Comparison of Whole-Word and Phoneme Scoring for Monosyllabic Words of Shahid Beheshti University of Medical Sciences in Presence of Speech- Spectrum Noise: A Psychometric Function Study

Abstract

Background and Aim: Speech audiometry assesses functional hearing beyond pure-tone thresholds, reflecting real-world speech perception. The present study compared whole- word and phoneme scoring methods for the Persian monosyllabic word lists of Shahid Beheshti University of Medical Sciences (SBMU-1) presented in speech-spectrum noise to determine their psychometric equivalence and sensitivity to Signal-to-Noise Ratios (SNRs) changes. Methods: Twenty-two young adults with normal hearing participated. 150 Consonant- Vowel-Consonant (CVC) words were presented binaurally at six SNRs of −5, 0, +5, +10, +15, +20 dB in speech-spectrum noise at 60 dB HL. Recognition performance was analyzed using whole-word and phoneme scoring. List equivalency and scoring effects were examined using Friedman and Wilcoxon signed-rank tests with Bonferroni correction. Results: Speech recognition improved systematically with increasing SNR for both scoring methods. Phoneme scoring consistently yielded higher scores than whole-word scoring, especially under low SNRs, revealing a 10–20% performance advantage. At high SNRs, the two methods converged. Across lists, differences were minor and list-specific, confirming the general equivalency of SBMU-1 lists in noise. Conclusion: Phoneme scoring provides a more sensitive measure of perceptual performance in noise by capturing partial recognition, whereas whole-word scoring better represents functional communication ability. The SBMU-1-word lists are psychometrically balanced and suitable for both clinical and research applications in Persian speech audiometry.