Characterizing ChatGPT's Feedback for FYW: Analyzing Feedback Responses to Inquiry-Driven Essays
Saved in:
| Title: | Characterizing ChatGPT's Feedback for FYW: Analyzing Feedback Responses to Inquiry-Driven Essays |
|---|---|
| Language: | English |
| Authors: | Kirkwood Adams, Maria G. Baker |
| Source: | Thresholds in Education. 2025 48(2):159-181. |
| Availability: | Academy for Educational Studies. 2419 Berkeley Street, Springfield, MO 65804. Tel: 417-299-1560; e-mail: cqieeditors@gmail.com; Web site: http://academyforeducationalstudies.org |
| Peer Reviewed: | Y |
| Page Count: | 23 |
| Publication Date: | 2025 |
| Document Type: | Journal Articles Reports - Research |
| Education Level: | Higher Education Postsecondary Education |
| Descriptors: | Freshman Composition, Artificial Intelligence, Man Machine Systems, Natural Language Processing, Feedback (Response), Writing Evaluation, Essays, Technology Uses in Education |
| Geographic Terms: | New York (New York) |
| ISSN: | 0196-9641 2381-5485 |
| Abstract: | In response to (1) studies finding that essay feedback generated by ChatGPT might be useful for student writers and (2) studies observing ChatGPT's tendency to adhere to narrow genre definitions when producing writing, our study seeks to examine whether ChatGPT can provide useful feedback in a first-year writing learning environment that targets a more nuanced genre definition of the academic essay as a form of inquiry. Our study employed three sequentially constructed feedback request prompts to assess how ChatGPT would go about generating outputs when asked to respond to a set of 50 published student essays that successfully achieve the goals of Columbia University's FYW class. The use of exemplary essays as input data revealed a durability in ChatGPT's feedback priorities under the influence of more robust prompting strategies. We found that in two-part responses, ChatGPT provided a descriptive as well as a prescriptive portion of feedback, the former largely affirmative, the latter critical with a focus on revision suggestions that guide toward narrow genre markers. The descriptive feedback mode seemed more adaptable to our capacious definition of genre, and the model responded less critically to exemplary essays when in descriptive mode, suggesting that further study of descriptive feedback capacities might be helpful in developing ChatGPT's potential to give useful feedback to students in FYW courses. However, in two brief stress tests of our results, we found cause for amending our findings. |
| Abstractor: | As Provided |
| Entry Date: | 2025 |
| Accession Number: | EJ1483971 |
| Database: | ERIC |
Be the first to leave a comment!
