J Am Acad Audiol 2007; 18(10): 813-825
DOI: 10.3766/jaaa.18.10.2
Articles
American Academy of Audiology. All rights reserved. (2007) American Academy of Audiology

Intra- and Inter-session Test, Retest Reliability of the Words-in-Noise (WIN) Test

Richard H. Wilson
,
Rachel McArdle
Weitere Informationen

Publikationsverlauf

Publikationsdatum:
07. August 2020 (online)

Retest stability and retest reliability were assessed for the Words-in-Noise Test (WIN) in two experiments involving older listeners with sensorineural hearing loss. In Experiment 1, the 70-item WIN protocol was administered during two sessions 12 months apart to examine retest stability on a sample of 315 veterans from four VA Medical Centers. The mean 50% points on the WIN were 12.5- and 12.8-dB S/N for the two sessions with a critical difference of 3.5 dB and an intra-class correlation coefficient of 0.88. [Normal recognition performance on the WIN (50% point) is ≤6-dB S/N.] In Experiment 2, intra- and inter-session retest reliability was examined for the two 35-word WIN protocols on 96 veterans, 48 of whom had mild-to-severe hearing loss (Group 1) and 48 of whom had a moderate-to-severe hearing loss (Group 2). The mean 50% points on the WIN during the two sessions (separated by 40 days) were 13.0- and 13.4-dB S/N (Group 1) and 15.3- and 15.8-dB S/N (Group 2) with no significant intra-session differences. A 3.1-dB critical difference was calculated for the groups combined with intra-class correlations of 0.89 and 0.91 for Group 1 and Group 2, respectively. La estabilidad y la confiabilidad para el re-test se evaluaron para la Prueba de Palabras en Ruido (WIN) en dos experimentos que involucraron sujetos con hipoacusia sensorineural. En el experimento 1, se administró el protocolo WIN de 70 ítems en dos sesiones con 12 meses de separación, para examinar la estabilidad del re-test en una muestra de 315 veteranos de nuestros Centros Médicos del VA. Los puntos medios 50% del WIN fueron 12.5 y 12.8 dB S/N para las dos sesiones, con una diferencia crítica de 3.5 dB y un coeficiente de correlación de intra-clase de 0.88. [El desempeño normal de reconocimiento en el WIN (punto 50%) es ≤ 6-dB S/N.] En el experimento 2, se examinó la confiabilidad del re-test tanto intra- como intersesiones para los 2 protocolos de 35 palabras del WIN, en 96 veteranos, 48 de los cuáles tenían una hipoacusia leve a severa (Grupo 1), y 48 tenían una hipoacusia moderada a severa (Grupo 2). Los puntos medios 50% del WIN durante las dos sesiones (separadas por 40 días) fueron de 13.0 y 13.4 dB S/N (Grupo 1) y 15.3 y 15.8 dB S/N (Grupo 2) sin diferencias significativas intra-sesión. Se calculó una diferencia crítica de 3.1 dB para los grupos, combinada con correlaciones intra-clase de 0.89 y 0.91 para el Grupo 1 y para el Grupo 2, respectivamente.