acoustic field section and using dBA, dBC and linear metering characteristics on the measurement amplifier. The following ambient noise levels were obtained: (1) 33.0 dBA (2) 43.5 dBC (3) 44.0 dB linear A low amplitude, low frequency hum was noted in the test capsule and although these levels appear high, the character of the hum was such that it did not interfere with preliminary testing. The source of this hum was 51 assumed to be the power amplifier. This was confirmed when ambient noise measurements were repeated with the power amplifier turned off. The following results were obtained: (1) 24.5 dBA (2) 33.0 dBC (3) 37.0 dB linear EXPERIMENTAL PROCEDURES Subject Screening Subjects were required to read a statement of purpose and sign an informed consent release form (Appendix A). All subjects completed a case history form and underwent otoscopic, audiometric and impedance screening. The equipment used for subject screening was an audiometer (Tracoustics Program III) and a middle ear analyzer (Grason Stadler Model 1723). Subjects were screened in an IAC sound treated chamber in accordance with ANSI 53.6-1969. Subjects who met criteria (Appendian) were given the pre- test and scheduled for data collection. Subjects reported no exposure to excessive noise for at least one hour prior to screening. RunningVCalibration The signal presentation system was calibrated for level each time the system was used (i.e., for each subject). This was accomplished electronically by (1) driving the system with tape-recorded level calibration signals (1000 Hz tone 52 and white noise), (2) setting intermediate voltage ampli- fiers, the equalizer, and the audiometer attenuator to their preset positions, and (3) adjusting the level controls of the power amplifier to yield criterion voltages at the outputs of the amplifier with the loudspeakers in circuit. Voltage was monitored with a voltmeter (Bruel & Kjaer Model 2409) and a routing switch that allowed connection to either (or neither) of the power amplifier outputs. System presets were as follows: tape recorder output level - 0 VU; signal line amplifier output level - -4 VU; audiometer input level - -3 VU; and audiometer recording attenuation pen at 100 dB. Data Collection Procedures Test procedures followed those prescribed by ANSI 83.19-1974. Four groups of ten subjects each were tested. Group A consisted of ten subjects (five females; five males) who participated in the training program and experimenter fit of the hearing protection devices. Group B subjects consisted of ten subjects (five females; five males). This group also participated in the training program, but used subject fit of the protectors. Group C (five females; five males) did not participate in the training program and were fit by the experimenter. Group D (five females: five males) did not participate in the training program and used a subject fit strategy. Subjects reported no exposure to excessive noise 24 hours prior to data collection. Subjects were given instructions 53 (written and verbal) and permitted to ask questions. Sub- jects were seated in the soundfield using the head position sensing system. No signals were present for five minutes prior to testing. A brief training session consisting of two, one-minute threshold tracings using white noise were completed. Any subject who presented an average excursion size greater than 15 dB was dismissed. Groups A and C were fit with the protection device by the experimenter in accordance with manufacturer's instructions and ANSI 83.19- 1974. Subjects were instructed not to manipulate the protector in any way. Groups B and D (subject fit) were given the manufacturer's directions and instructed to fit the device. They were allowed to manipulate the device in order to obtain a good seal prior to testing. A white noise was presented at approximately 60 dB SL for purposes of manipulation. Once the hearing protection device had been manipulated and attenuation found satisfactory to the subject, further manipulation was not allowed. The first measurement obtained following the brief training session was an unoccluded measurement. The order of occluded and unoccluded measurments were alternated. Three separate trials of each measure (open and occluded) were obtained at each of the nine test bands. The test bands were also randomized. Rest periods were provided to the subjects. The psychophysical method of adjustment was used for all trials. Each separate occluded trial included a refit of the hearing protector using a new pair 54 of protectors. After the fitting of hearing protection devices (experimenter and subject), subjects were asked to engage in vigorous jaw motions to insure proper fit. The run protocol used in this study is found in Appendix B. DATA REDUCTION Form and Volume of Subject Data Following data collection, there were 40 pre- and post-tests. These were scored and tallied as percentage correct scores. Two thousand one hundred and sixty fixed frequency Bekesy tracings were also obtained. These included three trials for each subject for unoccluded and occluded thresholds at each of the nine test bands. Pre- and Post-Test Scores Raw data gathered from the pre- and post-tests were in the form of percent-correct responses. Pre-tests were scored such that distractor items were ignored. Information gain (the signed difference score between post-test and pre- test results) was computed for each subject. Attenuation.Data Raw data were in the form of 45—second fixed frequency Bekesy tracings. Thresholds were defined as the mean of the mid-point of the last ten excursions. Attenuation was computed by subtracting the mean threshold for each occluded condition from the mean threshold for each unoccluded con- dition. 55 Noise Reduction Ratings (NRRs) Finally, Noise Reduction Ratings (NRRs) were computed for each subject group. These were summarized in tabular form and compared to each other and the manufacturer's listed NRR for the hearing protection devices. CHAPTER III RESULTS INTRODUCTION In September, 1979, the Environmental Protection Agency ruled that all domestic hearing protection devices must bear a label containing the devices Noise Reduction Rating (NRR). Studies have indicated that the laboratory-generated NRR relates poorly to industrial field generated NRR values. It has been suggested that the NRR underestimates the actual protection provided in industrial environments. This study sought to determine whether the method of fitting hearing protectors (experimenter vs. subject) and subject training (present vs. absent) affect attenuation and NRR data obtained in a laboratory testing situation. Addition- ally, the study sought to assess the effect of a particular mediated training package upon knowledge of basic auditory function and noise-induced hearing loss among initially naive subjects. Forty normal hearing adult listeners (20 females; 20 males) served as subjects. Subjects were randomly assigned to one of four groups. Group A consisted of ten subjects who participated in a commercial training program and whose hearing protectors were fit by the experimenter. Group B consisted of ten subjects who participated in the training 56 , 57 program and who fitted their own hearing protectors. Group C consisted of ten subjects who did not participate in the training program and who utilized an experimenter fit strategy. Group D consisted of ten subjects who utilized a subject fit strategy. Group D subjects did not participate in the training program. All forty subjects were given a 40 question pre-test and a 20 question post-test designed to assess the informational and affective effects of the training program. Alternate forms of the tests were administered at the time of audiometric screening and again directly following acquisition of real- ear attenuation thresholds. Attenuation data was gathered in accordance with ANSI 83.19-1974 "Method for the Measurement of Real-Ear Protection of Hearing Protectors and Physical_ Attenuation of Earmuffs." DATA REDUCTION AND ANALYSIS STATISTICAL PROCEDURES The following descriptive statistics were used: (1) mean (2) variance (3) standard deviation (4) range (5) standard error of the mean (6) confidence interval of the mean (7) skewness (8) kurtosis 58 (9) Pearson product-moment correlation coefficient (10) mean absolute deviation Computation of these statistics was accomplished by hand and by Egg (Econometric Linear Forecasting), an Apple II computer program published by the Winchendon Group of Alex- andria, Virginia. The analyses of variance (ANOVAs) were computed using the ELF and ANOVA computer programs. The ANOVA is an Apple II computer program published by Human Systems Dynamics of Northridge California. These programs provide normal ANOVA summary tables, plus estimates of the exact probability of occurence of observed F-ratios assuming the null hypothesis is true. Other statistical procedures included the Fisher r - Z, z - r transform, omegaz, and eta2 (Hays, 1973). These computations were made as indicated by the outcome of the ANOVAs. The significance criterion for all inferential statistics was Pa < .05. The NRRs were computed by a computer program written by M.R. Chial, Ph.D. This program followed the EPA NRR method and was run on the Apple IIe computer (see Appendix G for source code). INFORMATION GAIN Informational tests were scored in terms of percent correct. Information gain was computed as the difference between the pre- and post-test scores for each subject. 59 Appendix H contains the pre-test, post-test and information gain scores for each subject. Only the content items on the pre-test were scored and used for analysis. The distrac- tor items were disregarded. Description of Outcomes Table 3.1 presents means, standard deviations, and standard errors of the mean for pre-test scores, post-test scores, and information gain scores obtained from all four groups of subjects. Figures 3.1 and 3.2 graphically present the same information. Appendix I presents additional summary information for these data. Mean pre-test scores ranged from 65.5% (Group B-trained) to 71.5% (Group D-untrained) across groups, with a grand mean of 67.5%. This suggests that the four groups did not differ appreciably in pre-experimental knowledge of hearing, hearing loss and hearing protection. Mean post-test scores ranged from 62.0% (Group C-untrained) to 82% (Group A-trained). Mean information gain was 15.5% and 12.0% for the two trained groups, and -4.0% and -1.5% for the two untrained groups. Thus, training appeared to increase the difference between post-test and pre-test performance. Standard deviations for all three measures were moderate across groups. Pearson product-moment correlation coefficients were computed to assess the degree of relationship among (1) pre- and post-test scores, (2) pre-test scores and information gain scores, and (3) post-test scores and information gain 60 Table 3.1. Means, standard deviations and standard errors of the mean for pre-test, post-test, and infor- mation gain scores (percent correct) across groups. Pre-Test Post-Test Information Gain Group A (experimen- ter fit; training) Mean 67.0 82.5 15.5 Standard Deviation 16.0 12.5 23.7 Standard Error of the Mean 5.07 3.96 7.51 Group B (subject fit; training) Mean 65.5 77.5 12.0 Standard Deviation 15.5 13.7 14.0 Standard Error of the Mean 4.91 4.36 4.42 Group C (experimen- ter fit; no training) Mean 66.0 62.0 -4.0 Standard Deviation 13.3 10.4 15.9 Standard Error of the Mean 4.20 3.27 5.04 Group D (subject fit; no training) Mean 71.5 70.0 -1.5 Standard Deviation 10.6 8.2 13.6 Standard Error of the Mean 3.34 2.58 4.28 Groups A - D Mean 67.5 73.0 5.5 Standard Deviation 13.7 13.5 18.7 Standard Error of the Mean 2.16 2.13 2.95 Percentage Correct 61 100 T 1 l I H ,= Pre-Test Scores H = Post-Test Scores 80 _. '- 60 '_ _. 401— fi 20 _— ‘- VI 1 l L 0 l l A B C D Figure 3.1. Experimental Group Pre-test and post-test means for all groups. (Group A=experimenter fit and training; Group B=subject fit and training; Group C = experimenter fit and no training; Group D = subject fit and no training.) Percentage Correct 90 70 50 30 10 -10 62 F l I l J— —l —— “W L I l .L A B C d D Figure 3.2. Experimental Group Information gain means for all groups. (Group A=experimenter fit and training; Group B= subject fit and training; C=experimenter fit and no training; D=subject fit and no training.) Group Group 63 scores. Table 3.2 summarizes these results. As can be seen when viewing Table 3.2, the correlation between pre- test and post-test scores was statistically significant for Group B and for Groups A - D. Statistically significant correlations were also obtained for Groups A, C and D and Groups A - D for the post-test to information gain relation. Analysis of Outcomes An analysis of variance (ANOVA) was performed on pre- test scores to assess potential bias in the assignment of subjects to experimental groups.' Table 3.3 reports an F- ratio of .38 indicating no statistically significant differences among groups. Table 3.4 reports a two-way ANOVA where information gain (the difference between post-test and pre-test scores) was the dependent variable. Hearing protection device fit and training were the independent variables. F-ratios were .01 for fit, .3 for the fit-training interaction and 9.1 for training. Strength of statistical association (wz) was computed for the effect of training. The resulting wz value was .17, indicating 17% of the total variance in information gain is accounted for by training. REAL-EAR ATTENUATION DATA Attenuation data initially were in the form of 2160, 30-second fixed-frequency Bekesy tracings. Threshold was defined as the mean of the midpoints of the last ten excur- sions. Attenuation was computed by subtracting the mean 64 Table 3.2. Correlation coefficients relating pre-test, post-test and infromation gain scores for four groups of subjects. Groups A and B received training; Groups C and D did not. A B C D A-D Pre-Post Test -.37 .55 .11 -.03 .55 Pre-Information Gain -.87 -.57 -.77 -.79 -.59 Post-Information Gain .78 .37 .55 .63 .68 df = 9 for Groups A, B, C, and D Significance criterion for Pa 1 .10: r 1 .52 df = 39 for Groups A - D Significance criterion for P3 1 .10: R 1 .26 65 mm comm Hmuoe mmm.oma om mnoh Canoe: In In mm. mm m mmm smmsumm 3 a m mumsvm sow: Eoommum mo mmmummo mmuosqm mo Esm mOHSOm m .mmuoom ummuumum no ¢>oz< mo muasmmm .m.m manna 66 mo. 5 m ocowmn Ho um uchAMAcmHma mm o.ommma Hobos Hm.mmm om. o.mh>oa Hmsoflmmm 1:: m. o.om H o.om cofiuomuwucH 5H. H.m. m.-e~ H m.~me~ meanness In: #0. m.~ a m.~ uflm N3 m mumsvm :mmz Eoowmum mo mmmumma mmumsvm «0 55m mousom .mmuoom :fimm coflumEu0msfl mo ¢>oz¢ aozlosu mo muasmmm .v.m manna 67 threshold for each occluded condition from the mean threshold for each unoccluded condition. Individual subject atten- uation data (means and standard deviations across trials) are presented in Appendix J. Reliability of Data Reduction Reliability of data reduction was assessed as follows. Four subjects (10%) were selected at random from the larger group of forty listeners. Threshold tracings from these subjects were independently re-analyzed by an experienced audiologist who determined mean attenuation across three trials at each noise band. These results were compared to those produced by the experimenter (see Appendix K). The worst-case difference in measured attenuation was less than 2 dB. Mean absolute deviations (subsummed across test bands within subjects) ranged from 0.2 dB to 0.6 dB. Correlation coefficients (r) were computed to index the consistency of measured attenuation across test bands. For each of the four subjects so considered, correlations were r = .99. Thus it appears that data reduction methods were highly reliable. Description of Outcomes Table 3.5 presents the group attenuation data (means, standard deviations and standard errors of the mean) across subjects. Figures 3.3 - 3.6 display these results graphically. Table 3.6 presents the group attenuation data (means, standard deviations and standard errors of the mean) across trials and subjects. This is the descriptive statistical method 68 mo.~ co.~ nH.~ mn.~ oh.~ cm.d mm.~ om.H om.~ com: ecu «0 sebum ouuocoum m.v c.v v.9 c.h H.m h.m m.m n.v m.v cofiumd>ma oumocmum m.nm m.nm H.om H.om H.v~ m.v~ H.va ~.m m.oa com: Amcecfimuu 0: mafia oomw32w. a m30uo no.~ cm.m mv.a om.~ wo.a om.~ mv.a mv.~ mm.a com: mcu uo uOuum ouoocoum m.v m.n m.v 5.9 ~.m o.m m.v v.v m.v cOMumw>va ouoocoum a.~m o.mn o.~m v.~m ~.m~ m.h~ vavu m.o~ v.c~ com: AmcmcMMuu 0: “new umucmswumdxmv O docuo om.N o~.m om.~ mm.~ cm.~ nc.N mc.~ om.a od.~ com: ago we uOuum oumocmam n.m m.m m.o o.e v.o ~.o H.G m.m m.o cowuts>mo oumoceum o.m~ c.cm n.h~ ¢.w~ m.o~ n.~d o.ad m.h m.h coo: Amehcsmuu “new hoowmmmc m mwoto oa.u o~.H co. no. oo.a >5. mm. mm. mo.a com: mcu uo uOuum oumocoum m.m o.m v.~ m.~ ~.m m.~ m.~ m.~ ~.m scauow>oo oumocmom o.mm o.vm a.~m ~.om m.m~ w.o~ ~.v~ ~.m A.m cow: Amcflcwnuu “own usucmsdusaxvv 4 macaw coco comm oocv oman ccow coca can own mma musmmmz accosqoum .Adsoum Hod muomflosm oav muomflosm mmouom sumo cofiumscwuus .m.m manna 69 Figure 3.3. 60 l T IL I T l I I I so _.' .— I - : 40 L... — 0 on U m 2 me~ 30 __ Lam -q +Jo ‘V c m m E 20 h— _ 10 L— .— 0 I l- 1 l l I I L J 125 250 500 1k. 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) 15 .3 1 l l l I l l r I .3 12 .. A—A = SD's across subjects .4 > _ I - . 3" 9)- E}_{j - SD 3 across trials and subjects_ 11% “V 6 8 g 3 u 0 l m 125 250 500 1k 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) Means and standard deviations of attenuation data across ten subjects -- Group A (experimenter fit and training). Mean Attenuation 70 Standard Deviations 60 l7 I I I I l I *I r 40 b - 23 30 __ 3 — 20 _. __ 10 — .J 0 I I I I I I J I I 125 250 500 1k 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) 15 1 I I I I I I I I 12 - H = SD's across subjects su 'cts _1 - SD's across trials 5 9 _. .— 3 A 23 6 _. ‘..L .4 3 (L al- 0 I J I I 1 1 I 1 I 125 250 500 1k 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) Figure 3.4. Means and standard deviations of attenuation data across ten subjects -- Group B (subject fit and training). 71 60 II _—I If I I7 71* I I I so .. .4 40 _. .— c O -H 4.} 3 c 30 __ 0A lam 44o dv c 8 20 _ z — 10 - 0 I I I I I I I I I 125 250 500 1k 2k 3.15k (“C 6.3k 8k Center Frequency (Hz) m c 3 15 3 III I I I If I I I I '3 12 - = SD's across subjects _‘ = ' ' d b' t g 9 _ 0.0 SD 3 across trials an su jec s _ Eco E 31L ‘- m 0 I I I J .l l I I. L 125 250 500 1k 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) Figure 3.5. Means and standard deviations of attenuation data across ten subjects -- Group C (experimenter fit and no training). 72 Standard Deviations 60 I’ ’I I I’ l I I I I 50 _. _‘ o -H 4.) m 330 MA— — +Im u'o “V c M £20— ... 10 - .— 0111111111 125 250 500 1k 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) 15 7 I I F1 I I I I 12 - £3.£§ = SD's across subjects _. = SD's across trials and subjects 9rm: m 361-3: 31" 0111111111 125 250 500 1k 2k 3.15k 4k 6.3k 8k Center Frequency (Hz) Figure 3.6. Means and standard deviations of attenuation data across ten subjects -- Group D (subject fit and no training). 73 cd.~ no.~ mm.~ cm.~ om.~ m~.~ m~.N mm.~ om.a com: on» no Houum oumocmum n.@ «.9 0.5 m.5 m.m 5.w 5.9 m.m 5.m cofiuow>mo oumocMDm m.mm m.mm A.om H.om A.v~ m.va «.vu ~.m m.o~ com: .mcficfionu 0: “new oomnosmv a msouo nH.~ mm.~ c~.N oc.~ Q5.~ ou.~ o5.H oo.~ om.~ cam: may no nouum ouaocMDm v.w m.m w.w c.o H.m m.o n.m o.m o.m cofiumd>mo oumocmum o.un m.mn m.~n v.~m «.mu m.5~ v.9H m.oa v.oH com: AmcflcmmHu oc uuwu umucmsauvdxmv U dsouu m~.m mv.m co.m om.~ om.~ m~.~ MN.~ oa.~ ov.~ coo: mnu mo uOuum oumosoum 5.m m.oH c.m m.m 5.m v.u 5.9 m.o «.5 cOMuMM>wo oumocmum o.m~ o.on 5.5m v.0u m.o~ m.~u o.~a m.5 w.5 com: .mesehmuu Luau antenna. a macho ca.~ ov.~ ow.~ co.~ mo.d m~.~ 5m. cm.~ mm.~ cam: msu «0 wouum oumocmum m.o v.5 c.m o.v m.v 5.n m.~ ¢.m m.v sewumfi>mo oumocoum c.nm w.vm «.mm ~.on n.mn 5.@H N.va ~.m a.m coo: AmcwaMLu “awn umucmeumdxo. d dsouo coco comm ooov Oman coo“ coca can own mm“ musmnmz aocmsqwuh .Amamauu m museum Hmm muomflnsm oav namwuu osm muomnosm mmouoo sumo :OHuossmuud .o.m magma 74 defined in ANSI 83.19-1974 where results are described based upon 30 observations (i.e., 10 subjects, 3 trials). Figures 3.3 - 3.6 also display these standard deviation results graphically. Table 3.7 shows the descriptive statistics on attenuation data for groups collapsed across fit and training for Groups A and C (experimenter fit), Groups B and D (sub- ject fit), Groups A and B (training) and Groups C and D (no training). Appendix L presents additional summary information for these data. The experimenter fit groups (A and C) displayed greater mean attenuation at all frequencies except 125 Hz than did the subject fit groups (B and D). The experimenter fit groups also displayed lower standard deviations and standard errors of the mean than did the subject fit groups. This result was consistent across all frequencies. The trained groups (A and B) showed greater mean attenuation results across all fre- quencies than did the untrained groups. The trained groups also showed lower standard deviation results at all frequen- cies except at 250 Hz, 2000 Hz and 3150 Hz. The standard errors of the means revealed no pattern of difference between the trained and untrained groups. Reliability of Subject and GroupiData Reliability was assessed in several ways. At the level of the individual subject, reliability was assessed using the Pearson product-moment correlation coefficient computed to index consistency of measured attenuation for each test band in Trials 1 and 2, Trials 2 and 3 and Trials 1 and 3. Tables 75 -e. ~em. new. ace. new. ems. nee. can. one. can: use Lo Lotta teeeeesm e.m s.e e.m e.m m.m e.m e.e m.v e.m conuns>me teeteesm ~.~m o.mm m.om e.m~ m.n~ n.ms o.ms a.e ~.m eta: Ammzcoo -vo oouocuom n.4m n.mm m.om m.m~ ~.nn m.v~ m.- m.5 m.5 can: chfiinué 02v a 62.. U madam mo.~ 5m. m~.~ ~m.~ mm. m~.H m~.~ mo.~ mo.H com: on» no uOuum oumocnum m.v n.v m.m m.m N.v m.m ~.m m.v o.v sequow>wo puoocoum ~.mm 5.mm v.~m m.om 5.v~ v.oa m.v~ ¢.m o.o~ com: Amc«cflou9v a com < mmmOuo oo.~ ~5.~ mv.~ wm.~ om.~ ~m.~ vn.~ . ca.~ m~.~ cam: and no uOuum ouoocoum H.5 5.5 9.9 H.5 0.5 a.m o.m ~.m 5.m nodum«>mo ouoocmbm m.~m m.~m m.m~ m.m~ o.- m.m~ m.- m.m m.m cum: lust nommnsmc a act a mmmoto cm. mm. 55. vm. H5. om. v5. vm. vm. com: uzu mo uOoom oonozmom ~.v 5.m v.m 5.m ~.m m.v m.m 5.m N.v :Ofiumd>va oococnom m.~m ~.vm m.~n m.o~ m.m~ m.5~ m.v~ v.5 ~.m cos: Auwm umucmeuomxmv 0 pct < mascoo ooom 00mm ooov Oman ooow coca com omm mwa vusnavz 50cmsqmum .muomflnsm mmouom MUQU COHUMSCQHH¢ .5.m manmE 76 3.8 - 3.11 show these results for individual subjects within groups. These data were then averaged using the Fisher r - Z transform (1) across trial pairs within subjects and (2) across subjects within groups. Reliability across trial-pairs and subjects were .90 for Group A, .85 for Group B, .88 for Group C and .88 for Group D. Statistically significant correlations (Pa 1 .10) were those which exceeded .55. Based on this, all subjects except one exhibited reliable results. All four groups of subjects showed Very high correlations indicating very dependable relations. At the group level, the indices of reliability were the standard deviations, standard errors of the mean and Pearson product-moment correlation coefficients (for each group and for each noise band). Standard deviations and standard errors of the mean were moderate suggesting reasonable stability in attenuation across trials within groups. The opposite result was obtained when reliability was assessed using Pear- son product-moment correlation coefficients. Tables 3.12 - 3.15 present correlation coefficients for each group and each noise band. These data were then averaged using the Fisher r - Z transform (1) across trial pairs for each noise band and (2) across noise bands for each group. Reliability across noise bands for each group was .01 for Group A, .68 for Group B, .40 for Group C and .57 for Group D. Statistically significant correlations (P i .10) were 01 those exceeding .55. Only Groups B and D (subject fit) 77 Table 3.8. Reliability of individual subject data -- Group A (Experimenter fit and training). Subject Tl - T2 T2 - T3 T1 - T3 Mean 1 .80 .92 .70 .83 2 .94 .84 .82 .87 3 .85 .91 .78 .85 4 .98 .87 .90 .93 5. .93 _ .88 .94 .92 6 .93 .91 q .96 .93 7 .77 .87 .96 .89 8 .90 .82 .89 .87 9 .95 .95 .89 .93 10 . .92 .91 .96 .93 Mean .90 .89 .88 .90 df = 8 significance criterion for Pa 1 .10: r > .55 ~78 Table 3.9. Reliability of individual subject data -- Group B (Subject fit and training). Subject T1 - T2 T2 - T3 T1 - T3 Mean 1 .36 -.46 -.20 -.34 2 .84 .94 .89 .89 3 .93 .91 .93 .92 4 .92 .99 .92 .96 5 .88 .86 .92 .89 6 .98 .98 .98 .98 7 .88 .91 .71 .85 8 .93 .90 .97 .94 9 .98 .97 .97 .82 10 .81 .98 .83 .91 Mean .85 .80 .79 .85 df = 8 significance criterion for Pa i .10: r > .55 79 Table 3.10. Reliability of individual subject data -- Group C (Experimenter fit and no training). Subject Tl - T2 T2 - T3 T1 - T3 Mean 1 .97 .98 .94 .96 2 53 86 74 73 3 .96 .84 .86 .90 4 .81 .80 .89 .83 5 .44 .91 .85 .79 6 .95 .97 .96 .96 7 .88 .92 .86 .88 8 .95 .95 .95 .95 9 94 .88 .89 .90 10 . .93 .89 .94 .92 Mean .84 .90 .89 .88 df = 8 significance criterion for Pa 1 .10: r > .55 80 Table 3.11. Reliability of individual subject data -- Group D (Subject fit and no training). Subject T1 - T2 T2 - T3 T1 - T3 Mean 1 .92 .97 .96 .95 2 .96 .92 .95 .94 3 .95 .95 .96 .95 4 .93 .72 .82 .84 5 .97 .94 .94 .95 6 .77 .75 .64 .72 7 .97 .98 .96 .97 8 .72 .71 .87 .78 9 .94 .96 .95 .95 10 .80 .74 .83 .79 Mean .89 .86 .89 .88 df = 8 significance criterion for P < .10: r > .55 Table 3.12. 81 Reliability across trials and noise bands for Group A (experimenter fit and training). significance criterion for Pa 1 .10: r > .55 Frequency T1 - T2 T2 - T3 T1 - T3 Mean 125 .22 .26 -.08 .13 250 .21 .23 -.18 .09 500 -.02 .21 .38 .19 1000 .03 .30 -.22 .04 2000 -.44 .07 .66 .09 3150 -.25 -.12 -.04 -.14 4000 -.15 -.56 .13 -.19 6300 .44 -.44 -.25 -.08 8000 -.30 .14 .14 -.007 Mean -.03 .01 .06 .01 df = 82 Table 3.13. Reliability across trials and noise bands for Group B (subject fit and training). Frequency Tl - T2 T2 - T3 T1 - T3 Mean 125 .81 .67 .49 .67 250 .41 .86 .37 .55 500 .82 .88 .80 .83 1000 .77 .93 .69 .80 2000 .84 .95 .86 .88 3150 .21 .85 .27 .44 4000 .67 .20 .81 .56 6300 .78 .95 .63 .79 8000 .47 .81 .59 .62 Mean ' .64 .79 . .61 .68 df = 8 significance criterion for Pa 1 .10: r .55 83 Table 3.14. Reliability across trials and noise bands for Group C (experimenter fit and no training). Frequency Tl - T2 T2 — T3 T1 - T3 Mean 125 .57 .81 .43 .60 250 .85 .43 .69 .66 500 .70 -.007 .52 .40 1000 .59 .50 .57 .55 2000 -.27 -.30 .58 .003 3150 .57 .46 .40 .48 4000 -.02 .19 .26 .14 6300 .20 .33 .65 .39 8000 .15 .48 .45 .36 Mean .37 , .32 .51 .40 df = 8 significance criterion for Pa 1 .10: r i .55 84 Table 3.15. Reliability across trials and noise bands for Group D (subject fit and no training). Frequency T1 - T2 T2 - T3 T1 - T3 Mean 125 .45 .63 .17 .42 250 .59 .59 .42 .53 500 .78 .73 .57 .69 1000 .71 .46 .61 .59 2000 .63 .56 .77 .65 3150 .68 .87 .87 .81 4000 .79 .60 .56 .65 6300 .34 .59 .32 .42 8000 .33 .54 .21 .36 Mean .59 .62 .50 .57 df = 8 significance criterion for Pa 1 .10: r > .55 85 exceeded this criterion across noise bands and trial pairs. Groups A and C (experimenter fit) did not. Thus it appears that the experimenter fit strategy produced less reliable results across trials and noise bands than did the subject fit strategy. To determine whether training or fit affected reliability, correlation coefficients were further averaged across group pairs, trial pairs and noise bands. Significant correlations were those which exceeded .38 at the Pa 2 .10 significance level. Table 3.16 presents the reliability results across trials and noise bands for the experimenter fit groups (A and C). The reliability across noise bands and within trial pairs for the experimenter fit groups was .24, indicating less than significant reliability. For the intra-noise band, inter- trial conditions, significant reliability was only found at 125 Hz, 250 Hz and 1000 Hz. Table 3.17 presents correlations for the subject fit groups (B and D). All of the inter-trial, intra-noise band and inter- noise band, intra-trial conditions achieved significant corre- lation coefficients. Thus, it appears that the subject fit groups were reliable across trials and noise band. Tables 3.18 and 3.19 present correlations for the training groups (A and B) and the no training groups (C and D), respectively. Most of the inter-trial, intra-noise band correlation coefficients and all of the intra-trial, inter-noise band correlation coefficients were significant. For the most part then, the training and no training groups were reliable across trials and noise bands. 86 Table 3.16. Reliability across trials and noise bands for - Groups A and C (experimenter fit). Frequency Tl - T2 T2 - T3 T1 - T3 Mean 125 .46 .54 .25 .42 250 .66 .36 .41 .47 500 .57 .04 .44 .35 1000 .47 .45 .41 .44 2000 -.33 -.16 .60 . .04 3150 .20 .20 .24 .21 4000 -.07 -.11 .21 .01 6300 .30 -.15 -.07 .03 8000 -.11 .29 .27 .15 Mean .24 .16 .31 .24 df = 18 significance criterion for Pa 1 .10: r > .38 87 Table 3.17. Reliability across trials and noise bands for Groups B and D (subject fit). Frequency T1 - T2 T2 - T3 T1 - T3 Mean 125 .64 .69 .38 .57 250 .48 .71 .42 .54 500 .67 .82 .57 .69 1000 .69 .65 .62 .65 2000 .74 .81 ' .76 .77‘ 3150 . .37 .87 .40 .55 4000 .73 .24 .52 .51 6300 .57 .63 .80 .67 8000 .44 .73 .51 .56 Mean .59 .68 ‘ .55 .61 df =1’8 significance criterion for Pa 1 .10: r > .38 88 Table 3.18. Reliability across trials and noise bands for Groups A and B (training). Frequency Tl - T2 T2 - T3 T1 - T3 Mean 125 .61 .54 .34 .50 250 .37 .64 .23 .41 500 .65 .80 .71 .72 1000 .64 .77 .57 .66 2000 .52 .80 .78 .70 3150 .02 .73 .11 .29 4000 .36 .11 .63 .37 6300 .67 .59 .33 .53 8000 .24 .60 .50 .45 Mean ' .45 .62 .46 .51 df = 18 significant criterion for Pa 1 .10: r > .38 89 Table 3.19. Reliability across trials and noise bands for Groups C and D (no training). Frequency Tl - T2 T2 - T3 T1 - T3 Mean 125 .46 .69 .20 .45 250 .73 .50 .49 .57 500 .65 .49 .48 .54 1000 .67 .48 .57 .57 ,2000 .32 .14 .64 .37 3150 .58 .65 .65 .63 4000 .46 .34 .37 .39 6300 .20 .41 .44 .35 8000 .22 .50 .34 .35 Mean .48 .47 ‘ .46 .47 df = 18 significant criterion for PG i .10: r > .38 90 Based upon the results presented in Tables 3.16 - 3.19, it can be concluded that: (1) experimenter fit of hearing protection devices yield unreliable results across trials and noise bands; (2) subject fit of hearing protection devices yield reliable results; and (3) the presence of a training program did not affect the reliability. Relation Between Information Gain and Real-Ear Attenuation Relations between information gain and mean attenuation for three test bands (125 Hz, 1000 Hz, and 8000 Hz) was assessed for each group using Pearson product-moment correlation coefficients. Table 3.20 presents these results. Significant correlations (Pa 1 .10) were those which exceeded .55. Based on this,it appears that there is no statistically significant relationship between information gain and attenuation at 125 Hz, 1000 Hz, and 8000 Hz for any of the experimental groups. Analysis of Outcomes A three-way ANOVA was performed on the attenuation data where the factors were frequency, hearing protector fit (experimenter vs. subject) and subject training (present vs. absent). The significance criterion for this ANOVA was .05. Table 3.21 summarizes results of this ANOVA. The two 91 Table 3.20. Relation between information gain and real- ear attenuation for all groups. Center Frequency (Hz) 125 1000 8000 Group A (experimenter fit and training) -.07 .05 -.35 Group B (subject fit and training) -.38 -.47 -.40 Group C (experimenter fit and no training) -.32 .18 .09 Group D (subject fit and no training) .32 .04 .09 df = 8 _ significance criterion for Pa 1 .10: r > .55 92 mo. N am ocoxmo no no unmoflwflcmfime mmm mea.aee~v mmmmm 111 1111 11111 5H~.HH mam mae.om~m nonnm 111 1111 mme. ~em.m m eom.ee tcmn1sn1mceenmnn1sh1uen 111 1111 New. mmo.m m mee.e~ tenh1sh1maneenne 111 5ma. mav.a oam.ma m mm~.5ma octo1>o1uem em. Hoo.v mam.~em1 mae.eeoe m esm.msmmm temn nmme a 111 1111 11111 Gmm.eeH em ~me.m~oe nonnm 111 1111 nae. ede.mse H ede.msa meeennne1sn1uee 111 ems. eme.~ mme.me~ H mme.mm~ Antenna .n> .mmnav meneemne 111 ANA. mee.m mme.m~e H mmo.-e Anomflnsm .n> .nxmv net mwwmmwm «mum mu m mumsvm Eoowmum mmumsqm mousom _ m com: mo mmwummo no How .mumo coHumscmuum mo ¢>Oz< >m31mmunu m0 muasmmm .HN.m manoe 93 main effects of fit and training approached but did not achieve significance. Similarly, the fit-by-training inter- action did not significantly influence mean attenuation. As anticipated, the main effect of test band was significant. The strength of association (etaz) for the main effect of test band was .87. Thus nearly 90% of total variation in attenuation can be attributed to the frequency of the test band. Because the main effect of test band accounted for such a large proportion of the variance in attenuation, and because the present study was motivated by an interest in the effects of fit and training, nine additional ANOVAs were performed, one for each test band. Each of these analyses was a two—way, randomized-blocks, fixed-effects ANOVA. Results are given in Tables 3.22 - 3.30. The main effect of hearing protection device fit was significant at only two frequencies (1000 Hz and 4000 Hz). In addition, the effect of fit approached significance at the test fre- quencies of 2000 Hz, 3150 Hz and 6300 Hz. The strength of association (etaz) of the fit effect for 1000 Hz and 4000 Hz was .09 indicating that only 9% of total variation in attenuation can be attributed to the factor of fit. The main effect of training and the interaction effect of fit and training were found not to be significant at any test frequency. It is believed that the significance of fit at the two frequencies was obscurred in the three-way ANOVA by the strength of effect associated with test band (87%). 94 mo. 1 am ocowmn we as ucmowwwomfimx mm o5m.mmm Hobos mmm.¢~ om Hmm.omm Hmsofimmm 11 111 vmo. mom.~ a mom.~ seauooumucH 11 mo. ooa.m vmm.55 H vom.55 mcacfimne 11 111 ooo. moo. a moo. uflm «3 mam m mumsvm sow: Eooomum mo mmmummo mmumsvm no How monsom .Nm mmH 1 sumo :oflumssmuuo mo ¢>Oz< >o3103u mo muasmmm .NN.m OHDMB 95 .0 mo. 1 m ocoxmn Ho no ucmoHuHcmHm4 am omm.om5 Hmuoe MHm.o~ mm H5~.HM5 Hnsoemmm II III omo. mum. H mam. coHuomumucH 11 5eH. omH.m omo.mw H omo.mq mchHmua 11 III mom. mvv.HH H mve.HH uHm me mu m mumsqm com: Eoommum mo mmmummo mwumsqm «0 Sam wousom .u: omm 1 sumo coHumscmuum mo ¢>oz¢ xozlozu uo muHommm .mm.m mHnmE 96 5 mo. n m ocoxmo Mo on HomonHsmHme mm m-.mHm Hmuoe mom.m~ om mmv.oom Hmsonmm 11 111 mom. omv.mH H omq.MH coHuomumucH 11 111 Hme. mmm.es H mme.ea meHeHmne II 111 How. mHm.H~ H mHm.H~ uHm N3 mam m mumsqm sow: Eoommum mo mmmummo mmuooqm no How mousom .N: oom 1 sumo coHumscmuum wo <>Oz¢ >m3103u mo muHsmmm .vm.m mHnoB 97 a mo. n m ocommn so am ucsonHcmHm« sauce mm mom.mvHH Hmo.5~ mm mom.m5m Hmsonmm 11 111 meH. Hm5.e H. Hm5.v coHuosnmn:H II m5m. om~.H www.mm H mo¢.mm mchHmoe mo. mmo. mom.¢« mmv.mMH H mmv.mMH uHm «3 mam m mumovm :smz Eoommum mo mmmummo moussqm mo Esm mousom .Nm oooHI sumo :oHusscsuus mo «>024 >s3103u mo muHsmmm .mm.m mHnsE 98 mo. 1 am oco>mo no as ussoHMHcmHme mm mom.vaH Hsuoe 5mo.m~ om o5o.5ooH Hssonmm II III «Hm. ooo.5m H ooo.5m COHuosHmUGH 11 III mom. oo5.mm H oo5.mm moHcHsue II oNH. mmv.m NH5.M5 H NH5.m5 uHm Na mam m mussvm ass: Eoommuh mo mmmnmmo monsoom mo Esm meadow .Nm ooom 1 sumo :oHusscmuus mo «>024 >m3103a mo muHsmmm .mm.m MHQsB 99 .0 mo. u m osoxso no as ucsOHwHomHm« I! .N: omHm 1 sumo soHusssmuus mo ¢>oz¢ >s3103u mo muHsmsm mm mm5.mo~H Hmuoe mom.Hm om Hv5.omHH Hssonmm 111 III mmv. mmm.mH H mmm.mH coHuosumucH 111 m5H. 5¢m.H moo.mm H mvo.mm mchHmue III mmH. vmo.~ qmo.mo H vmo.mo uHm N3 mu m mumsqm :smz Eoommum mo mmmummo mmumsom mo Eom mousom II .5N.m ansB 100 mo. 1 am occaso no as assOHuHcone mm MHN.mHNH Hnuoe 5o5.m~ om oeo.mmoH Hssonmm 111 111 So. 23.3 H 23.3 530335 111 111 mom. momoH H 393 9.33:. mo. 0N0. mHN.m« mmH.omH H mmH.omH uHm Na mom m mumovm csmz Eoommum mo mmmummo mmussvm «0 83m mousom .Nm ooov 1 sumo soHumscmuum mo 4>oz¢ >s3103u mo muHsmmm .mm.m mHose 101 6 mo. u m ocoawo no us ussOHMHcmHm4 mm Hmuoe omm.om om Hv~.5HmH Hsoonmm 111 ~5N. mm~.H mom.mv H mom.mv coHuosususH 111 111 mmm. mmm.mH H mmm.mH @5539 111 com. ovo.H mmo.oo H mmo.oo uHm N3 mam m. mussvm osmz Eoommum mo msmumso mmumsqm mo Esm mousom .Nm oomoI sumo coHussssuus mo ¢>oz< >s3103u mo muHsmmm . mm . m OHnsB 102 6 mo. u m ocoaso no us ucsoHMHcmHme mm. mmm.e~mH Hmuoe eem.em om mmm.mm~H Hmsmummm 111 mom. mmo.H mme.em H mmo.em coHuomnmneu 111 111 Hoe. mem.om H mem.om ocucumne 111 111 eve. Ho~.- H Ho~.- nun N3 mam m susswm csmz Eoommum mo mmmummo moussom mo Esm mousom .Nm oooml sumo coHumsssuus mo ¢>Oz¢ >s3103u mo muHsmmm .om.m mHosB Summary 103 To summarize: (l) (2) (3) (4) (5) (6) (7) (8) Data reduction techniques of real-ear attenuation data were highly reliable. Reliability across trial pairs, within subjects was high and statistically significant for all but one subject. Reliability across trial pairs, within noise bands was statistically significant for the subject fit groups, but not the experimenter fit groups. Reliability across trial pairs, within noise bands was statistically significant for both the trained and untrained groups. The relation between information gain and real- ear attenuation was not statistically significant. The main effect of test band was statistically significant. The main effect of hearing protector fit was significant for 1000 Hz and 4000 Hz. At these bands, experimenter fit produced greater mean attenuation than subject fit. 7 The main effect of subject training was not statistically significant. None of the two-way interactions among fit, training and test band were significant, nor was the three-way interaction. 104 NOISE REDUCTION RATINGS (NRRs) NRR values were computed by means of a computer program (Appendix G) which duplicated the procedures described by the Environmental Protection Agency (EPA, 1979). Table 3.31 presents NRRs for each subject group and for several com- binations of groups. At the group level, NRRs were calculated using the definition of standard deviation specified in ANSI 83.19-1974 (N = 3 trials x number of subjects). At the group level, and for the combined groups, NRR was cal- culated using the more traditional definition for standard deviation (N = number of subjects). In all cases, mean attenuation was calculated across trials and across subjects. The NRR for the Bilsom Propp-o-Plast device is 22.1 dB (labeled as 20 dB).* The NRRs computed from the present data were appreciably Smaller than the labeled NRR. NRR results showed that Group B (subject fit and training) and Group D (subject fit and no training) yielded the lowest NRRs (.9 and 5.1 dB, respectively). For the experimenter fit groups, the training group (A) yielded a higher NRR (11.8 dB) than the no training group (B-NRR 8.2 dB). It appears, then, that fit influenced NRRs, but training probably did not. When NRRs were computed on the basis of only fit or training (N = 20), the experimenter fit groups (A and C) and the training groups (A and B) yielded slightly larger *These numbers are based upon data generated by Paul Michael, Ph.D. at Pennsylvania State University and are offered by Bilsom per the EPA regulation. 105 Table 3.31. NRR results. (ANSI) Group(s) NRR NRR Group A (experimenter fit and training) 11.8 9.0 Group B (subject fit and training) .9 -.2 Group C (experimenter fit and no training) 8.2 8.2 Group D (subject fit and no training) 5.1 3.2 Groups A and C (experimenter fit) 9.6 --- Groups B and D (subject fit) 3.1 --- Groups A and B (training) 6.6 --- Groups C and D (no training) 5.3 --- Groups A - D 6.0 --- 106 NRRs (9.6 and 6.6 dB, respectively) than the subject fit groups (B and D) or the no training groups (C and D - 3.1 and 5.3 dB, respectively). The NRR for all 40 subjects was 6.0 dB. CHAPTER IV DISCUSSION INTRODUCTION The purpose of this study was to determine whether the variables of hearing protector fitting (experimenter versus subject) and subject training (present versus absent) affect attenuation and NRR data obtained in a laboratory testing situation. The following questions were asked: (1) (21' (4) (5) (6) Is information gain significantly affected by the presence of a training program? Do real-ear attenuation values differ significantly as a function of experimenter fitting versus subject fitting of hearing protection devices? Do real-ear attenuation values differ significantly as a function of trained versus untrained listeners? Do real-ear attenuation values differ significantly as a function of the interaction between fitting method and subject training? Do real-ear attenuation values differ significantly as a function of test band? What is the correlation between information gain and real-ear attenuation as a function of fitting method and subject training? 107 108 (7) Do NRR estimates differ as a function of experimenter fitting versus subject fitting of hearing protection devices? (8) Do NRR estimates differ as a function of trained versus untrained listeners? INFORMATION GAIN The two subject groups (A and B) exposed to the Bilsom training films "Nice To Hear" and "SOS" exhibited significant information gain when tested in a pre-test - post-test paradigm. For this reason, and because the two groups (C and D) not so exposed produced negative information gain, it is reasoned that the Bilsom films cause at least a short- term increase in subject information about hearing, indus- trial noise, and hearing conservation. OSHA (1983) acknowledges the potential value of employee training in these areas through the requirement of annual instruction (OSHA, 1983, p. 9739). Although the Bilsom materials produce information gain, and although they have been found by others to contribute significantly to the overall effectiveness of hearing conservation programs (Karmy and Martin, 1982), the present findings suggest a lack of effect upon real-ear attenuation and derived NRR values. This lack of impact is evidenced by (l) the small correlations between information gain and real-ear attenuation and (2) the non-significant ANOVA main-effect of training upon real- ear attenuation. 109 The Bilsom films devote little attention to the method by which hearing protectors should be fitted by users, prob- ably because the details of those methods vary with protector type and because the film producers chose not to assume the use of any particular protector by the viewer. Further, the training tests designed to assess information gain did not include items relating to the details of fitting protectors. In addition to viewing films, the trained subjects in Groups A and B were given instruction in the use of the par- ticular insert protector used in the study; Groups C and D received only the information printed on the protector package. Possibly, the additional instruction offered by the exper- imenter was ineffective. Alternatively, the information on the protector package was just as effective (or ineffective) as that provided by the experimenter. In any event, it appears that special measures are necessary to instruct wearers of insert hearing protectors in the correct fitting of those devices. REAL-EAR ATTENUATION AND NOISE REDUCTION RATINGS (NRRS) As noted in Chapter III, reliability was assessed at the level of the individual subject and at the group level. Reliability at the level of the individual subject (Tables 3.8 - 3.11) was significant for all but one subject. This indicates that subjects were consistent in their threshold criteria across trials and across noise bands. This was not fully the case at the group level (Tables 3.12 - 3.19) where 110 reliability was assessed across trials and across subjects, within noise bands. Only the subject-fit groups, B and D (trained and untrained, respectively) exhibited reliable results. This suggests that when subjects fit the devices themselves, they were consistent across trials. In other words, they probably developed some internalized criterion for correct fit of the protectors. When the experimenter fit the devices from trial to trial, reliability was poorer, suggesting a less consistent criterion for fit across trials. This difference may occur because subjects are able to tactually observe placement of insert protectors and ex- perimenters can only visually observe placement. When correlation coefficients were further averaged across group pairs, trial pairs and noise bands (Tables 3.16 0 3.19), the subject fit groups (B and D), the trained groups (A and B) and the untrained groups (C and D) achieved significance. This supports the contention that the subject fit strategy produced greater reliability than the exper- imenter fit strategy, regardless of training. The outcomes of the analysis of attenuation data revealed several issues. To aid in this discussion, Table 4.1 reviews the information on means, standard deviations and standard errors of the mean for the four experimental groups. Table 4.1 also presents the labeled means and standard deviations for the Bilsom Propp-o-Plast protector. These data were generated by Paul L. Michael, Ph.D. at the Penn- sylvania State University in accord with ANSI 83.19-1974. 111 s.m ~.m m.~ ~.~ o.m o.m m.~ m.~ o.m ecuunu>mo nonstanm m.mm m.H¢ m.oq N.mm o.vm N.©N N.©N m.vm o.mm :smE sumo m.Hmm£on mo.H ou.H MH.~ mm.~ o5.H oa.H om.H om.H om.H csmz szu mo uOuum ousossum m.v o.c v.0 o.5 H.m 5.m m.m 5.q m.e :oHusu>uo onsozsum m.mm m.mm H.om H.on H.vm m.vH H.¢H N.m m.oH csmz HmsHsHsuu oc uuHu uosfiosn. Q mmOHU mo.H om.H nv.H om.H co.H wo.H mv.H ov.H mo.H ass: mzu no usuum oumocsum m.v m.m m.v 5.v N.m w.m n.v v.v m.v cOHumH>vo ousocsum o.~n o.mm o.~m v.Hm «.mm m.5H v.vH m.oH v.oH cmsz AmchHsuu o: “uHu usuchHuvaxs. U macaw om.~ ou.n on.“ mm.~ om.~ mo.~ mo.~ om.H oH.~ cs9: mnu «0 sebum oumocsum 5.o w.m m.m o.5 v.o H.@ H.@ w.m m.o coHusH>mo oumocsum o.m~ o.om 5.5N v.o~ m.o~ n.~H o.HH m.5 o.5 cs0: AmchHmuu mun uovmosn. m macho oH.H o~.H om. no. mo.H e5. no. mo. mo.H ass: man no ocuum onsocnum m.m o.m v.~ m.~ ~.m m.~ m.H m.~ H.m cOHusH>sQ ousosmum o.mm o.vm m.~m ~.om m.m~ 5.oH N.VH ~.o H.o csmz HmcHsHsuu uuHu umuomsHumon. < docuo oooo oomw ooov omHm ooom oooH oom omu mmH musmsm: >ocmsqmum .muso m.Hmm:oHE Eoum oss wosum chu Eoum sumo coHusssmuu< .H.v OHnt 112 Evident from the results of this study is that atten- uation across frequencies tended to increase as frequency increased except at 8000 Hz where a small drop in attenuation was noted. Michael's data show the same trend. This was further confirmed by a three-way ANOVA which showed that the main effect of test band was statistically significant. Indeed, 87% of the total variation in attenuation was attri- buted to test band frequency. Table 4.2 presents means and standard deviations for the EAR insert hearing protector for comparison to a similar type of hearing protector (Abel, Alberti and Riko, 1982). These results represent an average of 347 subjects tested following a subject fit strategy. Although these data were gathered with different procedures and used a different. protector, the effect of frequency upon attenuation was similar to the present data: frequency increases, then drops off slightly at the higher frequencies. Results from the present study indicated that the exper- imenter fit groups (A and C) and the groups which received training(A and B) showed slightly greater mean attenuation and slightly smaller standard deviations than the subject fit groups (B and D) or the untrained groups (C and D). These effects were not statistically significant except for fit at 1000 and 4000 Hz. Martin (1982) compared attenuation on a pre-molded insert earplug in terms of experimenter fit versus subject fit. He found that the experimenter fit condition yielded higher mean attenuation and lower standard 113 Table 4.2. Attenuation results for the EAR hearing protection device (From Abel, Alberti and Riko, 1982, p. 320). Frequency Measure 125 250 500 1000 2000 3000 4000 6000 Mean 12.8 14.2 14.5 18.0 24.5 27.1 25.2 21.8 Standard Deviation 8.3 10.0 8.7 7.7 7.5 8.0 8.5 9.2 114 deviations than the subject fit group (see Table 1.2). It is not known whether these differences were statistically significant. Martin (1982, p. 275) stated This...does illustrate the need for rigorously defined fitting procedures in standard methods and, more importantly, the need for general agreement as to which type of fitting procedure should be specified. Closely associated with the attenuation results are NRRs. The published NRR for the Bilsom Propp-o-Plast device is 22.1 dB. The means and standard deviations used to compute this NRR are presented in Table 4.1 (Michael's data). NRRs obtained in this study were considerably smaller than those reported by the manufacturer of this device. As with the attenuation data, somewhat larger NRRs were obtained for the experimenter fit and training strategies. Because the NRRs were based upon the attenuation values, and because attenuation did not differ significantly as a function of fit or training, it was reasoned that NRR differences were not statistically significant. RELATION OF OUTCOMES TO PRIOR RESEARCH Berger, Kerivan and Mintz (1982) demonstrated differences in measured attenuation and NRRs for insert protectors tested by eight laboratories. These differences persist despite nominal conformity with the ANSI 83.19-1974 and EPA (1979) methods. Berger, Kerivan and Mintz (1982) attributed inter-laboratory differences to hearing protector fitting, subject selection, subject training, and data reduction techniques, but did not report details of methodological 115 differences among laboratories. Forshaw (1982) cited several reasons for variations in attenuation measurements from laboratory to laboratory. He stated that the ANSI and EPA methods are not sufficiently explicit on the selection and training of subjects or on fitting procedures. He further stated that differences may be significant from laboratory to laboratory when only ten subjects are used for testing. Because of the difficult listening task and the length of the testing (approximately 2.5 hours in the present experiment), changes in signal detection criteria may be a source of variance. Many of the subjects used in this study stated that the listening task was difficult and fatiguing. This may explain why standard deviations were higher than expected and why mean attenuation and NRRs were lower than expected. Although frequent rest periods were provided, it is felt that subject fatigue affected results. Neither the ANSI or EPA test methods provide detailed guidelines for subject selection or training. The ANSI document specifies only that subjects exhibit normal hearing bilaterally. Because more specific requirements are not provided, laboratories may differ in the rigor with which they select and motivate subjects. Paid subjects may be better practiced and better motivated than unpaid subjects; this in turn may influence the outcomes of experimental testing. Few guidelines are present in these standards with 116 regard to subject training. ANSI specifies that trained subjects are to be used. Presumably, training refers to threshold tracking, but no information is offered as to the amount of training that should be provided. The ANSI document specifies only that ...no listeners shall be selected as a subject for these tests whose variability for the open threshold of audibility...is such that a range on three successive open threshold measurements at any test band between the 250 and 4000 Hz bands is greater than 6 dB (p. 4). Using this criterion for unoccluded thresholds, five of the forty subjects used in this study would have been rejected. One of these came from Group A (experimenter fit and training), one came from Group C (experimenter fit and no training), and three came from Group B (subject fit and training). Because of this distribution, it is felt that the increased variability of these subjects probably did not affect the attenuation results in terms of the experimental factor of fit. Differences among laboratories should be expected, and indeed have been found in the presence of such ambiguities. Tobias (1982, p. 171) stated: Neither the American standard nor the EPA computational procedures says anything substantive about how to select the human subjects for testing or about how to fit the hearing protectors to the subjects' ears. Again, measurers are making choices. Some choose their subjects more or less randomly, from the belief that only with that sort of selection can the variability values give a reasonable approximation to the ways in which the protector will work away from the laboratory. Others, suggesting that the increased variability one gets with a hetero- geneous group of subjects lead to unreliable 117 results -- that is, the results are not precisely repeatable -- began to collect experienced listeners for their tests of hearing protectors. Assuming validity of Tobias' statements, and because subjects used in this study were not "professional" listeners, it is reasoned that the results obtained here give a "reason- able approximation to the ways in which the protector will work away from the laboratory" (Tobias, 1982, p. 171). This is further supported by the observation that the atten- uation and NRR results obtained in this study closely parallel those found by others in field tests (see Tables 1.3 and 1.4). Work reported by Padilla, 1976, Regan, 1977, Edwards, gt_al., 1978, and Alberti, gt_al., 1979 indicated that attenuation results obtained in the industrial field are considerably lower than laboratory results primarily because better results tend to be found when devices are fit by experimenters and because in-field studies utilize a subject fit strategy. In addition to differences in subject selection and training, factors related to hearing protector fit may serve to explain the present outcomes. These factors include: (1) the experimenter did not optimally fit the devices and the package label instructions for fitting were insufficient or not followed properly; or (2) both methods were equally effective and published data are erroneously large; or (3) proper fit was not possible for reasons of device design. Several observations are possible with regard to device 118 design and the interaction of such design with subject variables. In the present study, equal numbers of males and females were used in each group. Although subject sex is typically not reported in published studies of hearing protector effects, and although the EPA and ANSI procedures are silent on this issue, it is known that males and females differ in ear canal dimensions. It is possible that subjects with relatively large or small canals were not adequately protected by the devices used here. Second, it was noted that several of the protectors could not be used because the polyethelene cover separated from the cotton-like filling when the protectors were removed from the dispensing package. It is possible that similar separation occured after the devices were inserted, thus reducing the effectiveness of the seal between the outer surface of the protector and the canal wall. Other reasons which may explain why the attenuation measurements and NRRS obtained in this study were lower than published results relate to variations from acoustical Specifications of the test environment. The test chamber used in this study had: (1) slightly shorter reverberation times than specified by ANSI 83.19-1974; (2) higher overall ambient noise levels than specified by ANSI 83.19-1974; and (3) undetermined levels of total harmonic distortion. ANSI 83.19-1974 specifies a reverberant sound field 119 primarily to simulate the diffuse characteristics of industrial sound fields. Reverberation times are to be between 500 and 1600 msec for each test band. Measured reverberation times of the test chamber used in this Study were marginally below criteria, ranging from 300 - 460 msec. Because of the departure from specifications was minor, it is reasoned that the Shorter reverberation times measured in the laboratory had a minimal effect on measured attenuation. Total harmonic distortion could not be measured be- cause of instrumental limitations. If appreciable har- monic distortion had been present in the system, it is expected that the unoccluded thresholds would have been better (lower) than they were. Therefore, it is unlikely that harmonic distortion had an effect. As stated in Chapter II, ambient noise levels could not be measured at individual noise bands. Overall ambient noise was measured in the test chamber using linear, A-weighted and C-weighted characteristics. To determine the possible effects of increased ambient noise level on measured attenuation, the maximum permissible ambient levels for individual octave bands specified in ANSI 83.19-1974 were converted to A- and C-weighted levels and then compared to the ambient levels of the test chamber used in this study. As noted in Table 4.3, the differences in maximum permissable levels specified by ANSI 53.19-1974 and the measured levels from the test chamber were 4 dBA, 15.0 dBC and 11.5 dB linear. If it is assumed that the effect of ambient noise upon unoccluded 120 Table 4.3. Converted ambient noise levels. dBA dBC Linear Ambient noise levels obtained in test chamber 33.0 43.5 44.0 Maximum permissable ambient noise levels per ANSI 83.19-1974 29.0 28.5 31.5 Difference 4.0 15.0 12.5 121 threshold is linear with respect to level and frequency, and further that ambient noise would not affect threshold standard deviations, then NRRS can be re-computed by adding a constant to the attenuation measured at each noise band. This was done for Group A (experimenter fit and training) and new NRRS were computed. Adjusted NRRS were 15.8 dB for the "corrected" dBA levels, 28.0 dB for the "corrected" dBC levels, and 24.3 dB for the "corrected" linear levels. The NRR originally computed for this experimental group was 11.8 dB, while the labeled NRR for the device was 20.0 dB. While it cannot be known with certainty whether the increased ambient noise in the test chamber adversely affected results, it is probable that lower ambient noise levels would have produced greater mean attenuations and NRRS. The ANSI 83.19-1974 document requires fairly Strict controls on the acoustic field used for testing hearing protection devices in the hope of producing stable results. Industrial environments almost certainly do not conform to such optimal control, and there may be interactions among acoustic fields, bodies and.hearingprotection devices. These sources of variance are expected to impact the effectiveness of hearing protection devices in real environments. Although the test chamber used in this study was not completely within specifications for all parameters, and although these limitations seem to have produced less than optimal results, it appears the test environment may have been more like an industrial field environment. 122 This study was not designed to investigate the effects of inter-trial differences upon measured attenuation, but instead followed the requirement of ANSI 83.19-1974 that data be averaged across three trials (within-trial dispersion was not measured). As an ad hoc analysis, raw attenuation data (Appendix J) were studied to determine the number of subjects who gave maximum attenuation values during the last trial. Tallies are shown in Table 4.4. With only a few exceptions, at least a third of the subjects in each group and at each test band produced greater attenuation results in Trial 3. Group B (subject fit and training) yielded the greatest count across test bands; Group D (subject fit and no training) gave an intermediate count; and Groups A (ex- perimenter fit and training) and C (experimenter fit and no training) produced the smallest counts. Instances of greater attenuation for Trial 3 were similar (about 15 subjects across four groups) for the nine test bands, the exception being the 250 Hz band which yielded a count of 25. This learning effect is more potent in the subject fit groups than in the experimenter fit groups. The factor of training seems to have little impact upon the trend toward greater attenuation in Trial 3, however. Had the Trial 3 data been used instead of the mean across trials, the factor of fit may have been Significant at more than two test bands. Further, the learning effect may (in part) explain why the present attenuations and NRRS 123 Table 4.4. Ad hoc analysis of Trial 3 attenuation data. (Numbers are the numbers of subjects per group which exhibited greater attenuation at Trial 3). Frequency _ 125 250 500 1000 2000 3150 4000 6300 8000 X Group A* 6 6 3 2 6 4 3 3 l 3.8 Group B* 3 5 2 4 l 4 4 5 5 8.1 Group C* 2 6 4 4 3 4 2 4 5 3.8 Group D* 6 8 4 4 4 4 6 2 4 4.7 Total 17 25 13 l4 l4 16 15 14 15 15.9 *Group A = experimenter fit and training; Group B = subject fit and training; Group C = experimenter fit and no training; Group D = subject fit and no training. 124 were lower than those reported by Michael. Regardless, the possibility of a learning effect in subjects who other- wise satisfy response stability requirements suggests the need for study of short-term learning effects. It is apparent that there exists a wide range of variation in attenuation and NRRS generated among laboratories and industrial fields and a variety of explanations for these differences. It is felt that the ANSI and EPA testing methods should be expanded to include better ways of Similating industrial field settings. It is felt that more research is needed to determine why inter-laboratory differences exist, what specifically the differences are, and ways to resolve the differences. If this is not done, users of the devices cannot really know what protection can be expected from a particular device and consequently may be underprotected. FINDINGS This study sought to determine whether the variables of hearing protector fit and subject training affect atten- uation and NRR data obtained in the laboratory testing situation. Based on the results and analysis of the results, the present study found as follows: (1) Information gain was significantly increased by a training program. (2) Reliability at the individual subject level was high and statistically Significant for all but one subject. (3) (4) (5) (6) (7) (8) (10) 125 Reliability at the group level was statistically significant for the subject fit groups and for trained and untrained groups, but not for the experimenter fit groups. Real-ear attenuation values differed significantly as a function of test band in patterns similar to what has been reported elsewhere. Real-ear attenuation values and NRR estimates obtained for the Bilsom Propp-o-Plast device were -considerably lower than the manufacturer's labeled attenuation values and NRR estimate. At least in part, this was due to inter-trial learn- ing effects and problems of ambient noise. Real-ear attenuation values differed Significantly as a function of fit only at 1000 Hz and 4000 Hz and approached significance at 2000 Hz, 3150 Hz, and 6300 Hz. Real-ear attenuation values did not differ sig- nificantly as a function of the presence or absense of subject training. Real-ear attenuation values did not differ significantly as a function of the interaction between fitting method and subject training. The correlation between information gain and real- ear attenuation was not Significant as a function of fitting method or subject training. NRR estimates appeared to differ as a function of 126 experimenter vs. subject fit slightly. Increased NRR estimates were found for the experimenter fit strategy. IMPLICATIONS AND SUGGESTIONS FOR FURTHER RESEARCH The results and outcomes of this study (as well as others reviewed herein) present some important implications for the industrial sector. Several issues about laboratory testing of hearing protection devices, in general, remain unresolved and deserve further attention. Several issues about the data generated from this study remain unresolved as well. It has been demonstrated that differences among lab- oratories and industrial field environments exist with regard to measured attenuation and NRRS. Industries using hearing protection devices should be aware that these differences exist and that the protection devices they are purchasing and using may reflect inaccurate protection values. Further research should focus on the test methods used for measuring hearing protection devices. There is a need for an "objective" test method to account for human subject effects related to variances associated with fit and training and also those variances associated with laboratory versus field effects. In particular, future research should focus upon subject selection and training procedures and upon laboratory fitting practices. An appropriate goal for such work is to identify 127 a balance between (1) the problem of stability in measured attenuation, and (2) the problem of validity in predicting effectiveness in the field. One approach to this goal would be to employ subject fit methods and to study patterns of change in measured attenuation across trials (i.e., as initially naive listeners become more practiced through experience). The resulting "learning curve" may allow more accurate pre- diction of protector effects in real-world situations. It is expected that inter-trial effects will vary for different types of protectors. Similarly, the long-term effects of subject training should be investigated. It is expected that "professional" subjects will develop threshold criteria and self-fit Strategies which differ appreciably from those of less experienced subjects and from those of workers in field environments. The effects of sex on measured attenuation and NRRS obtained in the laboratory Should be investigated. Although industrial environments typically involve more males than females, women are present in the work force. It would be appropriate to address these issues and their effects on hearing protection device testing methods and performance. Industry assessments of hearing conservation program and monitoring audiometry are indeed important. Because NRR estimates are variable across laboratories and, perhaps, invalid, further research is needed to determine these effects on hearing conservation programs. 128 CONCLUSIONS In addition to the findings and implications discussed above, several conclusions can be stated: (1) (2) This study demonstrated that two of Bilsom, International's training films ("Nice To Hear" and "505") provide a significant amount of short- term information gain regarding hearing, industrial noise and hearing conservation. Although these training films do not appear to affect the outcome of attenuation measurements and NRRS, it is felt that these and similar training materials would provide substantial information to industrial employees and would be appropriate for use in hearing conservation programs. It was Shown that subjects exposed to the training program and those who had hearing protection devices fitted by the experimenter demonstrated slightly greater mean attenuation values and lower standard deviations; these were statistically Significant for only two frequencies. Based on this outcome and previous research, it is concluded that experimenter fit strategies employed in labor- atory test environments yield greater attenuation and NRR results than do subject fit strategies. This study supported the previous, somewhat dis- couraging finding that different laboratories produce different NRR estimates, despite (4) (5) 129 considerable effort to manage test signals, the test environment, subject selection, subject training, hearing protection device fitting method and psychophysical method. Although there are two standards which specify methods for the measurement of hearing protection devices, differences are still found when nominally similar methods are followed. Further study is indicated to determine why inter-laboratory differences exist, the magnitudes and ranges of those differences, and ways to resolve them. In its present form, the ANSI 83.19-1974 test method produces highly variable results (a) among laboratories and (b) between laboratory and industrial field settings. The method, therefore, requires further study and refinement to reduce this variability. CHAPTER V SUMMARY AND CONCLUSIONS INTRODUCTION It is well known that excessive noise can damage the human auditory system; excessive noise can cause other sorts of problems as well. Damaging noise is prevalent in indus- trial environments and protection from the harmful effects of industrial noise should focus on prevention. When appropriate engineering and administrative controls cannot sufficiently decrease noise to acceptable levels, hearing protection devices are often employed. Prior to 1979, the effectiveness of hearing protection devices were evaluated behaviorally by an absolute threshold shift procedure. In September, 1979, the Environmental Protection Agency (EPA, 1979) ruled that all domestic hear- ing protection devices must bear a label containing a single- number estimate of effectiveness designated Noise Reduction Rating (NRR). The NRR indicates the noise attenuation capability of a hearing protection device, weighted by an assumed noise spectrum and the statistical variations in band attenuation data obtained from a group of ten listeners (Juneau, 1982). The behavioral test methods underlying the NRR are described in American National Standards Institute 130 131 (ANSI) 83.19-1974 "Measurement of Real-Ear Protection of Hearing Protectors and the Physical Attenuation of Earmuffs." The EPA also requires an experimenter fit strategy for deter- mination of a devices NRR. A review of the literature indicates that the validity of the NRR is open to question. Because the NRR is obtained under optimal laboratory conditions, these may not accurately reflect effectiveness of the device in the industrial field setting where the device is used. Several studies (Padilla, 1976; Regan, 1977; Edwards, gt_al., 1978; Alberti, gt_§l., 1979; Abel, Alberti, and Riko, 1982; Berger, 1983) have shown that attenuation data (used to compute the NRR) generated in controlled laboratory settings do not accurately .reflect effectiveness of protection devices in the industrial field setting. Other studies (Berger, Kerivan and Mintz, 1982; Forshaw, 1982) have indicated that there are inter- laboratory differences among laboratories which conduct NRR tests. Reasons cited for these inter-laboratory and laboratory-field differences include subject selection and training, fit of the device, data reduction techniques and variations in acoustical parameters. This study was designed to assess the effects of hearing protector fit and subject training on attenuation and NRR data obtained in a laboratory testing situation. METHODS Subjects for this study were forty adult listeners 132 (20 females; 20 males) who exhibited otoscopic, audiometric and otologic normalcy. Subjects were randomly assigned to one of four groups consisting of ten subjects each. Two experimental groups (A and B) viewed a commercially available (Bilsom International) multi-media education program designed to emphasize several informational and affective topics related to hearing, industrial noise and hearing conservation. Groups C and D did not receive the training program. Two groups (A and C) had hearing protection devices fit by the experimenter, and the remaining two groups (B and D) utilized a subject fit strategy. All 40 subjects were administered a 40 question pre- test and a 20 question post-test. These tests were designed to assess the informational and affective effects of the training program. Information gain scores were obtained as the difference between the post-test and pre-test scores. The hearing protection device used for all measurements was the Bilsom "Propp-o-Plast" disposable, insert type plug. Real-ear attenuation at threshold measurements were taken in accord with ANSI 83.19-1974; three trials each of unoccluded and occluded measures were taken at each of nine frequencies (125 - 8000 Hz, l/3-octave bands of noise). Attenuation for a given test band was computed as the differ- ence between occluded and unoccluded threshold measures across trials. NRRS were also computed for the experimental groups. 133 RESULTS With regard to information gain, results showed that training increased the difference between post-test and pre— test performance. Higher mean information gain scores were found for the groups which received the training program (A and B). A two-way analysis of variance (ANOVA) performed on information gain with the main effects of hearing protection device fit and subject training showed that the effect of subject training was statistically Significant. Reliability of measured attenuation was assessed across trials at the level of the individual subject (across noise bands) and at the group level (across subjects). These results showed that at the subject level, all subjects but one demonstrated reliability that was high and statistically Significant. At the group level, reliability across noise bands and groups was statistically significant for the subject fit groups (B and D), the trained groups (A and B) and the untrained groups (C and D). The experimenter fit groups (A and C) did not demonstrate statistically significant reliability. Mean attenuation results showed greater attenuation and lower standard deviations for the experimenter fit groups (A and C) and for the trained subjects (Groups A and B). The relation between information gain and attenuation was assessed and found not to be statistically significant. A three-way ANOVA showed that the main effect of test band was statistically significant, but that the main effects of hearing protector fit and subject training were not. 134 Subsequent analyses revealed statistically significant results for the main effect of fit at the frequencies 1000 and 4000 Hz and approached significance at 2000, 3150, and 8000 Hz. Noise reduction ratings (NRRS) for the experimenter fit groups (A and C) and for the training groups (A and B) were greater than those computed for the subject fit groups (B and D) or the no training groups (C and D). These NRRS were compared to the labeled NRR for the device tested and were considerably lower than the labeled-manufacturer's data. CONCLUSIONS Based on the results and analysis of the results, the following conclusions are offered: (1) This study demonstrated that two of Bilsom, International's training films ("Nice To Hear" and "508") provide a Significant amount of Short- term information gain regarding hearing, industrial noise and hearing conservation. Although these training films do not appear to affect the outcome of attenuation measurements and NRRS, it is felt that these and similar training materials would provide substantial information to industrial employees and would be appropriate for use in hearing conservation programs. (2) It was shown that subjects exposed to the training program and those who had hearing protection devices fitted by the experimenter demonstrated (3) (4) (5) 135 slightly greater mean attenuation values and lower standard deviations; these were statistically significant for only two frequencies. Based on this outcome and previous research, it is concluded that experimenter fit strategies employed in labor- atory test environments yield greater attenuation and NRR results than do subject fit strategies This study supported the previous, somewhat discouraging finding that different laboratories produce different NRR estimates, despite considerable effort to manage test signals, the test environment, subject selection, subject training, hearing protection device fitting method and psychophysical method. Although there are two standards which specify methods for the measurement of hearing protection devices, differences are still found when nominally similar methods are followed. Further study is indicated to determine why inter-laboratory differences exist, the magnitudes and ranges of those differences, and ways to resolve them. In its present form, the ANSI 83.19-1974 test method produces highly variable results (a) among laboratories and (b) between laboratory and industrial field settings. The method, therefore, 136 requires further study and refinement to reduce this variability. APPENDICES APPENDIX A SUBJECT SCREENING FORMS 137 STATEMENT OF PURPOSE The experiment which you are about to participate in relates to the use of hearing protection devices commonly used in industrial settings. The purpose of this experiment is to determine differences in hearing protection device measurements. l) 2) 3) 4) 5) 6) 138 INFORMED CONSENT RELEASE I, , freely and voluntarily consent to serve as a subject in a scientific study of hearing protection device performance conducted by Kimberly A. Payne, B.A. working under the supervision of Michael R. Chial, Ph.D. I understand that the purpose of the study is to determine differences in hearing protection device measurements, which may be of future usefulness to professionals involved in hearing protection. I also understand that the procedures involved are experimental and that the results of this study will not be of any direct personal benefit to me. I understand that I will not be exposed to any experimental conditions which constitute a threat to hearing, nor to physical or psychological well being. I understand that the data gathered for this study are confidential, that no information uniquely identified with me will be made available to other persons or agencies, and that any publication of the results of this study will maintain anonymity. I engage in this study freely, without payment to me or from me, and without implication of personal benefit. I understand that I may cease participation in the study at any time. I have had the opportunity to ask questions about the nature and purpose of the study and I have been provided with a copy of this written informed consent form. I understand that upon completion of the study, and at my request, I can obtain additional explanation about the study. SIGNED: DATE: l) 2) 3) 4) 5) 6) 7) 8) 9) 139 SUBJECT CASE HISTORY Age: Sex: Do you have a history of familial hearing loss. If so, state relation and age of person. Have you ever had ear surgery? If so, what type and when? Do you have or have you ever had hearing loss, vertigo, or tinnitus? If so, please explain. Do you currently have a cold, ear infection or upper respiratory infection? Have you ever worked in a noisy industrial setting? If so, when and for how long? Were you required to wear hearing protection devices? ' Have you ever taken the following drugs? kanamycin gentamycin streptomycin dihydrostreptomycin Have you ever been treated for severe burns or a severe infection (i.e., meningitis, encephalitis, etc.)? Have you taken any drugs in the last 72 hours (excluding coffee and aspirin)? List all. l) 2) 3) 4) 5) 6) 140 CRITERIA FOR AUDIOLOGICAL AND OTOLOGICAL NORMALCY Normal bilateral otoscopy. No excess cerumen. Pure tone air and bone conduction thresholds (dB HTL) no greater than 10 dB (re: ANSI 83.6-1969) at test ~frequencies between 250 Hz and 6000 Hz and no greater than 15 dB at 8000 Hz. Type A tympanograms bilaterally. Acoustic reflex thresholds within a normal level of 70-90 HTL and 60-90 SL at 500 Hz, 1000 Hz and 2000 Hz. Absence of acoustic reflex decay at 500 Hz and 1000 Hz, bilaterally. No reported history of otologic surgery, familial hearing loss, or current URI'S, vertigo, tinnitus or hearing loss. Confirming data shall be obtained within 2 days of experimental testing. SUBJECT SCREENING FORM SUBJECT NAME 141 SCREENING DATE AGE PHONE SEX EXAMINER AUDIOMETER AIR AND BONE CONDUCTION THRESHOLDS: dB HTL 250 500 1000 2000 4000 8000 AC Right Ear ___ ___ ____ ____ ____ .____ Left Ear ___ ___ ____ ____ ____ ____ BC Right Ear ____ ___ _____ ____ ____ Left Ear ___ ___ ____ ____ .____ OTOSCOPIC EXAMINATION TYMPANOMETRIC RESULTS: SEE ATTACHED TYMPANOGRAM .BRIDGE ACOUSTIC REFLEX THRESHOLDS: 500 Right Ear ____ Left Ear ___ ACOUSTIC REFLEX DECAY: 500 Right Ear I___ Left Ear STATEMENT OF PURPOSE READ INFORMED CONSENT RELEASE FORM SIGNED 1000 1000 2000 TRAINING PROGRAM COMPLETED PRE-TEST ADMINISTERED 142 SUBJECT INSTRUCTIONS You will be listening to different sets of noises at different pitches. These sounds will be very faint. AS soon as you hear the noises, press the button. As soon as the noises go away, release the button. You must listen very carefully. Please keep your eyes focused on the lighted box in front of you. Your instructions will be on this box. Keep your head and eyes facing this box. If the clear light goes on, this indicates that your head is out of position. Place your head back into position until the clear light goes off, and the test will continue. There is an intercom system in this lab, so you may communicate with the experimenter. Do you have any questions? APPENDIX B RUN PROTOCOL 2143 RUN PROTOCOL SUBJECT o: __"_N _ GROUP: A a C D SUBJECT CODE SUBJECT NAME PHONE SCREENING DATE TESTING DATE STATEMENT OF PURPOSE READ INFORMED CONSENT RELEASE FORM SIGNED CALIBRATION: (1000 Hz pure tone) TAPE RECORDER OUTPUT LEVEL (0 V0) LINE AMPLIFIER OUTPUT LEVEL (-4 VU) AUDIOMETER INPUT LEVEL (‘3 V0) AUDIOMETER RECORDING ATTENUATOR PEN SET (100 dB) POWER AMPLIFIER OUTPUT VOLTAGES (White noise) RIGHT 500 mvolts LEFT 500 mvolts SUBJECT REPORTED NO EXPOSURE TO EXCESSIVE NOISE IN LAST 24 HOURS____ SUBJECT ACCLAMATED TO TEST ENVIRONMENT FOR 5 MINUTES RESPONSE TRAINING SESSION COMPLETED EXCURSION WIDTH ON TRAINING SESSION SUBJECT DISMISSED? UNOCCLUDED WHITE NOISE THRESHOLD TRAINING PROGRAM ADMINISTERED PRE-TEST ADMINISTERED FORM PRE-TEST SCORE POST-TEST ADMINISTERED FORM POST-TEST SCORE HEARING PROTECTOR DEVICE FIT: EXPERIMENTER FIT SUBJECT PIT DEVICE MANIPULATED Table 1. Random Schedule of Test Signal Presentation. MeaShre Randomization Schedule Unoccluded Trial 1 __. ___ ___ ___ Occluded Trial 1 Unoccluded Trial 2 Occluded Trial 2 . Unoccluded Trial 3 ___ Occluded Trial 3 1J44 Table 2. Threshold Data. Test Frequency Measure I25 250 500 1000 2000 3150 4000 6300 8000 Unoccluded Trial 1 Occluded Trial 1 Unoccluded Trial 2 Occluded Trial 2 Unoccluded Trial 3 Occluded Trial 3 Y Unoccluded SD i Occluded SD Table 3. Attenuation Data. Test Frequency Measure - 125 250 500 1000 2000 3150 4000 6300 8000 Trial 1 Trial 2 Trial 3 — i SD DATA FILED IN SUBJECTS FOLDER APPENDIX C TRAINING TESTS l) 3) 145 PRE-TEST FORM A Tears are caused by: a) b) C) d) e) irritation to the eye anxiety or emotion infection all of the above none of the above Cataracts affect: a) b) c) d) e) The a) b) c) d) e) the iris the retina the conjunctiva the lens none of the above lacrimal apparatus is responsible for: cleansing and lubrication production of visual purple production of vitreous fluid focusing none of the above Earplugs should be inserted in the ear by: a) b) c) d) e) reaching under the chin, pulling the ear down and back and inserting the plug reaching behind the head, pulling the ear back, and inserting the plug reaching over the head, pulling the ear up and back and inserting the plug reaching in front of the head, pulling the ear forward, and inserting the plug none of the above 5) 6) 7) 8) 146 FORM A A person with a permanent hearing loss: a) b) C) d) e) will generally have problems with both loudness and clarity of sounds will only have problems with loudness of sounds will only have problems with clarity of sounds generally will not have problems unless the hearing loss is profound none of the above Sound is composed of: a) b) c) d) e) time and mass minutes and weight frequency and intensity hearing loss none of the above Which of the following would be the least intense? a) b) C) d) e) rustle of a leaf jet aircraft drill press traffic none of the above Correction of visual impairment may involve: a) b) corrective lenses surgery medication all of the above none of the above 9) 10) ll) 12) 147 FORM A With regard to hearing protection devices: a) b) C) d) e) earplugs are the best type earmuffs are the best type earplugs are best used in recreational settings but earmuffs are best used in industrial settings a wide variety of hearing protectors exist none of the above Nystagmus is associated with: a) b) C) d) e) vision only balance only vision and balance vision, balance and hearing none of the above Which of the following statements is true? a) b) C) d) e) The a) b) d) e) all industrial work settings have excessive noise only industrial work settings in Chicago, New York and Los Angeles have excessive noise each and every industrial work setting must be measured for excessive noise levels employees of industrial work settings must provide their own hearing protection none of the above hearing cells are located in the: ear canal cochlea eardrum middle ear none of the above 13) 14) 15) 16) 148 FORM A Which of the following statements is true? a) b) C) d) e) The a) b) C) d) e) the cornea exists in the outer layer of the eye the cornea exists in the middle layer of the eye the iris exists in the outer layer of the eye the retina exists in the middle layer of the eye none of the above retina: translates light waves into neural impulses gives color to the eye is associated with the pupil produces tears none of the above Hearing loss could affect: a) b) C) d) e) a persons family life a persons social life a persons professional life all of the above none of the above Which of the following statements is false? a) b) C) d) e) hearing protection devices prevent hearing loss noisy machinery can be provided with sound insulating enclosures hearing loss from noise exposure is permanent many types of hearing protection devices exist none of the above 17) 18) 19) 20) 149 Soundwaves first reach: a) b) c) d) e) the brain the cochlea the eardrum the middle ear none of the above Blindness may cause: a) b) c) d) e) The a) b) c) d) e) The a) b) C) d) digestive disorders feelings of loss and anxiety loss of balance all of the above none of the above sense organ associated with vision is: the mouth the eye the ear the brain none of the above 3 bones of the ear are located in the: outer ear middle ear inner ear eardrum none of the above FORM A 150 FORM A 21) Hearing loss is measured: a) in a loss of decibels b) in errors of understanding speech c) as a % d) in all of the above e) none of the above 22) Which of the following statements is false? a) excessive noise may increase a persons blood pressure b) excessive noise may cause irritability c) excessive noise causes hearing 1055 d) excessive noise may cause digestive disorders e) none of the above 23) The photosensitive system of the retina is/are: a) the lacrimal apparatus b) the conjunctiva c) the cornea d) the rods and cones e) none of the above 24) Which of the following statements is true? a) the rods are sensitive to bright light b) the rods are sensitive to dim.light c) the cones are sensitive to dim light d) the cones are equally sensitive to bright and dim light e) none of the above 25) 26) 27) 28) The a) b) C) d) e) The a) b) c) d) e) 151 FORM A purpose of a hearing protection device is: to decrease the level of sound reaching the inner ear to restore damaged hearing cells to prevent ear infections to improve a persons hearing none of the above ability to communicate is dependent upon: hearing alone speech alone hearing and speech high IQ none of the above Myopia is synonomous with: a) b) c) d) e) conjunctivitis farsightedness cataracts astigmatism none of the above An effective hearing protector is: a) b) c) d) e) one which is comfortable one which is comfortable and is fit properly one which is disposable one which lasts a long time none of the above 29) 30) 31) 32) 152 FORM A The portion of the eye which gives color is: a) b) C) d) e) the retina the iris the conjunctiva the cornea none of the above Which of the following statements is true? a) b) c) d) e) hearing conservation programs consist only of training programs hearing conservation programs are required by federal law hearing conservation programs consist only of hearing protection device utilization hearing conservation programs are not required by federal law none of the above Harmful noise may be found in/on: a) golf courses b) steel stamping plants c) churches or synagogues d) grocery stores e) none of the above Each eye contains million rods: a) l b) 10 c) 20 d) 130 e) none of the above 33). 34) 35) 36) 153 The organ of hearing is/are the: a) b) c) d) e) cochlea ossicles eardrum middle ear none of the above Which of the following is true? a) b) C) d) e) Each eye contains a) B) c) d) e) sight is more important than taste sight is more important than hearing hearing is more important than taste taste is more important than hearing none of the above million cones: 6 8 10 100 none of the above Conjunctivitis is: a) b) C) d) e) progressive blindness acute blindness inflammation of the conjunctiva inflammation of the cornea none of the above FORM A 37) 38) 39) 40) The a) b) c) d) e) The a) b) c) d) e) 154 FORM A darkened portion at the center of the eye is the: cornea retina iris lens none of the above humors of the eye refer to: the lens the rods and cones the fluids of the eye the iris none of the above Which of the following statements is false? a) b) C) d) e) The a) b) c) d) e) hearing cells send signals to the brain damaged hearing cells are easily repaired or restored each hearing cell reacts to a different frequency or pitch hearing cells react to sound waves none of the above principle organ of sight is: the lens . the cornea the retina the conjunctiva none of the above 155 PRE-TEST FORM B l) Cataracts affect: a) the iris b) the retina c) the conjunctiva d) the lens e) none of the above 2) Which of the following statements is true? a) a person with a hearing loss does not need to wear hearing protectors b) persons over the age of 50 do not need to wear hearing protectors c) persons who have normal hearing do not need to wear hearing protectors d) once a person has gotten used to noise, hearing pro- tectors are no longer needed e) none of the above 3) The portion of the eye which gives color is: a) the retina b) the iris c) the conjunctiva d) the cornea e) none of the above 4) Nystagmus is associated with: a) vision only b) balance only c) vision and balance d) vision, balance and hearing e) none of the above 5) Hearing a) b) c) d) e) 6) Each a) b) c) d) e) 7) Which of a) b) C) d) e) is is is is 156 FORM B loss from noise exposure: intermittent always temporary usually progressive aggrevated by ear infections none of the above eye contains 6 8 10 100 million cones: none of the above the following statements is true? the cornea exists in the outer layer of the eye the cornea exists in the middle layer of the eye the iris exists in the outer layer of the eye the retina exists in the middle layer of the eye none of the above 8) The darkened portion at the center of the eye is the: a) b) C) d) e) cornea retina iris lens none of the above 157 FORM B 9) The humors of the eye refer to: a) b) C) d) e) the lens the rods and cones the fluids of the eye the iris none of the above 10) Noise may best be defined as: a) b). C) d) e) unwanted sound pleasant sound soundwaves pressure waves none of the above 11) Hearing protection devices: a) b) C) d) e) should rarely be worn should be worn wherever excessive noise is found (industrial and recreational settings) should not be worn until a hearing loss exists should only be worn in industrial field environments none of the above 12) Persons working in noisy industrial settings may: a) b) C) d) e) get used to excessive noise and consequently not acquire a hearing loss not utilize earmuffs because earplugs are known to work better choose to wear or not to wear hearing protection devices experience adverse effects from noise exposure other than hearing loss none of the above l3) 14) 15) 16) 158 FORM B Myopia is synonomous with: a) b) C) d) e) conjunctivitis farsightedness cataracts astigmatism none of the above Which of the following statements is true? a) b) C) d) e) hearing cells can be destroyed by noise and once destroyed cannot be restored hearing cells are not destroyed by noise hearing cells may be destroyed from noise exposure but can be restored with a hearing aid current surgical techniques can restore damaged hearing cells none of the above Which of the following is true: a) b) C) d) e) sight is more important than taste sight is more important than hearing hearing is more important than taste taste is more important than hearing none of the above Tears are caused by: a) b) c) d) e) irritation to the eye anxiety or emotion infection all of the above none of the above 159 1?) Hearing conservation programs consist of: a) b) C) d) e) noise level measurements regular hearing tests use of hearing protection devices all of the above none of the above 18) The lacrimal apparatus is responsible for: a) b) c) d) e) cleansing and lubrication production of visual purple production of vitreous fluid focusing none of the above 19) Excessive noise may: a) b) C) d) e) cause hearing loss cause insomnia cause anxiety cause all of the above none of the above 20) Hearing damage from noise exposure is: a) b) c) d) e) cured with surgery cured with a hearing aid cured with rehabilitation cured with hearing protection devices none of the above FORM B 160 21) Correction of visual impairment may involve: a) b) C) d) e) corrective lenses surgery medication all of the above none of the above 22) The sense organ associated with vision is: a) b) c) d) e) the mouth the eye the ear the brain none of the above 23) Industrial work environments typically: a) b) c) d) e) are quiet places have few sources of excessive noise have many sources of excessive noise have sound treated work stations .none of the above 24) The ability to hear is important: a) b) C) d) e) for communication for enjoyment of our environment to alert or warn us of danger all of the above none of the above FORM B 161 FORM B 25) Of the following which environment is most likely not to contain excessive noise levels? a) b) C) d) e) nightclubs rifle ranges steel stamping plants airports none of the above 26) Blindness may cause: a) b) c) d) e) digestive disorders feelings of loss and anxiety loss of balance all of the above none of the above 27) After earplugs are inserted, you should: a) b) c) d) e) do chewing motions to make sure they are in place be careful not to move your jaw anymore than necessary talk and chew normally avoid adjusting them in any way none of the above 28) The principle organ of sight is: a) b) c) d) e) the lens the cornea the retina the conjunctiva none of the above 162 FORM B 29) The cochlea is located in: a) b) c) d) e) the outer ear the middle ear the inner ear the eardrum none of the above 30) Which of the following statements is true: a) b) C) d) e) 31) The a) b) c) d) e) the rods are sensitive to bright light the rods are sensitive to dim light the cones are sensitive to dim light the cones are equally sensitive to bright and dim light none of the above retina: translates light waves into neural impulses gives color to the eye is associated with the pupil produces tears none of the above 32) Conjunctivitis is: a) b) C) d) e) progressive blindness acute blindness inflammation of the conjunctiva inflammation of the cornea none of the above 163 FORM B 33) Which of the following is not a consideration in choosing hearing protection? ‘ a) b) C) d) e) comfort attenuation hygiene type of none of 34) Which of the a) b) C) d) e) hearing hearing hearing hearing none of device the above following is false? loss generally has effects on personality loss may cause loss of self confidence loss may cause insecurity loss may cause anxiety the above 35) The human ear: a) can block out damageable noise b) reacts equally to all sounds c) is unaffected by excessive noise d) transforms sound waves into auditory impressions or messages e) none of the above 36) There are approximately hearing cells in each ear: a) 1 million b) 350 c) 10,000 d) 30,000 e) none of the above 164 FORM B 37) Which of the following is true? a) b) C) d) e) 38) Each a) b) c) d) e) earplugs provide the best protection from harmful noise earmuffs provide the worst protection from harmful noise earmuffs plus earplugs provide the best protection from harmful noise earmuffs provide the best protection from harmful noise none of the above eye contains million rods 1 10 20 130 none of the above 39) The simplest way to protect people from excessive industrial noise is: a) b) C) d) e) to provide them with rehabilitation to provide them with hearing protection devices to provide them with a hearing aid to fire them none of the above 40) The photosensitive system of the retina is/are: a) b) d) e) the lacrimal apparatus the conjunctiva the cornea the rods and cones none of the above 165 POST-TEST FORM A 1) Which of the following is false? a) b) c) d) e) hearing loss generally has effects on a personality hearing loss may cause loss of self confidence hearing loss may cause insecurity hearing loss may cause anxiety none of the above 2) Hearing damage from noise exposure is: a) b) C) d) e) cured with surgery cured with a hearing aid cured with rehabilitation cured with hearing protection devices none of the above 3) Which of the following is true? a) b) C) d) e) earplugs provide the best protection from harmful noise earmuffs provide the worst protection from harmful noise earmuffs plus earplugs provide the best protection from harmful noise earmuffs provide the best protection from harmful noise none of the above 4) Which of the following is not a consideration in choosing hearing protection? a) comfort b) attenuation c) hygiene d) type of device e) none of the above 166 FORM A 5) The simplest way to protect people from excessive industrial noise is: a) b) C) d) e) to provide them with rehabilitation to to to fire none of provide them with hearing protection devices provide them with a hearing aid them the above 6) Which of the following statements is true: a) b) c) d) e) hearing cells can be destroyed by noise and once destroyed cannot be restored hearing hearing but can current hearing none of 7) The cochlea is a) b) c) d) e) cells are not destroyed by excessive noise cells may be destroyed from noise exposure be restored with a hearing aid surgical techniques can restore damaged cells the above located in: the outer ear the middle ear the inner ear the eardrum none of the above 8) Hearing conservation programs consist of: a) b) C) d) e) noise level measurements regular hearing tests use of hearing protection devices all of the above none of the above 167 FORM A 9) After earplugs are inserted, you should: a) b) C) d) e) do chewing motions to make sure they are in place be careful not to move the jaw anymore than necessary talk and chew normally avoid adjusting them in any way none of the above 10) Which of the following statements is true? a) b) c) d) e) a person with a hearing loss does not need to wear hearing protectors persons over the age of 50 do not need to wear hearing protectors persons who have normal hearing do not need to wear hearing protectors once a person has gotten used to noise, hearing protectors are no longer needed none of the above 11) Excessive noise may: a) b) C) d) e) cause hearing loss cause insomnia cause anxiety cause all of the above none of the above 12) Noise may best be defined as: a) b) c) d) e) unwanted sound pleasant sound soundwaves pressure waves none of the above 13) 168 FORM A Of the following, which environment is most likely not to contain excessive noise levels: a) b) C) d) e) 14) a) b) C) d) e) 15) a) b) C) d) e) 16) a) b) c) d) e) There are approximately nightclubs rifle ranges steel stamping plants airports none of the above Hearing loss from noise exposure: is intermittent is always temporary is usually progressive is aggrevated by ear infections none of the above hearing cells in each ear: 1 million 350 10,000 30,000 none of the above Hearing protection devices: should rarely be worn should be worn wherever excessive noise is found (industrial and recreational) should not be worn until a hearing loss exists should only be worn in industrial field environments none of the above 169 FORM A 17) The human ear: a) b) C) d) e) can block out damageable noise reacts equally to all sounds is unaffected by excessive noise transforms sound waves into auditory impressions or messages none of the above 18) Persons working in noisy industrial settings may: a) b) C) d) e) get used to excessive noise and consequently not acquire a hearing loss not utilize earmuffs because earplugs are known to work better choose to wear or not to wear hearing protection devices may experience adverse effects from noise exposure other than hearing loss none of the above 19) The ability to hear is important: a) b) C) d) e) for communication for enjoyment of our environment to alert or warn us of danger all of the above none of the above 20) Industrial work environments typically: a) b) c) d) e) are quiet places have few sources of excessive noise have many sources of excessive noise have sound treated work stations none of the above 170 POST-TEST FORM B 1) With regard to hearing protection devices: a) b) C) d) e) earplugs earmuffs earplugs earmuffs are are are are the best type the best type best used in recreational settings but best used in industrial settings a wide variety of hearing protectors exist none of the above 2) The organ of hearing is/are the: a) b) C) d) e) cochlea ossicles eardrum middle ear none of the above 3) An effective hearing protector is: a) b) c) d) e) one which is comfortable one which is comfortable and is fit properly one which is disposable one which lasts a long time none of the above 4) Which of the following is false? a) b) c) d) e) hearing protection devices prevent hearing loss noisy machinery can be provided with sound insulating enclosures hearing loss from noise exposure is permanent many types of hearing protection devices exist none of the above 171 FORM B 5) Which of the following would be the least intense? a) b) C) d) e) rustle of a leaf jet aircraft drill press traffic none of the above 6) Sound is composed of: a) b) C) d) 6) time and mass minutes and weight frequency and intensity hearing loss none of the above 7) Which of the following statements is true? a) b) c) d) e) all industrial work settings have excessive noise only industrial work settings in Chicago, New York and Los Angeles have excessive noise each and every industrial work setting must be measured for excessive noise levels employees of industrial work settings must provide their own hearing protection none of the above 8) The 3 bones of the ear are located in the a) b) C) d) e) outer ear middle ear inner ear eardrum none of the above 172 FORM B 9) Which of the following statements is false? a) b) c) d) e) hearing cells send signals to the brain damaged hearing cells are easily repaired or restored each hearing cell reacts to a different frequency or pitch hearing cells react to soundwaves none of the above 10) Hearing loss could affect: a) b) c) d) e) a persons family life a persons social life a persons professional life all of the above none of the above 11) Soundwaves first reach: a) b) C) d) e) the brain the cochlea the eardrum the middle ear none of the above 12) Which of the following statements is false? a) b) C) d) e) excessive noise may increase a persons blood pressure excessive noise may cause irritability excessive noise causes hearing loss excessive noise may cause digestive disorders none of the above 173 FORM B 13) Which of the following statements is true? a) hearing conservation programs consist only of training programs b) hearing conservation programs are required by federal law c) hearing conservation programs consist only of hearing protection device utilization d) hearing conservation programs are not required by federal law e) none of the above 14) Harmful noise may be found in/on: a) golf courses b) steel stamping plants c) churches or synagogues d) grocery stores e) none of the above 15) The purpose of a hearing protection device is: a) to decrease the level of sound reaching the inner ear b) to restore damaged hearing cells c) to prevent ear infections d) to improve a persons hearing e) none of the above 16) ‘The ability to communicate is dependent upon: a) hearing alone b) speech alone. c) hearing and speech d) high IQ e) none of the above 17) a) b) C) d) e) 18) a) b) C) d) e) 19) a) b) C) d) e) 20) a) b) C) d) e) 174 FORM B A person with a permanent hearing loss: will generally have problems with both loudness and clarity of sounds will only have problems with loudness of sounds will only have problems with clarity of sounds generally will not have problems unless the hearing loss is profound none of the above The hearing cells are located in the: ear canal cochlea eardrum middle ear none of the above Hearing loss is measured: in a loss of decibels in errors of understanding speech as a % all of the above none of the above Earplugs should be inserted in the ear by: reaching under thecfifimn.pulling the ear down and back and inserting the plug reaching behind the head, pulling the ear back, and inserting the plug reaching over the head, pulling the ear up and back and inserting the plug reaching in front of the head, pulling the ear forward, and inserting the plug none of the above l) 2) 3) 4) 5) 6) 7) 8) 9) 10) ll) 12) l3) 14) 15) 16) 17) 18) 19) 20) 175 PRE-TEST ANSWER SHEET FORM 21) 22) 23) 24) 25) 26) 27) 28) 29) 30) 31) 32) 33) 34) 35) 36) 37) 38) 39) 40) l) 2) 3) 4) 5) 6) 7) 8) 9) 10) ll) 12) l3) 14) 15) 16) 17) l8) 19) 20) 176 POST-TEST ANSWER SHEET FORM l) 2) 3) 4) 5) 6) 7) 8) 9) 10) ll) 12) 13) 14) 15) l6) 17) 18) 19) 20) 21) 22) 23) 24) 25) 26) O” O‘ O‘ (D D.) 177 ANSWER KEY PRE-TEST FORM A 27) 28) 29) 30) 31) 32) 33) 34) 35) 36) 37) 38) 39) 40) U‘ 0‘ 0‘ U‘ (D (D l) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15) 16) 17) 18) 19) 20) 21) 22) 23) 24) 25) 26) 178 ANSWER KEY PRE-TEST FORM B 27) 28) 29) 30) 31) 32) 33) 34) 35) 36) 37) 38) 39) 40) 1) 2) 3) 4) 5) s) 7) a) 9) 10) 11) 12) 13) 14) 15) 1s) 17) 1a) 19) 20) Q: D.- Q 0' O.) 0 O 179 ANSWER KEY POST-TEST FORM A 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) ll) 12) 13) 14) 15) l6) 17) 18) 19) 20) (D Q: U‘ O” O U‘ U‘ (D 9) 180 ANSWER KEY POST-TEST FORM B APPENDIX D SPEAKER WIRING DIAGRAMS 181 RIGHT CHANNEL b gin‘io‘lifier J’ Figure 0.1. Speaker wiring diagram for the right channel. 182 7 Q :ch 1 L (Doooooooo ocpooocpocpc L APPENDIX E INSTRUMENTATION, METHODS AND RESULTS OF REVERBERATION TIME MEASUREMENTS 183 Appendixli- Instrumentation, Methods and Results of Reverberation Time Measurements. Instrumentation and Methods Figure:E-l presents the block diagram of the instrumentation used for reverberation time measurements. A free- field microphone (Bruel & Kjaer Model 4145) was routed to a pre-amplifier (Bruel & Kjaer Model 2619) and powered by a microphone power supply (Bruel & Kjaer Model 2804). The output was routed to a measurement amplifier (Bruel & Kjaer Model 2607) and displayed on a graphic level recorder (Bruel & Kjaer Model 2305). The paper speed and writing speed of the graphic level recorder were set at 100 mm/second and 250 mm/second, respectively. The microphone was calibrated with a level calibrator (GenRad Model 1986). Prior to recording reverberation times, the rise/fall time of the electronic switch was set to a minimum of 100 microseconds and the attenuator of the Bekesy audiometer was set to the 100 dB position. The tape recorder was put into playback mode and reverberation time was measured for each of the nine test bands and the white noise. Measurements were made without a subject present as Specified in ANSI 83.19-1974. Results Table 3.1 presents the measured reverberation times of the test capsule. ANSI 53.19-1974 specifies that reverberation times should be between 500 and 1600 msec. As can be seen .mucmEUusmmmE mafia cofiumumnuw>mu MOM coflumucmesuumcfi mo Emummwo xoon .H.m musmflm 184 Hmouoomm _WUHHMHQE¢ hammsm Hm>mq acme , Hm30m .mEmlmum LVIIIIIIAHMW .rmusmomz 0H2 oflnmmuo 185 Table E.l. Measured reverberation times of test capsule. Signal Reverberation Time White Noise 320 msec 125 Hz 300 msec 250 Hz 300 msec 500 Hz 300 msec 1000 Hz 390 msec 2000 Hz 460 msec 3150 Hz 450 msec 4000 Hz 420 msec 6300 Hz 400 msec 8000 Hz 360 msec 186 from Table E.l, measured reverberation times are shorter than ANSI 83.19-1974 specifications. Although reverberation times are not within specifications, this result is not expected to significantly affect the outcome of this study. APPENDIX F INSTRUMENTATION, METHODS AND RESULTS OF SPATIAL UNIFORMITY OF ACOUSTIC FIELD MEASUREMENTS 187 Appendix F. Instrumentation, Methods and Results of Spatial Uniformity of Acoustic Field Measurements. Instrumentation and Methods The spatial uniformity of the acoustic field was defined in terms of (1) field diffussion and (2) random incidence field measurements. Figure F.l presents the block diagram of the instrumentation used for the field diffussion and random indicidence field measurements. A pressure microphone (Bruel & Kjaer Model 4144) was routed to a pre-amplifier (Bruel & Kjaer Model 2619) and powered by a microphone power supply (Bruel & Kjaer Model 2804). The output was routed to a measurement amplifier (Bruel & Kjaer Model 2607). The micro- phone was calibrated with a level calibrator (GenRad Model 1986). Measurements were made at six positions around the center position of a subject's head (subject absent) for the field diffussion measurements (as specified in ANSI 83.19-1974). For all field diffussion measurements, the attenuator on the Bekesy audiometer was set at the 90 dB position. The 20 dB pad was engaged and disengaged at various measurement points (noted in results section). The reference SPL for white noise was 80 dB. Readings on the power amplifier for the right and left channels were -40 dB and -30 dB, respectively. The tape recorder was set into playback mode and SPL readings were obtained on the measurement amplifier for each of the nine test bands and the white noise bands preceeding and following the noise bands for each of the six positions specified in 188 new mamaamsa COME Imusmmwz .mucmemusmmme onHw menopause Eoosmu :0ammsmwfio pamww How cofiumucmesuumsw mo Emummao onHm mammom bosom UH: A .mEmumHm .H.m musmnm @ 189 ANSI 83.19-1974. Random incidence field measurements were also obtained. Instrumentation and methods were identical to the field diffussion measurements with the exception of microphone position. The microphone was placed in the center of the area representing the position of a subjects head and rotated in three directions: (1) microphone facing ceiling; diaphram parallel to floor; (2) microphone facing front wall; diaphram perpendicular to floor; and (3) microphone facing door; diaphram perpendicular to floor. Results Table FUJ.presents the results of the field diffussion measurements for the six microphone positions.) ANSI $3.19- 1974 specifies that the SPL measured at these six positions must remain within a range of 6 dB for all test bands. It is further specified that the difference in SPL between the left-right positions shall not exceed 2 dB. As can be seen from these results, the measured values are within speci- fications of ANSI 83.19-1974. Table F~2 presents the results of the random incidence field measurements. These results are within specifications of ANSI 83.19-1974. 190 .ommmmcm one mo oma mo o.mm mo m.mm mo m.mm mo o.mm mo o.mm mo o.mm wmfioz mafia: mo m.hoa mo c.5oH mo o.>oa mo m.ooa mo o.hoa mo m.hoa ammwoz been: mo o.mm mo o.aoa mo m.ha mo o.ooa mo c.mm mo o.mm «a: ooom mo m.mm mo c.5m mo m.om mo m.nm mo m.em mo m.>m «um comm mo o.voa mo m.mo~ mo m.moa mo c.moa mo o.moH mo m.voa sum ooov mo o.mca mo m.voa mo m.>oa mo.m.moa mo m.eca mo m.ooa «N: omam mo m.>m mo o.mm mo m.vw mo m.mm mo m.vm mo o.mm um ooom mo o.vm mo m.om mo o.~m mo m.mm mo o.Hm mo o.mm a: coca mo m.mm mo m.~m mo m.mm mo m.am mo m.mm mo o.vm a: com mo m.>m mo m.>m mo o.mm mo o.wm mo c.5m mo m.mm um omm mo m.mm mo o.mm mo mnmm mo 0.5m mo m.mm mo m.hm um mma mo o.mm mo m.¢m mo m.vm mo c.vm mo o.vm mo m.vm mmfloz mafia: m .mom m .mom v .mom m .mom _ N .mom H .mom HUCOAm mmsHm> owuomomz .muasmmn ucmEmuommUE COHmODMMHo onflm .H.m mabme 191 .ommmmcm one mo om « mo m.~m mo c.mm mo c.mm mmfioz mufinz mo c.moH mo m.moH mo m.moa «mmwoz muwzz mo m.mm mo m.mm mo m.mm an: ooom. mo c.mm mo m.mm mo m.~m «u: come mo m.ooa mo m.aoa. mo o.~oa «um ooov mo o.moa mo m.~oa mo m.mOH «N: omam mo o.~m mo o.mm mo m.~m mm ooom mo o.mh mo m.mh mo o.mh um oooa mo o.om mo m.om mo m.om mm com mo o.¢m mo o.mm mo o.vm N: omm mo o.mm mo o.¢m mo o.vm um mmH mo m.am mo m.Hm mo m.Hm mmfloz mafia: noon mcflomm oflz HHOE accum mcwomh ca: mafiafiwo mcflomm ca: Hmcmflm mosam> consume: .muasmmu DCUEmHsmmmE onHm mocmoflocfi Eoocmm .~.E manna APPENDIX G COMPUTER PROGRAM FOR COMPUTATION OF NRR 1:110 100 REM 110 : 120 REM 130 REM 140 REM 150 : 160 REM 170 ' 180 REM 190 REM 200 REM 210 REM 220 REM 230 REM 240 REM 250 260 REM 270 REM 280 REM 290 REM 300 REM 310 REM 320 : 330 : 340 REM 350 360 REM 370 REM 380 REM 390 REM 400 REM 410 REM 420 REM 430 REM 440 REM 450 REM 460 REM 470 REM 480 REM 490 REM 500 REM 510 REM 520 REM 530 REM 540 REM 550 REM 560 REM 570 REM 580 REM S90 REM 600 REM 610 REM 620 REM 630 I 1000 : 1010 REM 1020 : 1030 TEXT 1040 HOME 1J92 EPA NRR LALLULATOR AN APPLESOFT BASIL PROGRAM MICHAEL R. CHIAL 07/08/83 ABSTRACT THIS PROGRAM ASSUMES THE METHOD SPECIFIED BY THE EPA (1979) [FEDERAL REGISTER, VOL 42, NO. 190, 40 CFR PART 211, PP. 56139-56147]. SPECIFICALLY. A HYPOTHETICAL NOISE IS ASSUMED TO HAVE EQUAL LEVELS AT OCTAVE BANDS (OB) BETWEEN 125 HZ AND 8000 H2 (I.E., PINK NOISE IS ASSUMED). FURTHER, A- AND C-NEIGHTINGS ARE ASSUMED IN THE MEASUREMENT OF ANY SUCH NOISE. PROGRAM PROMPTS FOR MEANS AND STANDARD DEVIATIONS OF REAL-EAR ATTENUATIONS TAKEN FROM A GROUP OF LISTENERS. PER EPA. IT IS ASSUMED THAT AT LEAST 10 TRAINED. NORMAL- HEARING PERSONS HAVE BEEN TESTED VIA METHOD OF EXPERIMENTER FIT OF HEARING PROTECTORS PROGRAM OUTPUTS COMPUTED NOISE REDUCTION RATING (NRR) FOR THE HEARING PROTECTOR SPECIFIED. VARIABLE LIST BM(I) I OCTAVE BAND MEAN REAT FOR [=9 BANDS CENTERED AT THE FOLLOWING FREQUECIES: BM(I) I 125 HZ BM(2) I 250 HZ BM(3) I 500 HZ BM(4) I I KHZ BM(5) I 2 KHZ BM(6) I 3 KHZ BM(7) I 4 KHZ BM(B) I 6 KHZ BM(9) I B KHZ OCTAVE BAND STAND. DEVIATION REAT FOR 1-9 BANDS As NOTED ABOVE. 8A(1) a A-uEIGHTED DB LEVELS FOR UNPROTECTED EAR, PER EPA STEP 4. THESE CONSTANTS APPEAR AT LINE 10210. at . OVERALL C-UEIGHTED LEVEL OF HYPOTHETICAL NOISE: 108 03(0), A CONSTANT. THIS Is EPA STEP 3. INTERMEDIATE VALUES PER EPA STEP 7 88(1) A(I) I LS I LOG SUM OF PROTECTED LEVELS PER EPA STEP 8. OS I DATE STRING KS I GROUP CODE STRING NO I HEARING PROTECTOR NAME STRING Re I USER RESPONSE STRING FN LG I BASE-10 LOG FUNCTION PIS I OUTPUT DEVICE DRIVER ISSUE GREETING, GET HPD & TEST GROUP ID INFORMATION NORMAL CLEAR : RESTORE 193 I050 DEF FN LGIX) = LOG (X) / LOO 110’ 1060 HTAB (9): PRINT “E-EPA NRR CALCULATORE-": PRINT : HTAB 7': PRINT 'BY MICHAEL R. LHIAL. PH.D.‘: HTAB t4): PRINT ..... _________________ ‘1070 POKE 34,4: REM PROTECT TOP 4 LINES 1080 HOME : VTA8 (I0) 1090 INPUT ‘ENTER NAME OF PROTECTOR: ';N$ 1100 INPUT “ENTER EXP. OR GROUP CODE: ';KS 1110 INPUT 'ENTER DATE (MM/DD/YY): ‘;Ds 1120 PRINT 1130 INPUT 'OUTPUT TO PRINTER 0R SCREEN (3 0R P)? 'gRs 1140 IF Rt < ) '8‘ AND Rs ( > 'P' THEN GOTO 1130 1150 IF Rt I '8‘ THEN SS I '0': GOTO 1180 1160 IF R! I ‘P' THEN INPUT 'PRINTER SLOT (NORMALLY I)? ”:83 1170 IF VAL (SS) ( 0 OR VAL (SS) > 7 THEN GOTO 1160 1180 P1. I 'PRI' o S! 1190 : 5000 REM GET MEAN ATTENUATION DATA (EPA STEP 5) 5010 : 5020 FOR I I I TO 9 5030 READ 8$(1) 5040 NEXT I 5050 HOME 5060 PRINT : PRINT 'REAL-EAR ATTENUATION DATA ENTRY': PRINT 5070 PRINT 'IF THERE ARE NO DATA FOR 3 KHZ & 6 KHZ.‘: PRINT “ENTER 0 (Z ERO).' 5080 PRINT I PRINT 5090 FOR I I I TO 9 5100 PRINT ‘ENTER MEAN FOR ';8$(I); 5110 HTAB 30: INPUT ";8M(I) 5120 NEXT I 5130 : 5140 REM GET ARITH. MEANS FOR 3 & 4 KHZ AND FOR 6 & 8 KHZ 5150 : 5160 IF 8M(6) I 0 THEN GOTO 5190 5170 8M(7) I (8M(6) P 8M(7)) / 2 5180 8M(6) I 0 5190 IF BM(8) I 0 THEN GOTO 5220 5200 8M(9) I (8M(8) 9 8M(9)) / 2 5210 8M(8) I 0 5220 : 6000 REM GET STD. DEV. DATA & DOUBLE VALUES (EPA STEP 6) 6010 : 6020 HOME 6030 PRINT : PRINT 'REAL-EAR ATTENUATION DATA ENTRY': PRINT 6040 PRINT 'IF THERE ARE NO DATA FOR 3 KHZ & 6 KHZ,': PRINT “ENTER 0 (Z ERO).' 6050 PRINT : PRINT 6060 FOR I I I TO 9 6070 PRINT 'ENTER STD. DEV. FOR ';8$(I): 6080 HTAB 30: INPUT ";88(1) 6090 88(1) I 88(1) I 2 6100 NEXT I 6110 : 6120 uREM GET ARITH. MEANS FOR 3 & 4 KHZ AND FOR 6 & 8 KHZ 6130 8 6140 IF 88(6) I 0 THEN GOTO 6170 6150 88(7) 5 (88(6) + 85(7)) /'2 6160 85(6) I 0 6170 IF 88(8) I 0 THEN GOTO 6200 6180 83(9) I (88(8) 0 88(9)) ’ 2 6190 85(8) I 0 6200 : - 7000 REM COMPUTE A-NEIGHTED 08 SOUND LEVELS IEPA STEP 7: 7010 REM STEP 4-STEP 5+STEP 6) 7020 : - 7030 FOR I I I TO 9 7040 READ 8A(I) 7050 IF BA(I) I 0 THEN GOTO 7070 7060 A(I) I 8A(1) - 8M(I) * 88(1) 7070 NEXT 1 7080 I 8000 REM COMPUTE LOG SUM OF 08 SOUND LEVELS FROM STEP 7 (EPA STEP 8) 8010 I 8020 LS I 0 8030 FOR I I 1 T0 9 8040 IF A(I) I 0 THEN GOTO 8080 8050 Y(I) I A(I) / 10 8060 A(I) I 10 “ Y(I) 8070 LS I LS + A(1) 8080 NEXT 1 3090 LS I 10 I FN LG 8100 I 9000 REM COMPUTE NRR AS STEP 3-STEP 8- 3 D8 (EPA STEP 9) 9010 : 9020 DC I 108 9030 NRR I CC - LS - 3.0 9040 I 10000 REM OUTPUT RESULTS AND PROMPT FOR REPEAT 10010 I 10020 HOME 10030 VTAB (10) 10040 PRINT CHRS (4);P1$ 10045 PRINT I PRINT 10050 PRINT “HEARING PROTECTOR: “: TA8( 25>;N3I PRINT “EXP. OR GROUP CO DEI“: TA8( 25)IK$ 10060 PRINT “DATE OF ANALYSIS: “3 TA8( 25);D$I PRINT I PRINT 10070 NRR I INT (NRR D 10 0 .5) / 10 10080 PRINT “THE COMPUTED NRR ISI“: TA8( 30)INRR;“ 08“ 10090 PRINT “CONSERVATIVE NRR ISI“; TA8( 30)INRR - 73“ 08“ 10100 PRINT I PRINT . 10110 PRINT CHR’ (4);“PRIO“ 10120 INPUT “PRESS RETURN TO CONTINUE “:R‘ 10130 PRINT “DO AGAIN? '3 10140 INPUT “(Y FOR YES, N FOR NO) “IRS 10150 IF R! I “Y“ THEN POKE 34,0: GOTO 1040 10160 IF RS < ) “Y“ AND RS < > “N“ THEN GOTO 10120 10170 I 10180 REM DATA FOR LABELS AND COMPUTATIONS 10190 I 10200 VDATA “125 H2“, “250 H2“, “500 HZ“, “ 1 KHZ“, “ 2 KHZ“, “ 3 KHZ“, “ 4 KHZ“, “ 6 KHZ“, “ 8 KHZ“ 10210 DATA 83.9, 91.4, 96.8, 100.0, 101.2, 0.0, 101.0, 0.0, 98.9 10220 I 10230 REM CLOSING ROUTINE 10240 3. 10250 HOME I VTAB (10) 10260 PRINT “SO LONG!“ 10270 POKE 34,0: REM UNPROTECT TOP 4 LINES 10280 END' 194 APPENDIX H PRE-TEST, POST-TEST AND INFORMATION GAIN SCORES (RAW DATA) 195 Table ILJ” Ere-test, post-test and information gain scores (raw data). Subject Pre-Test Post-Test Information Gain Group A 1 60 80 20 2 70 70 0 3 60 90 30 4 75 100 25 5 90 30 -10 6 55 100 45 7 75 35 10 8 40 85 45 9 90 60 -30 10 55 7s 20 Group B 11 35 55 20 12 70 85 15 13 75 65 +10 14 85 - 95 10 15 75 85 10 16 70 95 25 17 70 6O -10 18 45 75 30 19 75 80 5 20 55 8O 25 Group C 21 80 75 -5 22 75 45 -30 23 75 75 O 24 55 65 10 25 45 65 20 26 70 7 0 27 75 65 -10 28 80 50 -30 29 50 55 -5 30 55 SS “0 196 Table H.l. Continued. Subject Pre-Test Post-Test Information Gain Group D 31 85 80 -5 32 70 70 0 33 80 7S -5 34 60 60 -0 35 75 70 -5 36 60 65 5 37 70 65 -5 38 75 7O -5 39 85 60 -25 4O 55 85 30 APPENDIX I DESCRIPTIVE STATISTICS FOR PRE-TEST, POST-TEST, AND INFORMATION GAIN SCORES 197 Table I51” Descriptive statistics for percent-correct pre-test scores, post-test scores and information gain scores (N = 40). Pre-Test Post-Test Information Gain Variance 187.2 181.8 348.5 Minimum 35 45 -30 Maximum 90 100 45 Range 55 55 75 Skewness .502 .136 ' .073 Kurtosis .384 .459 .164- 198 Table 1.2. Descriptive statistics for percent-correct pre-test scores, post-test scores and information gain scores across groups. Pre-Test Post-Test Information Gain Group A (experimenter fit; training) Variance 256.7 156.9 563.6 Minimum 40 60 -30 Maximum 90 100 45 Range 50 40 75 Skewness .053 .185 .615 Kurtosis .544 .133 .013 Group B (subject fit; training) Variance 241.4 190.3 195.6 Minimum 35 55 -10 Maximum 85 95 30 Range 50 40 40 Skewness 1.033 .377 .553 Kurtosis .247 .912 .703 Group C (experimenter fit; no training) Variance 175-7 107.7 254.4 Minimum 45 45 -30 Maximum 80 75 20 Range 35 30 50 Skewness .489 .295 .612 Kurtosis 1.639 1.068 .019 Group D (subjeet fit; no training) Variance 111.4 66.7 183.6 Minimum 55 60 -25 Maximum 85 85 30 Range 30 25 55 Skewness .227 .524 1.024 Kurtosis 1.154 .288 4.026 APPENDIX J RAW ATTENUATION DATA 199 Table J.1. Raw attenuation data (table values are decibels). Crager Ffsaaen°xlé"2’ Measure 125 250 500 0 O 4000 6300 8000 Subject 1F Trial 1 6.3 8.3 18.6 14.2 19.3 17 7 22.3 33.4 40.1 Trial 2 8.3 12.7 17.5 14.8 29.1 27.5 33.4 34.7 30.5 Trial 3 16.6 14.1 18.8 11.0 30.7 30.4 29.5 32.5 29.7 10.4 11.7 18.3 13.3 26.3 25.2 28.4 33.5 33.4 SD 5.4 3.0 .7 2.0 6.1 6 6 5.6 1.1 5.7 Subject 2M Trial 1 3.2 4.2 13.2 15.1 23.8 26.5 31.9 33.7 33.3 Trial 2 1.7 .2 9.3 18.4 19.3 27.9 24.4 28.2 24.7 Trial 3 6.7 8.5 15.3 24.0 24.5 33.4 36.4 61.5 25.7 X 3.8 4.3 12.6 19.1 22.5 29.2 30.9 41.1 27.7 SD 2.5 4.1 3.0 4.4 2.8 3.6 6.0 17.8 4.8 Subject 3 F Trial 1 14.8 11.2 12.2 22.6 19.7 24.1 36.0 25.8 37.3 Trial 2 10.9 8.4 13.8 21.9 22.8 35.9 29.0 33.9 39.1 Trial 3 4.9 5.6 9.3 12.0 23.5 34.3 34.3 35.5 28.0 X 10.2 8.4 11.7 18.8 22.0 31.4 33.1 31.7 34.8 SD 4.9 2.8 2.2 5.9 2.0 6.4 3.6 5.2 5.9 Subject 4 F Trial 1 11.1 9.5 13.1 16.8 21.9 29.8 36.7 24.5 32.7 Trial 2 16.7 13.9 20.3 19.2 26.2 34.0 37.4 25.8 35.3 Trial 3 12.7 13.4 14.4 17.1 24.7 33.0 30.9 35.3 34.0 X 13.5 12.2 15.9 17.7 24.2 32.2 35.0 28.5 34.0 SD 2.8 2.4 3.8 1.3 2.1 2.1 3.5 5.8 1.3 Subject 5 M Trial 1 1.7 2.2 13.0 16.6 28.6 31.6 33.5 35.6 32.6 Trial 2 9.8 10.3 15.6 23.2 24.7 30.5 38.4 40.5 45.2 Trial 3 12.5 10.9 15.1 20.4 34.6 30.4 30.9 33.8 40.0 X 8.0 7.8 14.5 20.0 29.3 30.8 34.2 36.6 39.2 SD 5.6 4.8 1.3 3.3 4.9 .6 3.8 3.4 6.3 Subject 6 M Trial 1 3.8 5.4 15.8 16.3 26.0 34.3 39.4 42.6 45.7 Trial 2 2.4 10.1 17.9 15.6 23.4 25.6 32.0 37.5 26.3 Trial 3 1.1 9.1 10.7 15.7 22.5 25.3 28.2 26.6 30.1 X 2.4 8.2 14.8 15.8 23.9 28.4 33.2 35.5 34.1 SD 1.2 2.4 3.7 .4 1.8 5.1 5.6 8.1 10.2 Subject 7 M Trial 1 8.0 8.8 18.2 16.9 24.3 34.8 29.9 38.8 35.9 Trial 2 12.5 10.9 10.4 9.7 31.2 23.6 34.1 32.9 22.8 Trial 3 5.1 4.9 13.3 16.3 29.5 36.1 36.6 38.6 34.0 2‘ 8.5 8.2 13.9 14.3 28.3 31.5 33.5 36.7 30.9 SD 3.7 3.0 3.9 3.9 3.5 6.8 3.3 3.3 7.0 Subject 8 F Trial 1 13.3 9.1 15.1 20.7 25.1 32.2 32.8 36.0 28.7 Trial 2 -.4 9.6 11.6 13.4 19.8 31.6 28.0 27.9 28.9 Trial 3 10.2 9.9 9.6 16.7 22.1 23.3 38.5 28.1 24.5 Y 7.7 9.5 12.1 16.9 22.3 29.0 33.1 30.6 27.3 SD 7.1 .4 2.7 3.6 2.6 4.9 5.2 4.6 2.4 20C) Fest Frequency Measure 125 250 500 1000 2000 3150 4000 6300 8000 Subject 9 F' Trial 1 10.6 8.4 14.0 17.4 21.4 33.3 36.0 42.2 35.3 Trial 2 8.9 8.5 14.7 12.9 27.0 31.5 33.5 36.9 40.0 Trial 3 6.0 -1.8 11.9 11.6 21.1 24.9 23.3 24.0 31.0 Y 8.5 5.0 13.5 13.9 23.1 29.9 30.9 34.3 35.4 so 2.3 5.9 1.4 3.0 3.3 4.4 6.7 9.3 4.5 Subject 10 M Trial 1 5.1 8.9 16.4 22.7 35.4 35.3 43.6 41.8 39.4 Trial 2 7.8 1.6 14.8 13.8 20.1 28.1 28.0 31.1 21.6 Trial 3 10.7 9.4 14.8 16.4 38.0 40.0 40.2 39.2 40.0 7.8 6.6 15.3 17.6 31.1 34.4 37.2 37.3 33.6 SD 2.8 4.3 .9 4.5 9.6 5.9 8.2 5.5 10.4 Subject 11 M Trial 1 5.7 -.2 5.5 -4.9 11.2 30.9 118.9 20.8 20.8 Trial 2 1.5 4.8 1.8 .s 3.8 1.3 30.9 4.8 20.6 Trial 5 16.9 -2.8 -3.2 1.1 -3.6 .5 -11.7 -3.7 2.4 x 8.0 .6 1.3 —1.1 3.8 10.9 12.7 7.3 14.6 so 7.9 3.8 4.3 3.3 7.4 17.3 21.9 12.4 10.5 Subject 12 F Trial 1 18.3 9.9 14.4 12.4 30.9 36.4 30.5 57.4 36.0 Trial 2 6.8 7.8 13.2 14.6 24.2 22.3 30.8 33.5 23.5 Trial 3 6.4 10.3 14.8 15.2 28.5 32.8 31.3 32.3 26.1 Y 10.5 9.3 14.1 14.0 27.8 30.5 30.8 34.4 28.5 SD 6.7 1.3 .8 1.4 .3.3 7.3 .4 2.6 6.5 Subject 13 M Trial 1 27.2 10.2 21.0 20.5 31.8 28.1 39.9 41.3 46.9 Trial 2 23.8 19.4 27.1 18.5 33.4 33.8 39.1 39.9 41.1 Trial 3 2 .0 19.9 21.1 24.2 33.2 37.3 37.0 42.4 48.6 T 24.3 16.5 23.0 21.0 32.8 36.4 38.6 41.2 45.5 SD 2 6 5.4 3.4 2 8 .9 2.2 1.4 1.2 3.9 Subject 14 M Trial 1 0 -1.3 16.3 13.5 19.0 28.6 29.7 32.9 26.9 Trial 2 1.9 3.0 9.8 13.5 18.4 25.7 26.4 34.4 35.8 Trial 3 4.1 8.9 11.4 16.1 22.7 30.4 29.2 36.3 38.3 X 2.0 3.5 12.5 14.3 20.0 28.2 28.4 34.5 33.6 SD 2.0 5.1 3.3 1.5 2.3 2.3 1.7 1.7 5.9 Subject 15 M Trial 1 4.9 12.9 15.7 13.0 21.8 32.7 26.0 32.4 29.9 Trial 2 .7 2.0 6.4 6.6 11.4 19.5 23.1 22.6 11.6 Trial 3 -z.7 .4 o 4.8 9.9 16.9 18.4 14.0 18.9 X .9 5.1 7.3 8.1 14.3 23.0 22.5 23.0 20.1 SD 3.8 6.8 7.8 4.3 6.4 8.4 3.8 9.2 9.2 Sub ect 16 Trial 1 8.2 6.0 15.4 14.7 20.2 25.5 24.6 29.7 30.1 Trial 2 7.4 3.1 9.8 15.3 20.5 23.6 23.5 31.6 31.9 Trial 3 3.7 3.6 10.1 12.3 30.0 2..0 26.9 31.6 33.7 T 6.4 4.2 11.7 14.1 20.2 24.0 25.0 30.9 31.9 so 2.5 1.5 3.1 1.5 .3 1.3 1.7 1.0 1.8 2(11 Test Frequency Measure 125 250 500 1000 2000 3150 4000 0300 8000 Subject 17 F Trial 1 1.2 2.8 16.5 16.4 28.6 26 0 30.6 20.9 17.9 Trial 2 6.2 8.2 16.3 16.8 25.6 21 9 30.6 28.5 27.5 Trial 3 12.6 10.0 11.1 16.6 27.1 24 2 27.7 31.3 33.3 X 6.6 7.0 14.6 16.6 27.1 24 0 29.6 26.9 26.2 SD 5.7 3.7 3.0 2 1.5 2 O 1.6 5.3 7.7 Subject 18 F Trial 1 9.7 7.0 10.4 8.8 17.9 26.7 26.9 28.8 26.5 Trial 2 5.6 -1.4 1.0 9.3 11.2 18.5 24.5 25.9 29.0 Trial 3 4.0 4.4 2.9 8.4 15.4 30.4 27.5 31.4 23.0 X 6.4 3.3 4.7 8.8 14.8 25.2 26.3 28.7 26.1 SD 2.9 4.3 4.9 .5 3.3 6.0 1.5 2.7 3.0 Subject 19 F Trial 1 5.6 7.5 12.0 12.1 23.2 26.1 31.5 39.8 35.3 Trial 2 2.1 6.6 11.7 9.5 22.6 32.3 31.3 41.6 37.4 Trial 3 2.3 4.5 9.3 12.0 19.9 27.1 24.1 36.6 39.4 x 3.3 6.2 11.0 11.2 21.9 28.5 28.9 39.3 37.3 SD 1.9 1.5 1.4 1.4 1.7 3.3 4.2 2.5 2.0 Subject 20 F Trial 1 5.8 11.9 17.1 25.6 23.0 35.3 33.1 34.1 32.5 Trial 2 9.5 21.7 14.6 12.3 29.4 34.0 34.1 35.2 32.5 Trial 3 8.6 19.9 17.1. 12.2 27.9 32.7 35.6 32.3 32.0 Y 7.9 17.8 16.2 16.7 26.7 34.0 34.2 33.8 32.3 SD 1.9 5.2 1.4 7.7 3.3 1.3 1.2 1.4 ’ .3 Subject 21 F Trial 1 11.3 7.6 15.1 19.6 30.7 31.7 33.5 28.5 32.7 Trial 2 4.3 5.6 14.3 20.3 30.6 31.9 31.6 31.4 39.7 Trial 3 3.3 7.8 10.7 16.4 26.5 32.9 32.6 34.2 41.1 Y 6.2 7.0 13.3 18.7 29.2 32.1 32.5 31.3 37.8 SD 4.4 1.2 2.3 2.0 2.3 .6 .9 2.8 4.5 Subject 22 M Trial 1 2 .7 16.8 14.6 9.1 22.7 29.5 31.3 36.6 20.4 Trial 2 12.2 7.8 22.0 19.4 27.5 27.1 27.3 33.9 37.5 Trial 3 9.0 14.2 13.7 10.1 20.3 23.5 17.0 35.3 25.5 X 13.9 16.2 16.7 12.8 23.5 26.7 25.2 35.2 27.8 SD 6.0 1.8 4.5 5.6 3.6 3.0 7.3 1.3 8.7 Subject 23 F Trial 1 16.1 17.9 28.2 32.3 30.2 34.1 42.3 40.5 40.3 Trial 2 19.6 19.8 25.7 35.2 28.3 30.0 39.1 39.2 38.7 Trial 3 18.5 18.0 17.1 27.3 36.3 33.9 39.1 39.1 44.8 Y 18.0 18.5 23.6 31.6 31.6 32.6 40.1 39.6 41.2 SD 1.7 1.0 5.8 3.9 4.1 2.3 1.8 .8 3.1 Subject 24 F Trial 1 12.2 10.2 15.0 15.9 28.3 2 .1 42.4 34.8 28.0 Trial 2 5.7 1.4 7.5 14.3 19.3 21.8 36.8 13.4 27.5 Trial 3 6.1 16.8 22.1 24.7 34.3 34.0 38.9 36.5 32.7 Y 8.0 9.4 14.8 18.3 27.3 28.3 36.0 28.2 29.4 SD 3.6 7.7 7.3 5.6 7.5 6.1 8.1 12.8 2.8 2202 Test Frequency Measure 125 250 500 1000 2000 3150 4000 6300 8000 Subject 25 F Trial 1 1.3 8.4 4.2 20.6 16.6 34.8 22.5 27.8 16.1 Trial 2 11.0 8.9 10.0 14.1 29.1 24.4 29.6 34.7 30.7 Trial 3 5.5 9.3 9.0 16.8 20.4 30.8 31.5 33.5 31.5 x 5.9 8.8 7.7 17.1 22.0 30.0 27.8 32.0 26.1 so 4.8 .5 3.1 3.2 6.4 5.2 4.7 3.6 8.6 Subject 26 F Trial 1 1.9 3.6 13.4 9.4 15.7 29.4 24.9 32.4 30.8 Trial 2 1.9 5.1 10.7 13.7 26.4 36.5 38.3 38.2 33.9 Trial 3 6.5 10.2 18.3 14.8 23.2 33.7 36.3 39.9 29.7 Y 3.4 6.3 14.1 12.6 21.7 33.2 33.1 36.8 31.4 so 2.6 3.4 3.8 2.8 5.4 3.5 7.2 3.9 2.1 Subject 27 M Trial 1 25.0 11.5 6.6 18.5 22.8 48.9 41.8 34.4 37.3 Trial 2 15.7 13.3 8.9 10.6 30.7 39.7 30.3 31.7 29.6 Trial 3 12.8 14.8 12.2 4.8 20.3 38.5 33.2 33.0 2 .. X 17.8 13.2 9.2 11.3 24.6 42.3 35.1 33.0 31.2 so 6.3 1.6 2.8 6.8 5.4 5.6 5.9 1.3 5.4 Subject 28 M Trial 1 7.9 7.4 14.9 20.4 27.1 32.8 34.4 29.8 35.6 Trial 2 9.7 4.9 11.3 16.1 17.8 26.2 23.9 26.7 25.0 Trial 3 9.6 5.6 13.2 23.1 24.0 35.9 39.8 30.9 30.3 X 9.0 5.9 13.1 19.8 22.9 31.6 32.7 29.1 30.3 so 1.0 1.2 1.8 3.5 4.7 4.9 8.0 2.1 5.3 Subject 29 M Trial 1 8.4 6.5 17.0 21.3 27.2 31.5 32.7 35.6 38.0 Trial 2 10.4 6.3 11.7 11.9 25.7 24.3 26.2 32.3 37.4 Trial 3 14.2 7.4 17.2 19.6 19.0 18.9 27.7 34.1 31.1 X 11.0 6.7 15.3 17.6 23.9 24.9 28.8 34.0 35.5 so 2.9 .6 3.1 5.0 4.3 6.3 3.4 1.6 3.8 Subject 30 M Trial 1 14.2 12.9 15.9 20.7 28.1 35.0 32.2 43.4 38.9 Trial 2 8.8 17.1 17.9 16.7 21.7 26.0 28.3 35.0 35.8 Trial 3 9.0 9.5 16.2 19.7 28.1 37.5 43.7 39.7 39.1 X 10.6 13.1 16.6 19.0 25.9 32.8 34.7 39.3 37.9 so 3.0 3.8 1.0 2.0 3.6 6.0 8.0 4.2 1.8 Subject 31 F Trial 1 6.4 6.4 17.4 15.1 27.1 27.4 29.8 34.2 40.1 Trial 2 4.5 5.9 22.4 26.1 29.0 38.8 34.5 39.0 38.3 Trial 3 12.9 9.6 24.2 22.1 31.5 36.0 32.2 35.2 37.6 X 7.9 7.3 2 .3 21.1 29.2 34.0 32.1 36.1 38.6 so 4.4 2.0 3.5 5.5 2.2 5.9 2.3 2.5 1.2 Subject 32 F' Trial 1 10.2 9.9 11.4 15.3 2 .0 43.3 33.6 38.6 43.4 Trial 2 3.8 -1.2 11.3 8.2 20.5 29.8 24.8 35.1 37.2 Trial 3 13.8 14.7 7.6 13.1 27.9 35.4 35.9 34.2 33.7 Y 9.2 7.8 13.4 12.2 25.8 36.1 31.4 35.9 38.1 so 5.0 8.1 3.6 3.6 4.6 6.7 5.8 2.3 4.9 7 T 203 M ' Test Frequency edsure 125 250 500 1000 2000 3150 4000 6300 8000 Subject 33 M Trial 1 4.3 7.7 8.5 8.1 20.9 26.8 28.6 27.3 32.4 Trial 2 9.7 4.2 11.0 8.4 18.0 31.0 31.5 28.5 38.8 Trial 3 5.8 3.0 6.7 10.8 19.4 31.2 27.7 30.7 28.6 Y 6.6 4.9 8.7 9.1 19.4 29.6 29.2 32.1 33.2 SD 2.7 2.4 2.1 1.4 1.4 2.4 ' 1.9 5.7 5.1 Subject 34 M Trial 1 1.4 2.6 8.2 1.4 14.8 16.0 13.0 27.3 38.1 Trial 2 1.2 -4.8 6.4 -.8 6.1 15.1 5.9 13.6 28.9 Trial 3 6.8 8.9 2.9 8.6 21.7 22.8 25.3 25.3 28.9 Y 3.1 2.2 5.8 3.0 14.2 17.9 14.7 22.0 31.9 SD 3.1 6.8 2.6 4.9 7.8 4.2 9.8 7.4 5.3 Subject 35 F Trial 1 12.7 11.1 9.3 18.7 27.8 29.4 32.7 35.0 25.1 Trial 2 7.9 7.4 10.5 15.0 27.3 26.1 36.2 34.4 23.5 Trial 3 8.1 6.0 13.8 16.8 25.9 29.1 31.4 42.6 29.7 Y 9.5 8.1 11.2 16.8 27.0 28.2 33.4 37.3 26.1 SD 2.7 2.6 2.3 1.8 .9 1.8 2.4 4.5 3.2 Subject 36 F Trial 1 14.6 7.8 17.8 12.7 16.1 28.0 18.3 33.0 25.3 Trial 2 11.0 13.0 15.8 19.0 28.1 26.4 23.7 30.9 30.3 Trial 3 9.8 13.4 15.6 17.7 20.7 32.2 35.0 26.4 25.6 Y 11.8 11.4 16.4 16.4 21.6 28.8 25.6 30.1 27.0 SD 2.4 3.1 1.2 3.3 6.0 2.9 8.5 3.3 2.8 Subject 37 M Trial 1 12.2 1.7 7.2 10.0 17.1 12.9 28.5 33.2 30.5 Trial 2 15.9 5.7 9.3 01.4 19.5 21.3 27.4 35.6 31.4 Trial 3 18.1 9.8 8.8 14.6 22.0 21.7 29.0 34.5 35.0 Y 15.4 5.7 8.4 12.0 19.5 18.6 28.3 34.4 32.3 SD 2.9 4.0 1.0 2.3 2.4 4.9 .8 1.2 2.3 Subject 38 M Trial 1 14.5 13.4 17.2 18.9 29.8 32.5 26.6 32.9 33.8 Trial 2 17.0 14.0 20.9 21.6 29.5 36.4 38.2 36.4 20.4 Trial 3 9.4 14.8 6.2 13.2 25.3 32.0 33.7 26.0 32.5 Y 13.6 14.0 14.7 17.9 28.2 33.6 32.8 31.7 28.9 SD 3.8 .7 7.6 4.2 2.5 2.4 5.8 5.2 7.3 Subject 39 M Trial 1 6.7 9.9 12.7 19.7 27.0 35.3 34.4 34.7 31.9 Trial 2 16.6 14.1 21.1 20.7 31.8 35.4 36.8 42.2 42.1 Trial 3 17.0 14.6 16.9 19.9 28.5 38.6 38.8 35.9 40.5 Y 13.4 12.8 16.9 20.1 29.1 36.4 36.6 37.6 38.1 SD 5.8 2.5 4.2 .5 2.4 1.8 2.2 4.0 5.4 Subject 40 E' Trial 1 9.5 11.3' 1652 18.3 26.1 36.5 32.3 29.6 31.5 Trial 2 20.6 18.0 31.1 11.7 29.1 38.3 41.2 43.3 38.8 Trial 3 25.1 24.6 26.7 31.1 28.5 38.1 37.3 41.5 46.5 18.4 17.9 24.6 20.3 27.9 37.6 36.9 38.1 38.9 SD 8.0 6.6 7.6 9.3 1.5 .9 4.4 7.4 7.5 NOTE: Subjects 1-10 are from Group A (Experimenter fit and training); Subjects 11-20 are from Group 8 (Subject fit and training); Subjects 21- 30 are from Group C (Experimenter fit and no training); Subjects 31-40 are from Group 0 (Subject fit and no training). APPENDIX K RESULTS OF RELIABILITY CHECK 0N EXPERIMENTER DETERMINED ‘ THRESHOLDS 204 Table K.l. Results of reliability check on experimenter determined thresholds. Mean Mean Attenuation* Attenuation* Mean Frequency Experimenter Verifyer Differencet Deviation* r Subject 1 10.4 10.1 .3 253 11.7 11.7 o 500 18.3 18.3 o 1000 13.3 13.4 .1 2000 26.3 27.6 1.3 3150 25.2 27.1 1.9 4000 28.4 28.7 .3 6300 33.5 33.9 ,4 8000 33.4 33.5 .1 .488 .99 Subjecr 2 125 10.5 10.9 ,4 250 9.3 8.9 ,4 500 14.1 13.9 ,2 1000 14.0 14.3 .3 2000 27.8 27.8 o 3150 30.5 30.7 ,2 4000 30.8 30.8 0 6300 34.4 34.2 ,2 3000 28.5 28.6 .1 .200 .99 Subject 3 125 6,2 6.4 .2 250 7.0 6.5 .5 500 13,3 15.1 1.8 1000 18.7 18.8 .1 2000 29,2 29.3 .1 3150 32.1 32.2 .1 4000 32.5 31'? '3 31.3 - . , 3338 37.8 36.1 1,7' .622 .99 Subject 4 7,9 7.7 .2 13.3 7.3 9.0 1.7 500 21,3 21.5 .2 1000 21,1 21.0 .1 2000 29,2 27.9 1.3 3150 34,0 34.6 .6 4000 32,1 31.7 .4 6300 36,1 36.4 .3 8000 38.6 37.8 .8 .522 .99 *in dB. APPENDIX L CONFIDENCE INTERVALS OF MEAN ATTENUATION FOR ALL SUBJECTS 205 .mcwcwmuu 0: can uwm uomflnzmna no msouo “mswcfimuu one awn uoonnamnm macaw maouw umcacflmuu on one awn Houcoefiuomxm «mcacaeuu can uflm “mucoswummxmud msouoa 1‘14 o.m~- m.v~- ~.Nv m.m~- m.Hv m.mv m.o~- «.mm ooom n.vm- N.~v o.o~- m.ov v.NH- m.mv o.m~- H.Hv comm m.w~- m.H¢ n.v~- v.ov c.m~- m.ov m.m~- ~.nm ooov N.~H1 m.~v m.NN1 o.ov m.MH1 N.mm o.mN1 m.vm omam 5.5H- e.mm m.mH- o.Hm om.m- N.em «.ma- H.Hm ooem me.ei m.m~ o.n- H.m~ HH.H- v.m~ «.ma- m.o~ . coca m~.m- m.e~ m.e- N.N~ He.- a.- a.oH- e.aa com mm.- ”.BH me.~- m.wH em.~ m.aH He.m- a.- emu me.N- H.5H He.H- m.m~ am.e m.ma oe.e- 55.5 a mad «9 msouo «U msouw rm maouu admsoum. xocmdvoum new: mo new: mo new: no new: we .ucH mococwmcou .ucH mocwcflucoo .ucH mocoofimcou .u:H.meco©wmaoo .mo. n Hm>oH muwawnmnoum .muoanSm :wu mo masoum Aden now Amamnwomm CH. cowumacmuum same we mam>umucfi mococwmcow .H.q manme 206 .mcfiawmuu 0: can uwu uoonnsuua macho amcficaeuu 0: new paw umucoaaumdxm no msouo “mcwcwmuu can paw uoonnsmum macaw «mcwcaeuu can uau HmucmEHuomxmuc dsouo« . L n 1- 11 11:1 . a 4 i 1‘ W11 1444 1 ll.‘ 14 ill! 41" i n.~m- w.ee .o.a~- m.ee w.HH- m.ae e.-- m.ee ooem m.NN- e.ee “.mN- ”.me H.HH- m.we o.HN- H.me come H.6H- o.ee m.o~- e.ee N.HH- H.4e “.mN- o.~e oeoe m.-- m.nm e.o~- m.~e c.o~- h.~e e.-- m.mm omam ~.mH- 5.4m m.ma- m.em mm.e- m.em m.c~- N.em ooom He.~- H.5N mm.m- w.m~ em.o- e.e~ m.m- e.m~ coca Hm.a- m.e~ me.e- H.5N we. m.m~ mm.m- m.mH com ae.H o.e~ mm.a- e.oH He.e ~.mH mo.~- m.m~ emu me.c- m.HN mN. o.H~ mm.m a.e~ mm. m.e~. mma «a macho «U msouu «m macaw «A mnoum mocosvmum new: no new: no Gem: no new: no .ucH mocmnwmcoo .ucH mocmowucoo .ucH mocmowuaoo .ucH mflcoowmcou .mo. 0 Ho>oa aufiafinmnoum .muoonnau :0» mo mmaoum anew Ham mamauu one muomnnsm mmouom Amamnwomo aw. coHumscmHHanamaE mo mam>umucw mocmcfiwcoo 4w:H manna 207 Table: L.3. Confidence intervals of mean attenuation (in deci- bels) for all subjects (40). Probability level = .05. Frequency Confidence Interval of Mean 125 18.3 - .035 250 17.0 - .552 300 ' 22.3 - 4.80 1000 25-3 ‘ 5-60 2000 33.9 ~13.8 3150' - 40.0 -19.1 4000 41.1 - ~20.5 6300 44.2 -21.8 8000 42.8 -21.6 208 .mcucamnu 0:09 a U mmsouu “mcacfimnuum a c masonu uuwm nomnnamuo a m museum «paw HoucoEHummxouo a d mmcouoa . . 4 41411115 11314 3.04 11 14413 3‘...l 14‘ 44 ommH- - came m.¢~- m.~v m.m~- m.ev ~.m~- ,o.ov ooom e.wa- o.ee ”.mN- 6.55 5.5H- m.me e.a~- m.ee acme 5.5H- a.oe m.HN- m.ae m.ea- m.oe e.e~- o.mm oeoe N.mH- e.mm 5.55- b.5e m.m~- m.~e e.eN- e.am . omam o.HH- N.mm 0.53- e.Nm 55.5 - . «.mm e.ma- ~.Hm .oeem ee.m - w.m~ ha.m - o.a~ ma.~ - e.e~ Ne.m - N.m~ oeoa 55.5 - m.a~ Nm.e - e.m~ o5.H - a.m~ m~.m - m.o~ com Nee. “.ma me.a - H.ma m~.a ”.5H Ne.~ - N.eH omm m~.H H.5H ad.~ - o.a~ mH.H “.ma om.H - m.ea . mma «a a U mmsouo «m a 4 mmaouu «a a m banana «0 a 4 mmamwo mucoswmum .uchMWMmWMMcOU .uchMwflmWMMcou .uchMWmmwchoo .uchMwmowwucoo .mo. u Hm>ma wuflaflnmnoum .muoofinam cop we museum gnaw new Amaonaomo cwv sawumzcouue came mo mae>uouca oocwowmcoo .v.q manna REFERENCES Abel, S.M., P.W. Alberti and K. Riko. 1982. "User fitting of hearing protectors: attenuation results." In Personal Hearing Protection in Industry. P.W. Alberti, Ed. New York: Raven Press. Alberti, P.W. Ed. 1982. Personal Hearing Protection in Industry. New York: Raven Press. Alberti, P.W., K. Riko, S.M. Abel and R. Kristensen. 1979. "The effectiveness of hearing protectors in practice." The Journal of Otolaryngglogy. 8:354-359. American National Standards Institute. 1974. "Measurement of real-ear protection of hearing protectors and physical attenuation of earmuffs." $3.19 (American National Standards Institute, New York). American National Standards Institute. 1969. "American National standard specification for audiometers." 83.6 (American National Standards Institute, New York). American National Standards Institute. 1966. "Specification for octave, half-octave, and third-octave band filter sets." Sl.ll (American National Standards Institute, New York). Berger, E.H. 1983. "Using the NRR to estimate the real world performance of hearing protectors." Sound and Vibration. 17:12-l8. Berger, E.H. 1982. "Laboratory estimates of the real-world performance of hearing protectors." In Personal Hearing Protection in Industry. P.W. Alberti, Ed. New York: Raven Press. Berger, E.H. 1980a. EARLog # l "The threshold shift method of measuring hearing protector attenuation." EAR Division, Cabot Corporation. Berger, E.H. 1980b. EARLog # 4 "The performance of hearing protectors in industrial noise environments." EAR Division, Cabot Corporation. Berger, E.H. 1980c. EARLog # 5 "Hearing protector per- formance: how they work-and-what goes wrong in the real world." EAR Division, Cabot Corporation. 209 210 Berger, E.H., J.E. Kerivan and F. Mintz. 1982. "Inter- laboratory variability in the measurement of hearing piotector attenuation." Sound and Vibration. l6: - 9. Edwards, R.G., W.P. Hauser, N.A. Moiseev, A.B. Broderson and W.W. Green. 1978. "Effectiveness of earplugs as worn in the workplace." Sound and Vibration. 12: 12-22. Edwards, R.G., et al. 1978. "Effectiveness of earplugs as worn in the workplace." Sound and Vibration. 12:12-22. Environmental Protection Agency. 1979. "General pro- visions for product noise labeling requirements for hearing protectors, approval and promulgation." Federal Register 40 GET, Part 211, 56120-56147. Forshaw, S.E. "Standards for evaluating and classifying hearing protection devices." Paper presented at the Annual Convention of the American Speech-Language- Hearing Association. Toronto, Canada. November 18- 20, 1982. Gescheider, George A. 1976. P3 cho h sics: Methods and Theory. New York: John Wiley and Sons. Hays, William L. 1973. Statistics for the Social Sciences. New York: Holt, Rinehart and Winston, Inc. Humes, Larry E. 1983. ”A psychophysical evaluation of the dependence of hearing protector attenuation on noise level." Journal of the Acoustical Society of America. 73:297-310. Jerger, J. 1970. "Clinical experience with impedance audiometry. Archives of Otolaryngology. 92:311. Juneau, Roger P. 1982. "The NRR: laboratory test versus field performance - how do they relate." Hearing Instruments. 33:l4-15;51. Karmy, S.J. and A.M. Martin. 1982. "Employee attitudes towards hearing protection as affected by serial audiometry." P.W. Alberti, Ed. New York: Raven Press. Levitt, H. 1971. "Transformed up-down methods in psychoacoustics." Journal of the Acoustical Society of America. 49:467-477. 211 Martin, Alan. 1982. "How realistic are standard subjective test methods for evaluating hearing protection attenuation?" In Personal Hearing Protection in Industry. P.W. Alberti, Ed. New York: Raven Press. Melnick, William. 1978. "Temporary and permanent threshold shift." In Noise and Audiology. David D. Lipscomb, Ed. Baltimore: University Park Press. Michael, Paul L. and Gordon R. Bienvenue. 1980. ”Hearing protector performance - an update." American Industrial Hygiene Association. 41:542-546. Michael, P.L., R.L. Kerlin, G.R. Bienvenue, J.H. Prout, and J.I. Shampan. 1976. "A real-ear field method for the measurement of the noise attenuation of insert-type hearing protectors." H.E.W. Publication No. (NIOSH) 76-l8l. Nixon, Charles W. 1982. "Hearing protection standards." In Personal Hearing Protection in Industry. P.W. AlberEi, Ed. New York: Raven Press. Occupational Safety and Health Administration. 1981a. "Occupational noise exposure; hearing conservation amendment." Federal Register 29 CFR Part 1910, 4078-4179. Occupational Safety and Health Administration. 1981b. "Occupational noise exposure; hearing conservation amendment; rule and proposed rule." Federal Register 29 CFR Part 1910, 42622-42639. Occupational Safety and Health Administration. 1983. "Occupational noise exposure; hearing conservation amendment; final rule." Federal Register 29 CFR Part 1910, 9738-9785. Padilla, M. 1976. "Earplug performance in industrial field conditions." Sound and Vibration. 10:33-36. Regan, D.E. .1977. "Real-ear attenuation of personal ear protective devices worn in industry." Audiology and Hearing Education. 3:16-17. Smith, Curtis R., and Thomas E. Borton. 1981. "Premolded insert hearing protection: concerns and considerations." Hearing Instruments. 32:11-12:42. Smith, C.R., T.E. Borton, L.B. Patterson, M.T. M020 and R.R. Camp. 1980. "Insert hearing protector effects on sound pressure levels in a simulated external ear canal: preliminary results using KEMAR." Ear and Hearing. 1:26-32. 212 Stevens, 3.8. 1975. Psychophysics. New York: John Wiley and Sons. Tobias, Jerry V. 1982. "Measurers' standard and nonstandard testing of hearing protector effectiveness." In Personal Hearing Protection in Industry. P.W. Alberti, Ed. New York: Raven Press.