Computerized development of speech perception test

What is it about?

Hearing loss affects millions of people worldwide. Since hearing is fundamental to human communication and serves as the primary pathway through which spoken language is learned, understood, and shared, checking a person's ability to perceive and repeat spoken words is a core part of standard hearing tests. To ensure these test results are accurate and reliable over time, clinicians need to use multiple different lists of test words that are identical in difficulty and contain a similar balance of speech sounds. However, traditional methods for creating these lists rely on manual selection, which frequently forces compromises between language features and how difficult the words are to hear. The purpose of this study is to introduce a reliable, computer-driven approach to automatically design perfectly balanced word recognition tests, demonstrating its effectiveness by developing the first highly standardized Hebrew Word Recognition Test (HWRT).

Why is it important?

This research provides a direct dual contribution to both local clinical practice and global audiological science. For clinicians in Israel, the HWRT delivers a much-needed, scientifically validated tool that ensures ear-specific diagnostic testing, hearing aid fittings, and cochlear implant evaluations are highly accurate and completely free from learning or memory biases during repeated testing. Crucially, the standardized 25-word lists provide the unique clinical flexibility to shorten testing times without sacrificing diagnostic reliability. Globally, because the materials were balanced using strict speech-in-noise difficulty metrics, they possess the robust equivalence needed for cross-linguistic scientific research. Most importantly, this objective, data-driven framework serves as a freely accessible, replicable template that can be used to construct high-quality speech audiometry tests in any language worldwide.