How L2 Teachers Identify and Respond to GenAI-Generated Essays: A Sensemaking Theory Investigation
Description
This repository contains materials and anonymized data for a mixed-methods study on CSL writing teachers’ detection of GenAI-generated essays and their sensemaking and pedagogical responses to AI-related uncertainty. The dataset supports replication of identification accuracy analyses and secondary analyses of teachers’ cue use and decision-making. The study examines (a) teachers’ accuracy in classifying texts as GenAI-generated vs. student-written, (b) whether performance varies with teacher characteristics (e.g., experience; GenAI-use proficiency), (c) how text properties (e.g., score bands) affect detectability, and (d) how teachers construct “reasonable explanations” and response strategies when certainty is limited. What the files contain. Demographic Information of Interview Participants (DOCX): This DOCX file provides a participant characteristics table for the 20 interviewees in the qualitative phase of the study. Each row corresponds to one interview participant, identified only by an anonymous code (S1–S20). GenAI prompt template (DOCX): the standardized prompt used to generate banded GenAI essays; it instructs the model to compare four sample essays (Bands 2–5) and generate a new Band 2 essay under explicit constraints (e.g., 280–320 words; frequent character/lexical/grammar errors; lower fluency). Questionnaire instrument (DOCX): background items plus a 12-essay task; for each essay participants provide a binary source judgment, a brief written justification, and a 2–5 quality score using the provided rubric; plus an overall open-ended cue question. Interview protocol (DOCX): semi-structured guide eliciting reflections after the rating task, uncertainty and responsibility judgments, cue use and decision pathways, institutional support needs, and classroom response strategies.