Theoretical issues of validity in the measurement of aided speech reception threshold in noise for comparing nonlinear hearing aid systems

Naylor, Graham (2016) Theoretical issues of validity in the measurement of aided speech reception threshold in noise for comparing nonlinear hearing aid systems. Journal of the American Academy of Audiology, 27 (7). pp. 504-514. ISSN 2157-3107

[img] PDF - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (289kB)

Abstract

Background: Adaptive Speech Reception Threshold in noise (SRTn) measurements are often used to make comparisons between alternative hearing aid (HA) systems. Such measurements usually do not constrain the signal-to-noise ratio (SNR) at which testing takes place. Meanwhile, HA systems increasingly include nonlinear features that operate differently in different SNRs, and listeners differ in their inherent SNR requirements.

Purpose: To show that SRTn measurements, as commonly used in comparisons of alternative HA systems, suffer from threats to their validity, to illustrate these threats with examples of potentially invalid conclusions in the research literature, and to propose ways to tackle these threats.

Research Design: An examination of the nature of SRTn measurements in the context of test theory, modern nonlinear HAs, and listener diversity.

Study Sample, Data Collection, and Analysis: Examples from the audiological research literature were used to estimate typical interparticipant variation in SRTn and to illustrate cases where validity may have been compromised.

Results and Conclusions: There can be no doubt that SRTn measurements, when used to compare nonlinear HA systems, in principle, suffer from threats to their internal and external/ecological validity. Interactions between HA nonlinearities and SNR, and interparticipant differences in inherent SNR requirements, can act to generate misleading results. In addition, SRTn may lie at an SNR outside the range for which the HA system is designed or expected to operate in. Although the extent of invalid conclusions in the literature is difficult to evaluate, examples of studies were nevertheless identified where the risk of each form of invalidity is significant. Reliable data on ecological SNRs is becoming available, so that ecological validity can be assessed. Methodological developments that can reduce the risk of invalid conclusions include variations on the SRTn measurement procedure itself, manipulations of stimulus or scoring conditions to place SRTn in an ecologically relevant range, and design and analysis approaches that take account of interparticipant differences.

Item Type: Article
Keywords: experiment design, hearing aid, noise, signal-to-noise ratio, SNR, speech
Schools/Departments: University of Nottingham, UK > Faculty of Medicine and Health Sciences > School of Medicine > Division of Clinical Neuroscience
Identification Number: 10.3766/jaaa.15093
Depositing User: Naylor, Graham
Date Deposited: 31 Oct 2016 11:57
Last Modified: 12 Oct 2017 21:42
URI: http://eprints.nottingham.ac.uk/id/eprint/38052

Actions (Archive Staff Only)

Edit View Edit View