ASSESSING ORAL PROFICIENCY IN YOUNG EFL LEARNERS: A STUDY OF EIGHTH-GRADE PUPILS’ PERFORMANCE AT THE SERBIAN NATIONAL ENGLISH COMPETITION
ASSESSING ORAL PROFICIENCY IN YOUNG EFL LEARNERS: A STUDY OF EIGHTH-GRADE PUPILS’ PERFORMANCE AT THE SERBIAN NATIONAL ENGLISH COMPETITION

Summary/Abstract: The present study examines construct and predictive validity of oral assessment used at the 2024 Serbian National Competition in the English language for eighth-grade pupils. The study descriptively and critically evaluates the assessment of oral proficiency among top EFL learners at the end of compulsory education. A total of seventy-four competitors participated in the final speaking exam. Their performances were assessed holistically and analytically across four criteria: coherence and interaction, lexical accuracy, grammatical accuracy, and pronunciation. Descriptive statistics and correlational analysis were employed to compare holistic and analytic scores and to identify which component most strongly influenced overall ratings. The findings indicate that most candidates achieved scores between 17 and 20 points, suggesting a potential ceiling effect that may limit discrimination among top-level performances. The strongest predictors of the overall speaking score were coherence, interaction and lexical accuracy, whereas grammar and pronunciation played a secondary, though still meaningful, role. As the written test components showed a very loose relation with speaking performance (indicating weak predictive validity of the written exam for oral proficiency), the results of the study call for the refinement of performance descriptors and rubrics to enhance construct validity and differentiation of proficiency levels. These findings suggest that the speaking test functions as a reliable global measure of advanced oral proficiency and adds distinct value to the competition’s assessment system, but that the analytic criteria and written components may require certain refinement if finer-grained distinctions among top-level performers are desired. The study contributes empirical evidence to the discussion of nationwide standards for a reliable EFL oral proficiency assessment.

• Details
• Contents