Initial Development and Preliminary Validation of Khasi Word Lists for Speech Audiometry in Adults

Background: Speech audiometry requires linguistically and culturally appropriate test materials to accurately assess functional hearing ability. The present study aimed to develop and preliminarily validate equivalent Khasi word lists for adult speech audiometry in quiet and in noise conditions. Methods: This descriptive, cross-sectional observational study was conducted in the otorhinolaryngology department of a tertiary care teaching hospital in Shillong, Meghalaya, India, from February 2025 to July 2025 on adult Khasi participants. Adult native Khasi speakers with normal hearing on pure tone audiometry were recruited at each stage after obtaining written informed consent. An initial pool of 200 commonly used Khasi words (145 monosyllabic, 55 bisyllabic; consonant-vowel-consonant (CVC) and consonant-vowel-consonant-vowel (CVCV)) was identified from literature, dictionaries, and media. Ten native speakers rated word familiarity on a five‑point Likert scale, and 80 words (40 monosyllabic, 40 bisyllabic) with scores ≥4 were shortlisted. Comprehensibility (equality) testing was carried out in 10 normal‑hearing adults at a 20 decibel (dB) sensation level, and words correctly identified by ≥70% of participants were retained. Four lists of 20 words each (10 monosyllabic, 10 bisyllabic) were formed. List equivalence in quiet was assessed in five normal‑hearing adults at 40 dB sound pressure level (SPL) using live stimuli presented via a calibrated clinical audiometer in a sound‑treated room; word recognition scores (WRS) were analyzed using repeated‑measures analysis of variance (ANOVA). The same lists were then evaluated in another five adults in noise at signal‑to‑noise ratios (SNRs) of −3, −5, and −10 dB using speech‑shaped noise, and WRS were analyzed similarly. Results: The four lists showed high mean WRS in quiet (range = 18.6-19.8/20), with no statistically significant difference between lists (p = 0.1081). In noise, there were no significant differences in mean WRS between lists at any SNR (−3 dB: p = 0.3425; −5 dB: p = 0.577; −10 dB: p = 1.00), and all lists maintained high scores across SNRs. Descriptive phonemic analysis demonstrated comparable distribution of consonants and vowels across the four lists. Conclusion: Four preliminarily validated and equivalent word lists in Khasi have been developed for adult speech audiometry. These lists are suitable for use in both quiet and noisy environments at varying SNRs. These lists fill an important gap in regional audiological assessment tools and can be used clinically for evaluating speech recognition and monitoring outcomes of hearing rehabilitation among Khasi‑speaking adults, while further validation in larger and hearing‑impaired cohorts is warranted.