Meaningful, manageable, constant, and timely
Learners ask for a better way to improve their writing, yet writing is rarely taught successfully in KFL classrooms.
Learners come to Korean through family, friends, and Hallyu. Their reasons rarely involve writing.
Writing demands accuracy as well as fluency and complexity, with a low tolerance for error.
Courses stress oral communication, so writing gets a small share of class time.
A clear, efficient method that any language teacher can start using right away, without long-term training.
Since Truscott (1996), most studies find that WCF helps learners write more accurately.
Writing, feedback, and revision happen individually, outside the class period.
Each learner gets feedback on their own errors and becomes aware of their own patterns.
141 learners of German: no method of correction improved accuracy
WCF is ineffective and possibly harmful; abandon it
Learners revise more accurately with feedback
The question shifts from whether to how
Findings for focused feedback “may be too focused to be practical.”Evans et al., 2010, p. 386
No consensus in ESL or in KFL; a multi-step feedback-revision cycle works better than one-time feedback (Chandler, 2003).
Comprehensive, indirect, coded feedback on short, frequent writing, repeated until each text is error-free.Evans et al., 2010; Hartshorn et al., 2010 · skill acquisition theory, DeKeyser, 2007
A paragraph on a set topic in 10 minutes, in class
Teacher marks errors with codes; returns it next class
Learner corrects the draft and resubmits
Until the paragraph is error-free
Learners also keep tally sheets of their error types and error rates.
Holistic feedback; learners note the errors they make
One paragraph, written in 10 minutes
Writing several days every week
Feedback arrives the next class day
Feedback on one or two long essays a semester cannot do this.
| Evans et al., 2010 | ESL, intensive program |
| Hartshorn et al., 2010 | ESL, intensive program |
| Evans et al., 2011 | ESL, university |
| Hartshorn & Evans, 2012, 2015 | ESL, intensive program |
| Kurzer, 2018 | ESL, three levels |
| Akiyama & Fleshler, 2013 | Japanese: grammar items only |
Mostly advanced ESL learners. Never tested with learners of Korean, or at the intermediate level.
To what extent does DWCF increase linguistic accuracy for intermediate learners of Korean as a foreign language?
Will DWCF affect the fluency and complexity of their writing?
intermediate learners in a third-year Korean course at a private university in the western U.S.
Intact sections; students were not randomly assigned.
| Treatment · DWCF | Control · traditional | |
|---|---|---|
| Writing | Paragraph, in class | Essay, take-home |
| Time per draft | 10 min | 37.5 min |
| Number of writings | 30 | 6 |
| Total writing time | 300 min | 225 min |
| Feedback | Indirect, coded | Direct |
| Revision | Until error-free | Once |
24 codes in all, written for features English codes do not cover.
Codes point to the error; the learner finds the correction.
A new paragraph starts every class day while earlier ones are still being revised.
| Pre-test | Post-test |
|---|---|
| 내 인생에 영향을 준 인물A person who has had an impact on my life | 내 인생에서 가장 중요한 사람The most important person in my life |
| 내 인생의 목표The purpose of my life | 꼭 이루고 싶은 것The thing I really want to achieve |
| 역사적인 사건A historical event | 이 세상에 일어난 특별한 사건A special event that happened in the world |
Highlighted: measures not used in earlier DWCF research.
Letters stack into one syllable block, so syllables are easy to tell apart.
Small errors in spelling and inflection lower the score a little, not a whole clause.
Error-free syllables give learners credit for what they got right.
Both counted errors, error-free clauses, and error-free syllables in every test paragraph.
Pearson’s r between raters, all p < .001
Mixed-model repeated measures ANOVA: group (treatment, control) × time (pre, post).
The treatment group went from 4.08 to 2.38 errors per sentence. The control group did not improve.
Error-free clauses rose from .321 to .572 of all clauses with DWCF, and fell slightly in the control group.
The treatment group rose from .808 to .894. Control pre-test (.874) is recomputed from the group total; the thesis table prints .087.
Large effects on every measure: DWCF helped intermediate KFL learners write more accurately.
| Measure | Group × time | |
|---|---|---|
| Fluency | Sentences | F(1, 17) = 0.03, p = .875 |
| Eojeols | F(1, 17) = 0.13, p = .728 | |
| Syllables | F(1, 17) = 0.81, p = .381 | |
| Complexity | Clause-to-word ratio | F(1, 17) = 0.05, p = .822 |
| Subordinate clauses per clause | F(1, 17) = 0.97, p = .340 | |
| Clauses per sentence | F(1, 17) = 0.74, p = .402 |
Thirty short, timed paragraphs did not cost learners fluency or complexity.
The accuracy gains found in ESL hold for learners of Korean.
Systematic feedback on all errors was effective, matching the many errors learners make in real writing.
Error-free syllables offer a finer measure for Korean writing research.
The treatment group wrote and revised for longer, which may explain part of the gain.
Minutes over 12 weeks · * estimated by participants
Questions and comments are welcome.
Oh, S. (2020). The effect of dynamic written corrective feedback for learners of Korean [Master’s thesis, Brigham Young University]. BYU ScholarsArchive.