A Language Neutral Nonsense Speech-in-Noise Test for Dravidian Language Speakers: Development and Psychometric Evaluation

Prajwal Kumar  Eranna; Jayashree S  Bhat; Aradith  S

doi:10.18502/avr.v35i2.21203

Prajwal Kumar Eranna Nitte (Deemed to be University), Nitte Institute of Speech and Hearing, Mangalore, India
Jayashree S Bhat Nitte (Deemed to be University), Nitte Institute of Speech and Hearing, Mangalore, India
Aradith S Shravan Hearing Aid Centre, Shivamogga, India

DOI: https://doi.org/10.18502/avr.v35i2.21203

Keywords: Speech perception in noise; development; psychometric function; language independence; slope function; nonsense word

Abstract

Background and Aim: Assessing Speech Perception in Noise (SPIN) in multilingualcontexts like India is challenging due to the lack of linguistically appropriate test materials.Recognizing the limitations of existing SPIN tests in multilingual and clinically diversesettings, this study addressed critical need by developing a language-neutral, nonsenseSPIN test material tailored for Dravidian languages.

Methods: Nonsense word lists in the consonant vowel consonant vowel format weregenerated using a random combination of common phonemes in the Dravidian languages(Kannada, Malayalam, Telugu, Tulu, Tamil). These lists were recorded, and the Signal-to-Noise Ratio required to achieve 50% Speech Recognition (SNR50) was used to selectoptimized lists based on a criterion of mean ±0.15 SD. The final lists were administeredto 50 normal-hearing individuals at 0 dB SNR. Language independence was evaluated bycomparing performance across speakers of the five languages. Further performance was alsoassessed across eight SNR levels to establish a psychometric slope function and goodnessof fit was assessed. To evaluate test–retest reliability, 12 participants were retested within aone-week interval.

Results: The study resulted in 4 final optimized lists based on SNR50 selection criteria andfurther analysis. Lists showed sensitivity to varying SNR levels, as reflected by consistentpsychometric function slopes. Comparable performance across language groups confirmedthe language-independent nature of the test.

Conclusion: Developed test provides audiologists with a reliable and standardized tool toassess SPIN. By eliminating the influence of familiarity and ensuring language neutrality,the test is well-suited for clinical use across speakers of Dravidian languages.