A Driven Listening–Speaking Assessment Scale: Examining the Three-Dimensional Linkage of Usability-Acceptability-Practical Instructional Value!

Abstract

This study addresses limitations of the CED 4/6 listening–speaking scales, including insufficient subskill coverage, crude scoring items, and missing quantitative anchors and scenario exemplars. Based on psychometric reliability and validity theory, we designed a high‑precision assessment scale that integrates usability → acceptability → practical instructional value. The scale decomposes speaking ability into eight subdimensions: pronunciation clarity, prosodic rhythm, lexical diversity, syntactic complexity, emotional expression, cross‑cultural pragmatics, technical terminology mastery, and interaction strategies. Each subdimension includes a five‑level quantitative rubric with positive and negative exemplars and scenario demonstration videos. This design reinforces the operability of scoring anchors. The system incorporates a dynamic proficiency‑tiering mechanism. It matches instructional plans and tailored practice resources based on real‑time scores. This creates a closed‑loop assessment–feedback–intervention model. We assessed scale reliability and validity using Delphi consistency tests, KMO and Bartlett’s tests, exploratory factor analysis, and Cronbach’s α. In the empirical phase, we conducted Pearson correlation, paired‑sample difference tests, and hierarchical regression path analysis. These analyses confirmed that interface usability and exemplar clarity have significant positive effects on teacher trust and continued use. They also demonstrated that acceptability mediates the effect of usability on practical instructional value. Results indicate that the new scale significantly enhances inter‑rater consistency and teacher adoption. It supports personalized instruction in art education across regions. These findings offer a replicable methodological paradigm for the digital and intelligent transformation of educational assessment.