<![CDATA[Error analysis is a key tool for understanding second language development and designing targeted instruction in English as a Foreign Language (EFL) contexts. Despite the growing use of mobile-based assessment platforms, little is known about large-scale error patterns and their variation across proficiency levels and task types. This study analyzed 31,482 errors from 15,000 EnglishScore practice test sessions completed by 1,500 EFL learners, categorized into grammatical, lexical, syntactic, and functional domains using a modified Surface Strategy Taxonomy. A convergent parallel mixed-methods design combined descriptive statistics, ANOVA, and Pearson correlation with qualitative coding to examine error frequency, distribution, and interrelationships across Common European Framework of Reference (CEFR) proficiency levels and task types. Results showed that grammatical errors were most frequent overall, especially among beginners, while lexical and functional errors increased at higher proficiency levels. Task type significantly influenced error distribution, with gap-fill tasks eliciting the highest grammatical and lexical errors, and sentence reordering and discourse-based tasks revealing greater syntactic and functional weaknesses. Strong correlations among error types indicated that linguistic competencies are interconnected, underscoring the need for integrated, proficiency-sensitive, and task-informed instruction. The findings contribute to interlanguage development theory and offer practical guidance for educators and assessment designers integrating mobile-based tools into EFL learning]]>
Copyrights © 2026
