01Background
Learners of Korean call writing the hardest of the four skills, yet it is rarely taught well.
- Motivation: learners come through family, friends, and Hallyu, not writing (Jee, 2015; Wang, 2016).
- Difficulty: writing demands accuracy as well as fluency and complexity (Yoon, 2013).
- Time: courses stress oral skills; writing gets little class time (Jin, 2015).
- Training: teachers need a method they can use right away (Jin, 2009).
After Truscott (1996) argued against grammar correction, most studies found that written corrective feedback (WCF) improves accuracy (Chandler, 2003; Ferris, 2004). The question is now how to give it.
02Dynamic WCF
Comprehensive, indirect, coded feedback on short, frequent writing, repeated until each text is error-free (Evans et al., 2010; Hartshorn et al., 2010), grounded in skill acquisition theory (DeKeyser, 2007).
- Write10-minute paragraph in class
- CodeTeacher marks errors; returned next class
- ReviseLearner corrects and resubmits
- RepeatUntil error-free
Gap: DWCF had been tested mostly with advanced ESL learners, never with learners of Korean.
03Research questions
To what extent does DWCF increase the linguistic accuracy of intermediate KFL learners?
Will DWCF affect the fluency and complexity of their writing?
04Method
19 intermediate learners in a third-year Korean course at a private university in the western U.S.; native English speakers, most of whom had lived in Korea for 18–24 months. Intact sections over 12 weeks.
| Treatment | Control | |
|---|---|---|
| Writing | Paragraph, in class | Essay, take-home |
| Drafts × time | 30 × 10 min | 6 × 37.5 min |
| Feedback | Indirect, coded | Direct |
| Revision | Until error-free | Once |
Error codes built for Korean
24 codes in all (writing style and grammar). Codes point to the error; the learner finds the correction:
Tests and measures
Three 10-minute paragraphs before and after, on paired topics (내 인생의 목표 → 꼭 이루고 싶은 것). Two raters; Pearson’s r = .98–.99. Mixed-model repeated measures ANOVA.
| Accuracy | errors per sentence · error-free clauses · error-free syllables |
|---|---|
| Fluency | sentences · eojeols · syllables |
| Complexity | clause-to-word ratio · subordinate clauses · clauses per sentence |
New measures: a Korean syllable block (강 = ㄱ + ㅏ + ㅇ) is easy to count, so one spelling error costs a syllable, not a whole clause.
05Results
RQ1: DWCF improved accuracy on all three measures, with large effects.
Control pre-test for syllables (.874) is recomputed from the group total; the thesis table prints .087.
RQ2: Fluency and complexity did not change (all six group × time interactions, p > .33).
06Discussion
- The gains found in ESL hold for learners of Korean at the intermediate level.
- Comprehensive feedback worked, matching the many errors learners make in real writing.
- The error-free syllable is a finer unit for Korean writing research.
Limitations and next steps
- Small, intact groups (9 and 10); no random assignment.
- More time on task for the treatment group: revision ≈ 662 vs. 147 minutes over 12 weeks (self-estimated).
- The researcher taught the treatment section.
- Next: other institutions and levels, such as King Sejong Institutes; longer studies; more domains of accuracy.
Selected references
Chandler, J. (2003). Journal of Second Language Writing, 12, 267–296.
Evans, N. W., Hartshorn, K. J., McCollum, R. M., & Wolfersberger, M. (2010). Language Teaching Research, 14, 445–463.
Hartshorn, K. J., et al. (2010). TESOL Quarterly, 44, 84–109.
Kurzer, K. (2018). TESOL Quarterly, 52, 5–33.
Truscott, J. (1996). Language Learning, 46, 327–369.