Abstract
For decades, the concept of learning styles has influenced English language teaching (ELT), encouraging teachers to adapt instruction to assumed visual, auditory, or kinesthetic learner preferences. However, this approach has been increasingly criticized for its weak empirical foundation and limited pedagogical value. The purpose of this article is to justify a shift from learning styles to evidence-based psychometric diagnostics of individual differences in ELT. The study focuses on measurable learner characteristics such as L2 motivation, speaking anxiety, self-efficacy, and self-regulation. Methodologically, the article combines theoretical analysis with pilot data from an 8-week school-based quasi-experimental study conducted in Astana, Kazakhstan, where psychometric profiling was integrated into English language instruction. The results showed acceptable reliability of the main diagnostic scales, comparable baseline profiles between experimental and control groups, and stronger learning gains in the experimental condition, particularly in speaking performance. In addition, learners who received psychometrically informed instruction reported more supportive assessment experiences, greater willingness to communicate, and lower anxiety during assessed speaking tasks. These findings are significant because they demonstrate that psychometric diagnostics can provide teachers with more valid and actionable information than traditional learning-style categories. The article concludes that evidence-based diagnostics offer a more scientifically grounded and pedagogically useful approach to individual differences in ELT, improving instructional fit and supporting positive washback in classroom assessment.