From Baseline to Blended Feedback: A Within-Subjects Study of AI-Generated and Teacher-Delivered Feedback in EFL Writing

This 8-week mixed-methods study examined how English as a foreign language (EFL) learners navigate, interpret, and appropriate ChatGPT-generated and teacher-delivered feedback across sequential instructional phases, and what motivational and writing consequences this navigation produces. Using a within-subjects design, 24 intermediate university-level EFL learners completed 5 genre-based paragraph tasks across 3 feedback phases: baseline (no feedback), ChatGPT-only (GPT-4o mini), and blended (ChatGPT combined with teacher mediation). Data comprised rubric-based writing scores, phase-specific surveys, and focus-group interviews analyzed through the AI-Mediated Writing (AIM-W) framework. Qualitative findings revealed a developmental arc in feedback literacy: from prompt-bounded experimentation in Phase 2, through layered scaffolding recognition and selective agency, to ethical calibration of AI use in Phase 3. Motivational trajectories followed self-determination theory predictions: autonomy and competence dominated during ChatGPT-only interaction, while teacher mediation reintroduced relatedness, the missing motivational dimension, and unexpectedly strengthened trust in AI-generated feedback itself. Writing performance patterns showed micro-level stability across phases despite escalating genre demands, and macro-level gains under blended feedback, though these are reported cautiously given that genre complexity co-varied with feedback phase, constituting a design limitation that precludes causal attribution. The study contributes an empirically grounded account of how feedback literacy develops in multisource AI feedback settings, with specific implications for prompt-literacy instruction, feedback sequencing, and the pedagogical conditions under which AI-generated feedback can be responsibly integrated into EFL writing classrooms.

Authors

Institutions

Publication Details

Journal
Language Teaching Research
Published
2026-11-19
DOI
https://doi.org/10.1177/13621688261499965
Primary Topic
EFL/ESL Teaching and Learning
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

From Baseline to Blended Feedback: A Within-Subjects Study of AI-Generated and Teacher-Delivered Feedback in EFL Writing

Harun ÇİFTCİ
1 links
Language Teaching Research
EFL/ESL Teaching and Learning
article

From Baseline to Blended Feedback: A Within-Subjects Study of AI-Generated and Teacher-Delivered Feedback in EFL Writing

Harun ÇİFTCİ
article en

Abstract

This 8-week mixed-methods study examined how English as a foreign language (EFL) learners navigate, interpret, and appropriate ChatGPT-generated and teacher-delivered feedback across sequential instructional phases, and what motivational and writing consequences this navigation produces. Using a within-subjects design, 24 intermediate university-level EFL learners completed 5 genre-based paragraph tasks across 3 feedback phases: baseline (no feedback), ChatGPT-only (GPT-4o mini), and blended (ChatGPT combined with teacher mediation). Data comprised rubric-based writing scores, phase-specific surveys, and focus-group interviews analyzed through the AI-Mediated Writing (AIM-W) framework. Qualitative findings revealed a developmental arc in feedback literacy: from prompt-bounded experimentation in Phase 2, through layered scaffolding recognition and selective agency, to ethical calibration of AI use in Phase 3. Motivational trajectories followed self-determination theory predictions: autonomy and competence dominated during ChatGPT-only interaction, while teacher mediation reintroduced relatedness, the missing motivational dimension, and unexpectedly strengthened trust in AI-generated feedback itself. Writing performance patterns showed micro-level stability across phases despite escalating genre demands, and macro-level gains under blended feedback, though these are reported cautiously given that genre complexity co-varied with feedback phase, constituting a design limitation that precludes causal attribution. The study contributes an empirically grounded account of how feedback literacy develops in multisource AI feedback settings, with specific implications for prompt-literacy instruction, feedback sequencing, and the pedagogical conditions under which AI-generated feedback can be responsibly integrated into EFL writing classrooms.

Language Teaching Research
Isparta University of Applied Sciences (TR)
Quality education
EFL/ESL Teaching and Learning
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon