Participate and share: Poster 
This research was conducted within the framework of mixed methods research and adopts a social constructivist perspective. This framework was selected because it allowed for a comprehensive understanding of the research topic. Qualitative data was collected through teacher surveys, which provided valuable insights into their impressions and opinions. Additionally, quantitative data in the form of rubric scores was gathered for comparative analysis. This mixed methods approach enabled a well-rounded exploration of the research questions by considering both the subjective experiences of teachers and objective data for validation
Data Collection:
For this study, we began by selecting a sample of 30 student essays from various campuses within the school district. These essays were chosen based on their relevance to our research topic, specifically discussing the impact of the book '1984' from junior English courses.
Participant Selection:
To gather teacher input, we designed a pre-survey to understand their initial impressions of AI grading and AI technology in general. With the assistance of the district English coordinator, we reached out to teachers across the district, and participation was entirely voluntary. Those who chose to participate were asked to use a rubric to grade the selected essays as they would any other student work. Importantly, the identity of the papers was concealed from the teachers.
Grading Process:
While teachers were grading the essays, the researcher entered both the rubric scores and the essays into the generative AI model, ChatGPT. The results from ChatGPT's grading were recorded alongside the teachers' scores.
Analysis:
Upon completion of the grading process, we conducted a comparative analysis to assess the level of agreement between AI grading and teachers' grading. This comparison was based on specific grading criteria and statistical analyses, the details of which are available in the study's supplementary materials. The results were shared with the teachers and a post-survey was conducted to gather their final opinions.
My results will not be complete until the end of the first semester of school, mid-December. I anticipate that the AI's scores for the papers will, on average, be lower compared to those graded by the teachers. This expectation is based on the consideration that AI grading is not influenced by factors such as fatigue, distraction, or individual teacher bias. While these factors may affect teacher grading, AI grading remains consistent and objective. Additionally, the removal of teacher-student familiarity, which often impacts teacher grading, further contributes to the expected variance in scores.
The educational and scientific importance of this study lies in its response to the evolving landscape of education technology, where generative AI is increasing in usage. This study holds valuable insights for ISTE attendees, offering them a glimpse into the possibilities and implications of incorporating generative AI into educational practices.
By conducting this research, we aim to provide ISTE attendees with a nuanced understanding of how generative AI can, possibly, be effectively utilized in the grading process. If the experiment demonstrates a substantial similarity between AI and teacher grading, it can spark discussions among educators on how AI can serve as a supportive tool in their grading workflows while preserving the essential role of the teacher. Conversely, if the experiment reveals significant differences in grading outcomes, it will signal a cautionary note, suggesting that teachers may need to exercise discretion in adopting AI-based grading solutions. Ultimately, this study empowers educators to make informed decisions about integrating AI into their teaching practices, ensuring that it aligns with their pedagogical goals and the best interests of students.
Original research
Return
Trips and Tours