Lexical Adaptation during Formative Assessment under the Merdeka Curriculum Framework in Non-Formal EFL Classrooms

Abstract

The implementation of decentralized, student-centered curricula frequently demands that educators dynamically modify their instructional input during real time formative evaluation sequences. However, how these pedagogical demands alter the micro linguistic parameters of teacher talk remains under-researched, particularly within non-formal educational settings characterized by mixed ability learner groups. This study investigated the real time lexical adaptation of English as a Foreign Language (EFL) teachers during embedded formative assessments within a non-formal framework in Indonesia. Utilizing a quantitative corpus linguistic design, a specialized spoken micro corpus of 14,250 running tokens was compiled from authentic classroom interactions and segmented into general instruction and formative assessment sub-corpora. The text files were computationally processed to calculate variations in the Moving-Average Type-Token Ratio (“MATTR”) and lexical density percentages. The empirical findings revealed a systematic, statistically significant reduction in both linguistic metrics when instructors transitioned into active evaluation phases. The group mean for “MATTR” dropped from 0.63 to 0.52, while the mean lexical density declined from 46.1 percent to 37.5 percent during active testing loops. This structural contraction was driven by a massive increase in formulaic question prompts and the purposeful repetition of target lexical items to prevent comprehension breakdowns. These results imply that interactive pedagogical tasks automatically trigger structural language simplification, meaning that teachers require explicit training in micro linguistic management to ensure that temporary input simplification is balanced by deliberate phases of lexical enrichment.