Errors in writing remain one of the major challenges faced by English as a Foreign Language (EFL) students. Therefore, this study aimed to identify the types and possible sources of errors found in EFL students' recount texts. This study employed a qualitative descriptive design involving 32 sixth-semester students of the English Education Study Program at Universitas Indo Global Mandiri. The data were collected through a controlled writing task and stimulated recall interviews. The writing data were analyzed using Corder's (1967) error analysis procedure and classified based on the taxonomy proposed by Heywood et al. (2025), while the interview data were analyzed using Richards' (1970) framework to identify the possible sources of errors. The findings revealed a total of 557 errors in students’ writing. Among the three Tier 1 categories proposed by Heywood et al. (2025), Grammar Sentences was the most dominant category (51.89%), followed by Grammar Words (32.68%) and Spelling (15.44%). Verb Forms and Tenses was identified as the most frequent error type. The interview findings further showed that errors were associated with interlingual transfer, overgeneralization, ignorance of rule restrictions, incomplete application of rules, false concepts hypothesized, and performance factors. These findings indicate that students' errors result from the interaction of first language interference, developing grammatical knowledge, and performance limitations during writing. Therefore, English writing instruction should integrate explicit grammar instruction with meaningful writing practice to help students improve their overall writing accuracy.

Keywords: writing errors, error analysis, EFL writing, stimulated recall interview, recount text