An open-set sentence speech recognition in noise test in Cantonese Chinese, which is applicable on children and adults with hearing impairment, is proposed to be developed in this project.
INTRODUCTION: Speech recognition in noise is an important aspect of human communication. One's speech recognition ability in noise is inevitably affected by any hearing impairment, and may only be quantified with specific language at an appropriate difficulty. There is a niche among clinically available test materials between word recognition and 10- syllable sentence recognition in noise in Cantonese Chinese. Ten-syllable sentence items are commented to be too challenging to young children and elderly population. An assessment tool for pediatric patients, and the whole spectrum of patients with significant hearing impairment, compromised language or cognitive abilities, is deemed necessary. Both the research and clinical audiology communities need a valid and reliable outcome measurement tool for sentence recognition abilities in noise of the specific population. An open-set sentence speech recognition in noise test in Cantonese Chinese, which is applicable on children and adults with hearing impairment, is proposed to be developed in this project to bridge the gap. METHODS: Thirty three participants with typical hearing, from each of the 5 age groups from their 6 years of age to adulthood, will be recruited to contribute to the normative data of the proposed test. One hundred participants with various degree of hearing impairment will be recruited for the validation of the proposed new test as well. A sentence pool of at least 150 sentence items will be constructed based on the basic sentence syntax in Cantonese Chinese. Each sentence item will be 7 syllables in length on average. The words used in the sentences will be words frequently used in daily life by both persons with typical hearing and hearing impairment. A 4-talker babble in Cantonese Chinese will be generated as the competing noise in this proposed new test. OUTCOMES: By Rasch modelling, sentence items with high probability of consistently and accurately estimating sentence recognition ability in noise across different participants will be selected and distributed to a few parallel lists. Rater and test-retest reliability of the test items and lists will be established. Content-related and construct-related evidence of validity of lists will also be supported by psychometric analyses as well as expert consultation. APPLICATION: A valid and reliable assessment tool, of sentence recognition ability in noise in Cantonese, applicable to persons with hearing sensitivity ranging from profound impairment to typical, in both the research and clinical audiology fields, will be constructed.
Study Type
OBSERVATIONAL
Enrollment
113
Measurement of speech recognition performance in noise
The Chinese University of Hong Kong
Hong Kong, Hong Kong
Person Ability of Participants in Sentence Recognition in Noise
In accordance with Item Response Theory, the ability of individual participants in sentence recognition in noise, measured by the proposed new test, is generated through Rasch modelling or other relevant models.
Time frame: up to 1 month
Item Difficulty of Test Items in Sentence Recognition in Noise
In accordance with Item Response Theory, the difficulty of individual proposed test items in sentence recognition in noise, contributed by all participants, is generated through Rasch modelling or other relevant models.
Time frame: up to 3 years
Internal Consistency of Test Items
Cronbach's alpha or other relevant measurements will be used to estimate the consistency of the items on measuring the same dimension.
Time frame: up to 3 years
Rater Reliability of Test Items
A portion of the test sessions will be scored twice by more than one raters. The intra-class correlation (ICC) or other relevant measurements will be employed to estimate rater reliability.
Time frame: up to 1 month
Test-Retest Reliability of Test Items
A portion of participants will be tested twice. Pearson's coefficient of correlation or other relevant measurements will be computed to measure the correlation between the first and the second test results.
Time frame: up to 1 month
Content-related Evidence of Validity of Test Items
Construction of sentence test items with elements from various relevant language corpus, based on the basic sentence syntax in Cantonese Chinese, and evaluated with reference to expert opinions, will be regarded as content-related evidence of the test item validity.
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
Time frame: up to 3 years
Construct-related Evidence of Validity of Test Items
Unidimensionality and local independence, or other relevant measurements, in accordance with Item Response Theory, will be regarded as construct-related evidence of the test item validity.
Time frame: up to 3 years