Assessment-as-Learning in EFL: Effects of CEFR-Based Reflection on Achievement and Calibration

High-stakes tests often report outcomes without shaping what learners do next. In large, CEFR-aligned English as a foreign language (EFL) programs, evidence is limited on whether brief, scheduled reflection can turn assessments into learning and improve calibration (alignment between self-ratings and scores). A sequential explanatory mixed-methods study was conducted with 389 tertiary learners who completed two midterm exams 8 weeks apart. The intervention combined pre-exam CEFR “can-do” self-ratings with a short, teacher-led dialogue, and nonparticipants served as a comparison group. Quantitatively, achievement was compared with ANCOVA (Exam 2 adjusted for Exam 1), and calibration was tested per skill with χ²/Cramér’s V . Scores declined across the term in both groups, but the decline was smaller with reflection. Calibration patterns were skill-specific: writing and speaking remained aligned, grammar improved by the second cycle, while listening and reading weakened. Interviews explained why the routine mattered: students reported clearer awareness of strengths and gaps, concrete strategy shifts between exams, and, critically, concise teacher feedback that translated reflection into action; anxiety and vague item wording were recurring headwinds. Findings suggest that a simple, low-cost assessment-as-learning routine can modestly buffer performance decline and refine skill-level calibration at the program scale. Practical steps are outlined for embedding it within CEFR-aligned assessment cycles.