Abstract
Speech audiometry has become an essential component of the audiological test battery. Various speech audiometry materials have been developed in different languages and also revised later to meet criteria such as familiarity of the words, phonetic/phonemic balance, number of test items, and dialectal variations. The current study focused on developing a PB word list in Tamil for adults. The study was carried out in two phases. Phase I involved the development of a PB word list in Tamil. Phase II included evaluating the effectiveness of the developed word lists on adults with normal hearing sensitivity and hearing impairment. A total of 500 bisyllabic words were used in creating a master list. The collected words were subjected to familiarity testing with 40 native Tamil speakers. From the master list of 500 words, 90% of the participants rated 320 words as most familiar, and those words were included in the study. The 320 words were validated by eight experts working in the fields of audiology, speech-language pathology, linguistics, and psychology. The disagreed-with words were excluded from the list, and only 270 words that were agreed upon by all the experts were included in the final list. The final lists comprised four-word lists with 25 phonetically balanced (PB) words per list. The final word lists were recorded and tested for goodness. These final words were then administered to find SIS scores for two groups. Group I consisted of adults with 120 normal hearing sensitivity, and Group II consisted of 130 adults with hearing impairment. The results were analysed using SPSS software. The mean SIS in adults with normal hearing sensitivity ranged from 99.4 to 99.8%. The SIS of adults with normal hearing sensitivity revealed no significant difference (p > .05) with respect to gender, ears, or across the four-word lists. All four-word lists provided similar scores, establishing the equivalency of word lists in individuals with normal hearing sensitivity as well as hearing impairment. A psychometric curve of performance intensity function in adults with normal hearing sensitivity revealed that the SIS score increased as the sensation level increased from 10 to 40 dB SL. There is a linear increase in SIS scores from 10 to 20 dB SL. A plateau was attained at 40 dB SL with 100% SIS scores. Thus, the present study recommends the use of 40 dB SL for SIS testing. Mean SIS scores for adults with mild, moderate, and severe hearing loss were found to be 90.1%, 65.4%, and 41.8%, respectively. A significant decrease in SIS was observed as the degree of hearing loss increased.
Keywords: Phonemically balanced, Speech identification, Hearing impaired, Psychometric function
Introduction
Speech audiometry provides information on the degree of communication handicap caused by hearing loss, differential diagnosis, and selection of amplification devices, as well as monitoring the progress of aural rehabilitation [1]. One of the tests in speech audiometry is the Speech Identification Score (SIS). The SIS measures the ability of an individual to identify speech at suprathreshold levels and is generally assessed using phonetically balanced (PB) monosyllabic words. Ramkissoon, Proctor, Lansing, & Bilger [2] supported the idea that speech audiometry materials should be assessed in the patient’s native language and dialect to obtain accurate scores on speech audiometry testing.
In India, there is an increasing need to develop culturally and linguistically sensitive materials for assessment and rehabilitation because of the multilingual and multicultural diversity. Several authors [3–5] felt the need to revise or newly develop some of the speech audiometry materials for the following reasons: unfamiliarity of words in the existing word list, requirement of many word lists to assess an individual at different testing conditions, necessity to overcome dialectal variations, and because few Indian languages (Tamil, Telugu, Kannada) had very few monosyllabic words as these languages had vowel-ending words.
There have been several attempts made to revise or develop some speech audiometry materials for various reasons. Yathiraj and Vijayalakshmi [6] developed PB word lists for adults in Kannada that contained eight phonemically balanced word lists. Each word list consisted of 25 bisyllabic words. Later, Manjula, Antony, Shivarju, and Geetha [3] developed and standardised 21 PB word lists in Kannada to test an individual in different clinical conditions in addition to the previously available test material, attributing this to the fact that improvements in hearing device technology have increased the number of features in a hearing device. When these features or parameters have to be evaluated and compared under a number of conditions, a larger number of word lists is mandatory. Sreedhar, Venkatesh, Nagaraja, and Srinivasan [4] developed paired word lists for testing of the Speech Recognition Test (SRT) in Telugu and compared their developed word lists with the existing empirically untested bisyllabic word list developed by Padmaja [7]. The authors reported that the words in the existing lists were infrequently used in conversations and unfamiliar to the Telugu listeners. The authors also reported that the material did not contain cues for guessing, which made it difficult to repeat during testing.
Need
The PB word lists for adults were developed 40 years ago by Dayalan [8] in Tamil. It contained two-word lists with 25 monosyllabic PB words (consonant-nucleus-consonant combinations) in each list. This word list has four main concerns. The first concern is that native Tamil speakers find it difficult to repeat these words during testing as some of the words in the list are obsolete (e.g., /sa:r/, /ya:n/, /na:r/) and not used in everyday conversation. According to Zubick et al. [9], it is important that speech stimuli consist of words that are considered “familiar” in the native language. The accuracy of speech audiometry relies heavily on the subject’s knowledge of the test material. Hence, it is essential to develop speech audiometry material in Tamil with words that are used in everyday context in the present scenario and that are comfortable for adults to repeat during the testing. Second, the Tamil language is considered "diglossic” (formal Tamil and colloquial Tamil) in nature [10]. The colloquial spoken Tamil consists of very few consonant-ending monosyllabic words. The colloquial Tamil uses bisyllabic vowel-ending words in sentences and discourse. Thus, the monosyllabic words in the word list developed by Dayalan [8] are infrequently used in everyday conversation. Similarly, bisyallabic words are used for speech identification tests in Kannada [6] and Telugu [11]. Third, the word list previously developed by Dayalan [8] was standardised only on 30 native Tamil-speaking adults (mean age of 19.5 years) with normal hearing sensitivity. Fourth, the validity of these word lists was not assessed in individuals with hearing impairments. This study aims to develop a word list with familiar words and words used in everyday context. Since Tamil is considered a vowel-ending language, this study intends to develop PB word lists using bisyllabic words. Chinnaraj, Neelamegarajan, and Ravirose [12] developed a 25-word list and standardised the words lost in the normal hearing population in quiet and in the presence of noise. However, the authors did not validate the word list with hearing-impaired adults. Thus, the aim of this study is to develop a phonetically balanced word list in Tamil for speech identification score testing and evaluation of its effectiveness in adults. The developed word lists were standardised on a larger number of adults with normal hearing sensitivity. The effectiveness of the word list will be assessed in a clinical population with a larger number of hearing impairments.
Method
Ethical Consideration
The study was approved by the institutional ethics committee (REF: CSP/16/JAN/45/42). Oral and written consent was obtained from participants prior to data collection.
Study Design
Cross-sectional study.
Phase I: Development of Phonemically Balanced Word List in Tamil
The following steps were involved in constructing the Phonemically Balanced word list in Tamil.
Collection of Words
A master list was prepared, collecting 500 bisyllabic words with consonant vowel consonant vowel (CVCV) combinations in Tamil. Words in this master list were taken from different sources like newspapers, magazines, general books, and Tamil websites. It was ensured that these words were not borrowed from any other language and that they did not have any clusters. These CVCV words had all the phonemes in Tamil.
Familiarity Testing
Familiarity testing was carried out on 40 participants who were native Tamil speakers from the south and north regions of Tamil Nadu. It was ensured that an equal number of participants from urban and rural areas of Tamil Nadu participated in the familiarity testing to eliminate dialectal variations of Tamil. The participants were asked to rate each word based on a 3-point familiarity rating scale as "most familiar" (MF), "familiar" (F), and "unfamiliar" (UF). The words that 90% of the participants considered MF were taken to construct the final word list. From the master list of 500 words, 90% of the participants rated 320 words as the most familiar.
Content Validation
The list with the most familiar 320 words was given to eight experts working in the fields of audiology, speech-language pathology, linguistics, and psychology for content validation. The word list was given to these experts to check if these words fulfil the familiarity rating criteria: absence of emotional overlay, cultural and religious overlay, and bisyllabic words with CVCV combinations. The experts were requested to mark these 320 words as "agree" or "disagree" and write suggestions on the words if needed. The responses from each expert were collected, and their suggestions were discussed with each expert. The disagreeable words were excluded from the list. The 270 words that were agreed upon by all experts were included in forming the final word list.
Preparation of PB Words
Content-validated words were sorted into four word lists, each comprising 25 words. As there was no recent literature available on the percentage of occurrence of phonemes in Tamil, it was necessary for the author to use an older version developed by Ramakrishnan (1957). The phonemes from each list were matched with the percentage of phonemes in Tamil. It was ensured that the final four PB word lists had occurrences of each phoneme that were similar to those mentioned by Ramakrishnan, Nair, Chiplunkar, Rajaram [13]. Each word list was randomised to make eight-word lists. Thus, these word lists could be used in eight different conditions for clinical purposes such as speech perception measures, PI-PB function, hearing aid trials, functional hearing aid gain, and to monitor progress in auditory training. The final word lists are enclosed in Appendix 1.
Recording and Goodness Rating
The final word lists were audio recorded using CSL software in a sound-treated room. The words in the final lists were spoken by a female speaker whose native language was Tamil. The spoken words were picked up by a condenser microphone, which was kept at a 10 cm distance and positioned at a 0° angle with reference to the face. These words were recorded in 32-bit resolution at a sampling rate of 44,100. The final recordings were transferred to a compact disc using Nero burning software at a slow copying speed. The recordings were edited using Adobe Audition software (Version 3.0). The interstimulus interval between the two words was set to four seconds. These words were normalised to equate the loudness across the words. A calibration tone of 1 kHz was added at the beginning of the word list. The final list was subjected to a goodness rating by 10 normal-hearing individuals (native Tamil speakers). The stimulus was presented from the audiometer through the loudspeaker at 40 dB SL (with reference to SRT). Stimuli with 100% correct responses were included in the recorded version. The stimuli with incorrect responses were rerecorded and further evaluated by goodness testing. Words that were identified 100% by all individuals with normal hearing sensitivity were included in the final word lists.
Phase II: Evaluating the Effectiveness of the Word List on Adults with Normal Hearing Sensitivity and Hearing Impairment
Participants
Two groups of participants were included in the study. Group I consisted of adults with normal hearing sensitivity, and Group II consisted of adults with hearing impairments. Informed consent was obtained from the participants prior to the testing. Group: I included 120 native Tamil-speaking adults with equal distribution of males and females between the ages of 18 and 30 (mean age of 23.2 years). The normal hearing sensitivity of the Group-I participants was confirmed by testing the air conduction threshold, which was less than 15 dB HL from 250 Hz to 8 kHz at octave frequencies. An otoscopic examination was performed on all the participants to ensure a normal middle ear status. Group II included 130 native Tamil-speaking adults with hearing impairment between the ages of 18 and 50 (mean age of 42.5 years). None of the participants in both groups had articulatory errors or an intellectual disability.
Equipment
The middle ear status was examined using a handheld otoscope. A calibrated Piccalo Plus audiometer was used to estimate pure tone threshold and to assess speech audiometry testing.
Procedure
The participants were seated comfortably with TDH 39 headphones placed on their ears. Initially, Speech Recognition Threshold (SRT) was obtained in Tamil using a recorded version of the bisyllabic spondee word list developed by Dayalan (1976). These spoken words were presented through headphones. Participants were instructed to repeat the words even if they heard them very softly and to guess the word if they could. The lowest intensity at which the participant was able to repeat three out of five words was taken as SRT.
Speech identification score (SIS) was obtained using the final word lists that were developed in this study. These words were presented from a computer through a Piccalo Plus audiometer. These words were delivered through headphones at a presentation level of 40 dB SL with reference to SRT. The word lists were presented in a random order to avoid the order effect. The ear at which the words were presented varied among the participants to avoid ear effects. The participants were instructed to listen carefully through headphones and repeat each word that was heard. An open-ended oral response was obtained from the participants. After the participants repeated the words, the author immediately checked for the correct response using the scoring key. Each correct response was given a score of 1, and an incorrect response was given a score of 0. The raw score was then converted into a percentage.
Performance intensity functions were performed on 50% of the total participants. The performance intensity function is a graphical representation of the percentage of words correctly identified or recognised as a function of the intensity level of the word. To obtain the performance intensity function, the final word lists were presented to group-I participants with normal hearing sensitivity at 10 dB SL, 20 dB SL, 30 dB SL, and 40 dB SL with reference to SRT. SIS was obtained on each of the above sensational levels. Test–retest reliability was checked in 10% of the participants with normal hearing sensitivity and hearing impairment. The reliability testing was carried out after 3 weeks to check the reliability of the word list.
Statistical Analysis
The results were analysed using Statistical Package for Social Science (SPSS; version 17). The mean and standard deviation of SIS for each list were calculated for groups I and II separately. The sample did not follow a normal distribution curve, and the values were skewed towards the right side of the curve in group I. Skewed data were observed as group 1 adults with normal hearing sensitivity had a SIS of more than 90%. Hence, the Friedman test was carried out to find the significance difference in the main and interaction effects of gender, ears, and across the four word lists. A mixed ANOVA test was used to find the significance difference in performance between the ear, gender, and four lists in adults with hearing Impaired. A paired t test was carried out to find a significant difference in SIS between ears and among the list. An independent t test was done to find significant differences in speech identification performance between males and females. Performance intensity functions for each list were plotted for group I participants who had normal hearing sensitivity. (Table 1)
Table 1.
Mean and standard deviation of SIS (N = 120) in group I individuals with normal hearing sensitivity
| Gender | Ear | List 1 | List 2 | List 3 | List 4 |
|---|---|---|---|---|---|
| Mean (S.D) | Mean (S.D) | Mean (S.D) | Mean (S.D) | ||
| Male | Right | 99.6 (1.08) | 99.6 (1.2) | 99.6 (1.08) | 99.7 (1) |
| Left | 99.6 (1.08) | 99.4 (1.2) | 99.6 (1.2) | 99.5 (1.2) | |
| Female | Right | 99.6 (1.08) | 99.4 (1.2) | 99.6 (1.2) | 99.5 (1.2) |
| Left | 99.4 (1.36) | 99.6 (1.2) | 99.6 (1.2) | 99.8 (0.72) |
Results
Speech Identification Scores in Adults with Normal Hearing Sensitivity
The mean and standard deviation (SD) were calculated for 120 participants in group 1 (individuals with normal hearing sensitivity) across the four word lists for both ears in males and females. The values are depicted in Table 2. It can be observed from Table 1 that individuals with normal hearing sensitivity (group I) had speech identification scores of more than 95%.
Table 2.
Mean and standard deviation of SIS in group II adults with hearing impairment
| Degree of hearing loss | Speech identification scores (%) Mean (S.D) |
|---|---|
| Mild (N = 82) | 93.4 (4.18) |
| Moderate (N = 88) | 78.4 (8.62) |
| Severe (N = 30) | 53.8 (12.4) |
The Friedman test was carried out to find the significance difference in the main and interaction effects of gender, ears, and across the four-word lists. The test revealed that there was no significant difference in main and interaction effects with respect to gender, ears, or across the four-word lists (p > 0.05). A Mann–Whitney test was carried out to compare the mean SIS between gender and age. The results revealed that there was no significant difference (p > 0.05) between the scores of males and females between the right and left ears. The Wilcoxon signed-rank test was used to determine the difference in SIS among the four-word lists. The results revealed that there was no significant difference (p > 0.05) on SIS across lists in adults with normal hearing sensitivity. Results indicated that all the four-word lists provide equivalent scores for participants with normal hearing sensitivity.
Speech Identification Scores on Adults with Hearing Impairment
The mean and standard deviation were calculated for Group II participants (130 adults with hearing impairment), as shown in Table 2. Among 130 participants, 200 ears were found to be affected by sensorineural hearing losses, for which SIS was carried out. On these affected ears, 30 participants in Group II had unilateral sensorineural hearing loss. Among200 ears, 50 ears with sensorineural hearing loss, mild sensorineural hearing loss is seen in 82 ears, moderate sensorineural hearing loss in 88 ears, and severe sensorineural hearing loss in 30 ears. It can be seen from the table that the mean SIS decreases as the degree of hearing loss increases.
A mixed ANOVA was carried out to see the significant difference among the four-word lists and the degree of hearing loss. The results revealed that there is a significant difference (p < 0.05) in the main and interaction effects. Further, a paired t test was done to see the significant difference among the four lists. The results indicated that there was no significant difference among the four word lists in individuals with hearing impairment. In addition, an independent t test was carried out to find a significant difference in SIS among different degrees of hearing loss. The results revealed that there was a significant difference (p < 0.05) among the different degrees of hearing loss (mild, moderate, and severe degrees of hearing loss). The results indicated that these word lists are also sensitive in indicating the different degrees of hearing loss.
Comparison of Speech Identification Scores Between Adults with Normal Hearing Sensitivity and Hearing Impairment
Group I participants with normal hearing sensitivity have a higher score than group II participants with hearing impairment (Tables 1 & 2). An independent t test was carried out to find a significant difference between adults with normal hearing sensitivity and those with different degrees of hearing impairment. The results revealed a significant difference (p < 0.05) in SIS between adults with normal hearing sensitivity and hearing impairment of mild, moderate, and severe degrees of hearing loss.
Psychometric Curve of Performance Intensity Function in Adults with Normal Hearing Sensitivity
Performance intensity functions across the four word lists were carried out for 50% of group-I participants with normal hearing sensitivity (60 participants) at 10 dBSL, 20 dBSL, 30 dBSL, and 40 dBSL with reference to SRT. A psychometric function between mean SIS and sensational levels for each word list at sensation levels 10 dBSL, 20 dBSL, 30 dBSL, and 40 dBSL is depicted in Fig. 1. The results reveal that there is an increase in the SIS score as the sensation level increases. A plateau is achieved at 40 dB SL with 100% SIS scores. There is a linear increase in SIS scores from 10 to 20 dB SL. The SIS score reached between 90 and 100% at 30 dB SL. All individuals had 100% SIS at 40 dB SL, which is a reasonably comfortable listening level. SIS at lower sensational levels (10 dB SL and 20 dB SL) was reduced among the participants.
Fig. 1.
Psychometric function of SIS scores across sensational level for all four lists in individuals with normal hearing sensitivity
Test Re-Test Reliability
The test–retest reliability testing was performed on 10% of the individuals in groups I and II. The same tests were repeated on these individuals after 3 weeks of initial testing. Cronbach's alpha test was used to compare the mean difference between the test and re-test speech identification scores of these individuals. The results revealed that there is a high correlation between test and re-test SIS scores (0.96; p.001) in participants with normal hearing sensitivity and (0.9; p.001) hearing impaired.
Discussion
The present study found no effect of ear, gender, or word list on speech identification score. The results of the present study are in consensus with the results of Sreedhar, Venkatesh, Nagaraja, and Srinivasan [4]. The authors reported no significant difference between the speech recognition thresholds of males and females for the bisyllabic spondee word list in Kannada for adults with normal hearing sensitivity.
Bhimte and Rangasayee [14] developed monosyllabic word lists in Hindi for children. The results showed that there was no significant difference between male and female children from 4 to 6 years of age. The results of this study were found to be contradictory with the study done by Boominathan and Yathiraj [15], where there was a gender effect in children. The authors reported that females performed better than males on SIS. However, these results were obtained in children, unlike the current study, which involved adult participants. Similar to earlier studies by several authors [4, 11, 16] the present study also revealed no significant difference (p > 0.05) between the SIS of the right and left ear.
There is a significant difference in speech perception among different degrees of hearing loss, from mild to severe SN hearing loss. In individuals with sensorineural hearing loss, the SIS was linearly reduced with increasing degrees of hearing loss in the present study. Schoepflin [17] also reported that the SIS decreased as the degree of hearing loss increased from minimal to profound. The results of the present study were similar to the findings of the study done by Shetty, Mathai, and Uppunda [18]. The authors found significant the effect of degree of hearing loss on SIS in individuals with mild, moderate, moderately severe, severe, and profound degrees of sensorineural hearing loss. SIS was tested using the PB word list in Kannada at 40 dB SL.
The present study found significant differences in the SIS using developed word list between individual with normal hearing individual and hearing impaired. Dias, Devadas, and Rajashekhar [5] also reported similar results in their study. The authors developed monosyllabic word lists in Goan Konkini and administered them to individuals with mild to moderate SNHL. The authors found a significant difference between the SIS scores of individuals with normal hearing sensitivity (100%) and hearing impairment (89.3%). Thus, the present study concluded that SIS obtained using developed bisyllabic words differentiated individuals with normal hearing sensitivity from those with hearing impairments.
All four PB word lists are found to be equal based on the psychometric function in the present study. Manjula, Antony, Shivaraju, and Geetha [3] developed 25-word lists with bisyllabic words in Kannada. The authors found that among the 25 lists, four were not equivalent in performance to SIS. Thus, the authors eliminated those four lists and suggested using the remaining 21 lists for testing. The eliminated four-word lists were utilised as practise lists. Similar to the present study results, Manjula, Antony, Shivaraju, and Geetha [3]reported an increase in the SIS scores as the sensational levels also increased from 10 to 30 dBSL. SIS scores reached a plateau at 40 dBSL in the present study and also in earlier studies [5, 11]. Based on the above findings, the present study recommends that SIS to be carried out at 40 dB SL in clinical settings.
The developed word list had high test–retest reliability of SIS scores in the normal hearing population and the hearing impaired. Grange [19] found similar results to those in the current study. The author administered SIS using CID-W22 word lists to 40 participants with varying degrees of hearing loss. The results revealed that there was high test–retest reliability of SIS scores at 40 dB SL. Thus, the present study revealed that the developed four-bisyllabic word list in Tamil has high test–retest reliability in adults with normal hearing sensitivity as well as in adults with hearing impairment. Similar results were obtained by Dias, Devadas, and Rajashekhar [5], who verified the test–retest reliability of the developed two-word list in Konkini on 20 normal-hearing subjects after a gap of 1 week. The results revealed that subjects obtained 100% SIS during the test as well as the re-test.
Conclusion
The four developed PB word lists were found to be equivalent and provide similar SIS scores. Hence, all four lists can be interchangeably used for individuals with normal hearing sensitivity and hearing impairments to assess SIS in different clinical settings. It has a potential application in differentiating individuals with normal hearing from those who are hearing impaired. The developed PB list is sensitive in indicating the different degrees of hearing loss. Further, the developed four-word list has been randomised to make it eight lists so that it could be utilised in differential diagnosis, deciding on appropriate rehabilitative options, and determining the efficacy of rehabilitative devices.
Acknowledgements
The authors would like to thank Ms. Selvi, the coordinator for cochlear Implantation program of Sri Ramachandra Institute of Higher Education and Research, who helped us to recruit the patient for the study
Author Contributions
First Authors contributed in terms of material development, data collection, Literature review writing. Corresponding authors contributed in terms of conception of the study, design, supervision, analysis, interpretation, critical review.
Funding
No funding received for this project.
Declarations
Conflict of interest
The authors have no relevant financial or non-financial interest to disclose.
Ethics Approval
The current study was approved by the ethics committee of SriRamachandra Institute of Higher Education and Research (DU), Chennai. (REF: CSP/16/JAN/45/42).
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Waghmare P, Mohite J, Gore G. Development of Marathi speech recognition test (pediatric): a preliminary report. J Indian Speech Hear Assoc. 2011;25(1):59–64. [Google Scholar]
- 2.Ramkissoon I, Proctor A, Lansing CR, Bilger RC. Digit speech recognition thresholds (SRT) for non-native speakers of English.2002, ASHA publications [DOI] [PubMed]
- 3.Manjula P, Antony J, Kumar KS, Geetha C. Development of phonemically balanced word lists for adults in the Kannada language. J Hear Sci. 2015;5(1):22–30. doi: 10.17430/893515. [DOI] [Google Scholar]
- 4.Sreedhar JS, Venkatesh L, Nagaraja MN, Srinivasan P. Development and evaluation of paired words for testing of speech recognition threshold in Telugu: a preliminary report. J Indian Speech Lang Hear Assoc. 2011;25(2):128–136. [Google Scholar]
- 5.Dias MA, Devadas U, Rajashekhar B. Development of speech audiometry material in goan Konkani language. Lang India. 2015;15(2):268–280. [Google Scholar]
- 6.Yathiraj A, Vijayalakshmi CS (2005) Phonemically balanced wordlist in Kannada. University of Mysore
- 7.Padmaja. Spondees in Telugu for testing of SRT (1978) In: Kacker SK, Basavaraj V. (Eds), Indian speech, language and hearing tests: The ISHA battery-1990
- 8.Dayalan S (1976) Development and standardization of phonetically balanced test materials in tamil language. Unpublished Master’s Dissertation: University of Mysore
- 9.Zubick HH, Irizarry LM, Rosen L, Feudo P, Kelly JH, Strome M. Development of speech-audiometric materials for native Spanish-speaking adults. Audiology. 1983;22(1):88–102. doi: 10.3109/00206098309072772. [DOI] [PubMed] [Google Scholar]
- 10.Matiki AJ. A case review of Tamil diglossia. Lang India. 2010;10(11):392–397. [Google Scholar]
- 11.Kumar SR, Mohanty P. Speech recognition performance of adults: a proposal for a battery for Telugu. Theory Pract Lang Stud. 2012;2(2):193–204. doi: 10.4304/tpls.2.2.193-204. [DOI] [Google Scholar]
- 12.Chinnaraj G, Neelamegarajan D, Ravirose U. Development, standardization, and validation of bisyllabic phonemically balanced Tamil word test in quiet and noise. J Hear Sci. 2022;11(4):42–47. doi: 10.17430/JHS.2021.11.4.5. [DOI] [Google Scholar]
- 13.Ramakrishna BS, Nair KK, Chiplunkar VN, Atal BS, Rajaraman V. Statistical studies in some Indian languages with applications to communication engineering. J Res. 1957;4(1):25–35. [Google Scholar]
- 14.Bhimte MS, Rangasayee MR (2015) Topic: development of Hindi language monosyllabic speech perception test material for children
- 15.Boominathan P (1999) Picture speech identification test for children in Tamil, Unpublished Master’s dissertation: University of Mysore
- 16.Sinthiya K (2009) High frequency speech identification test in Tamil. Unpublished Master’s Dissertation. University of Mysore, Mysore
- 17.Schoepflin J (2012) Back to basics: speech audiometry. Retrived from http://www.audiologyonline.com/articles/back-to-basics-speech-audiometry-6828
- 18.Shetty HN, Mathai JP, Uppunda AK. Disability certificate for individuals with hearing impairment–time to rethink. Indian J Otol. 2017;23(1):7–12. doi: 10.4103/0971-7749.199509. [DOI] [Google Scholar]
- 19.Grange ME (2013) Test-retest reliability in word recognition testing in subjects with varying levels of hearing loss. Brigham Young University. All thesis and dissertation: Brigham Young University. Retrieved from http://scholarsarchive.byu.edu/cgi/viewcontent.cgi? article=4479&context=etd