🔄 First Visit: Redirecting automatically to original journal site...
⏸️ STOP REDIRECT & STAY ON PAGETEACHER-REVIEWED GENERATIVE AI LESSON PLANNING: A QUASI-EXPERIMENTAL ANALYSIS OF PERCEIVED PEDAGOGICAL QUALITY | EDUCATIONAL RESEARCH AND INNOVATION
Abstract
Abstract
Teachers increasingly use Generative Artificial Intelligence (GenAI) to draft lesson outlines, examples, classroom activities, and formative assessment prompts. The rationale for this study is that teacher-reviewed GenAI planning is spreading faster than classroom-level evidence on its pedagogical value. Existing approaches include GenAI lesson-plan assistants, AI-supported critique tools, automated feedback systems, and teacher professional development models. These approaches remain difficult to evaluate because many studies rely on reviews, design reports, or teacher-education settings rather than longitudinal classroom comparisons. This study contributes a quasi-experimental evaluation of teacher-reviewed GenAI lesson planning in Grade 12 instruction. Using survey data from 7 teachers, 14 class sections, and 431 students, it compares conventional planning with a human-in-the-loop workflow in which teachers review, adapt, and approve GenAI-generated lesson materials before classroom delivery. The personal information of the participants was not used in this study, and direct identifiers were removed and replaced with coded IDs before analysis to protect personal data. Difference-in-Differences (DID) models, adjusted for baseline academic average, digital literacy, attendance, teacher fixed effects, and class-section clustered standard errors, estimate a positive treatment contrast of +6.13 points on a 10-50 perceived pedagogy scale (p < 0.001). The largest construct-level gains are observed in formative feedback, lesson personalization, and active engagement. The findings imply that GenAI lesson planning can improve students' perceived classroom experience when it is governed as a teacher-controlled support process rather than as an automated instructional replacement.
Teachers increasingly use Generative Artificial Intelligence (GenAI) to draft lesson outlines, examples, classroom activities, and formative assessment prompts. The rationale for this study is that teacher-reviewed GenAI planning is spreading faster than classroom-level evidence on its pedagogical value. Existing approaches include GenAI lesson-plan assistants, AI-supported critique tools, automated feedback systems, and teacher professional development models. These approaches remain difficult to evaluate because many studies rely on reviews, design reports, or teacher-education settings rather than longitudinal classroom comparisons. This study contributes a quasi-experimental evaluation of teacher-reviewed GenAI lesson planning in Grade 12 instruction. Using survey data from 7 teachers, 14 class sections, and 431 students, it compares conventional planning with a human-in-the-loop workflow in which teachers review, adapt, and approve GenAI-generated lesson materials before classroom delivery. The personal information of the participants was not used in this study, and direct identifiers were removed and replaced with coded IDs before analysis to protect personal data. Difference-in-Differences (DID) models, adjusted for baseline academic average, digital literacy, attendance, teacher fixed effects, and class-section clustered standard errors, estimate a positive treatment contrast of +6.13 points on a 10-50 perceived pedagogy scale (p < 0.001). The largest construct-level gains are observed in formative feedback, lesson personalization, and active engagement. The findings imply that GenAI lesson planning can improve students' perceived classroom experience when it is governed as a teacher-controlled support process rather than as an automated instructional replacement.