Statistical evaluation and prediction of word recognition test scores of the Foneties Verteenwoordigende Eenlettergrepige Woordlyste in Afrikaans (FVEWA)

Abstract

Theoretical framework: In audiology, speech audiometry is generally accepted as clinically more sensitive than pure tone audiometry to detect reduced auditory analytical skill in terms of complex acoustic structure, as it includes higher-level linguistic function in the evaluation. Although speech audiometry is considered to be an integral part of the audiometric test battery, there is currently no phonetically balanced Afrikaans speech material available to measure speech recognition. In order to obtain an accurate picture of an individual's speech recognition abilities of monosyllabic words, it is important to perform speech recognition tests in a person's first language. Aim: The purpose of this study was to determine the reliability of the Foneties Verteenwoordigende Eenlettergrepige Woordlyste in Afrikaans (FVEWA) to measure speech recognition abilities in members of an Afrikaans-speaking population. Methodology:The research is based on a quantitative retrospective study of audiological records of participants assessed at a private audiology practice. The FVEWA is evaluated in terms of interlist equivalence as well as performance-intensity or psychometric function for use in clinical practice. Results: The six FVEWA wordlists were found to be statistically equivalent in terms of difficulty level. The average intensity at which participants obtained 50% word recognition was 14.2 dBGP. The slope, at 50% (%/dB), of the wordlists used in the FVEWA was 3.1%/dB.