L2 Writers' Feedback-Seeking from ChatGPT for Automated Written Corrective Feedback: Insights from a Multiple Case Study
6th International Postgraduate Colloquium (InPaC) 2023
L2 writers’ feedback seeking from ChatGPT
for automated written corrective feedback
Yan Da
Ph.D. candidate from
SoLLaT, Universiti Sains Malaysia
| Table of contents | ||
|---|---|---|
| 01 | 02 | 03 |
| Introduction Research Background, Problem Statement | Literature Review Key theoretical and conceptual insights | Method Method and materials applied in the study |
| 04 | 05 | |
| Results Results from the study | directions | Concluding Remarks Discussion, implication, and future |
01 Introduction
| Feedback in L2 learning/writing | ||
|---|---|---|
| ● | Recent advancement | (Carless & Young, 2023; Hyland & Hyland, 2019; Vuogan & Li, 2022) |
| ● | Shifting paradigm | (Jensen et al., 2023; Winstone et al., |
| 2022) | ||
| ● | Technology-enhanced Ranalli, 2021) | feedback (Koltovskaia, 2020; “Feedback seeking involves purposely seeking information about one’s own level of performance, interpreting it, and applying it to seek, generate, and use information for enhancement purposes. “(Carless & Young, 2023; Anseel et al. 2015) |
Research Context: GAI for feedback
What differences GAI has brought about to feedback
practices and research in L2 peadogy?
- Outstanding performance (Wu et al.,
- Emphasis on interactivity in human-AI communication (Giray, 2023)
- Calls for agentic engagement (Nieminen et al., 2021; Winstone et al., 2016)
| 02 | ||
|---|---|---|
| Review | Literature |
| Automated feedback: | from automation to AI-generation | |||
|---|---|---|---|---|
| Traditional Origin and shortcomings tradtional automatic feedback providers, eg., Pigai.com, Criterion, etc. (Bai & Hu, 2017) | of utilization | Comtemporary Improvement, providers (mostly AI-based), e.g., Grammarly. (Koltovskaia, 2020) | affordances, and of new feedback | Generative AI Why ChatGPT (or its likes) has the potential to be an effective feedback provider? (Mizumoto & Eguchi, 2023; Wu et al., 2023) A knowledge gap : Students’ behavior/agency |
| • • | How How environment. | Feedback seeking (behavior) is feedback seeking is feedback seeking feedback from teachers, peer learners and the learning | defined ? Avenue to reduce uncertainty but also subject to a lot of factors. conceptualized Students’ utilization of inquiry and monitoring strategies in seeking | A knowledge | (Anseel et al., 2015) in L2 learning? (Papi et al., 2019) gap : |
|---|---|---|---|---|---|
| Passive | feedback generator/co-constructor | receiver | vs. active feedback in the age of AI. |
Theoretical underpinnings
New feedback paradigm
(Jensen et al. 2023)
Sociomaterial/ecological perspective of learning/feedback
(Chong, 2021; Gravett, 2022)
Activity Theory/Human-technology interaction
(Kaptelinin & Nardi, 2018)
Research Questions
RQ1:
What characterizes RQ2:
the development of
What are L2 writers’ feedback seeking perceptions of the
abilities of the development of participants in using
feedback seeking ChatGPT for abilities? feedback?
03 Methods
and Materials
Subject, Context and Procedures
Participants The project
3 Chinese EFL Longitudinal study; each learners; multiple week 2 sessions of training
case study and 3 sessions of practice
Data collection Analysis
Document (ChatGPT Document analysis;
prompts); in-depth Content analysis interview
Some ChatGPT-related methodological considerations
- Official GPT 3.5 version
- A uniform prompting template
- Exploratory approaches to use ChatGPT for AWE/AWCF
04 Results*
**** Partially presented as the study is on-going***
Prompts Created by Toby
| Prompt | Words | Words per prompt | Rounds | Prompts per round | Time spent (min) | |
|---|---|---|---|---|---|---|
| Week 1 | 306 | 3213 | 10.5 | 48 | 6.4 | 251 |
| Week 2 | 426 | 4856 | 11.4 | 35 | 12.2 | 215 |
| Week 3 | 520 | 7124 | 13.7 | 31 | 16.8 | 177 |
| Week 4 | 589 | 9365 | 15.9 | 27 | 21.8 | 111 |
| Week 5 | 696 | 12389 | 17.8 | 26 | 26.8 | 96 |
| Sum | 2537 | 36947 | - | 167 | - | 850 |
| Mean | 507.4 | 7389.5 | 13.9 | 33.4 | 16.8 | 170 |
Prompts Patterns of Toby
05 Concluding
Remarks
Discussion, limitation, implication
Discussion Implications
1. How should we understand feedback seeking in the
For research and context of ChatGPT pedagogy
2. And its developmental trajectories?
Limitations Future research
What the present Directions for future study failed to research.
achieve?
| Thanks! Do you have any questions? | 谢谢! Terima Kasih! | |
|---|---|---|
| CREDITS: | This presentation template was created by : yanda 87 Flaticon includes icons by Flaticon | Slidesgo Slidesgo, and @student.usm.my Freepik, and infographics & images by Freepik : +60 12 466 7298 (also ) |