Collaborative Research [FW-HTF-RL]: Enhancing the Future of Teacher Practice via AI-enabled Formative Feedback for Job-Embedded Learning

点击翻译按钮获取中文摘要

This project envisions a future of work where advanced technologies provide automated, job-embedded, individualized feedback to drive professional learning of the future worker. To achieve this goal, it addresses a fundamental question: Are evaluative or non-evaluative feedback systems more effective in driving professional learning? This question will be tested on professionals where objective, fine-grained feedback is especially critical to improvement--the teaching professions. This research will be situated within English and language arts (ELA) instruction in middle- and high school classrooms, where underperformance and inequality in literacy outcomes are persistent problems facing the U.S. Current methods of supporting teacher learning through feedback are sparse, cumbersome, subjective, and evaluative. Thus, a major reconceptualization is needed to provide feedback mechanisms that- meaningfully affect teacher practice and are accessible to all. In partnership with TeachFX, an industry leader in technology-enabled instructional feedback, this project will work with teachers to design and test systems of automated feedback. Insights from the study will lead to feedback systems that empower teaching professionals, generate continued professional learning, and ultimately, increase student achievement. The scientific merits of the project are centered around the foundational question of whether instruction can be construed entirely along a continuum of ineffective to more effective practice. The hypothesis is that the richest opportunities for on-the-job feedback in the professions are agnostic technologically-driven feedback systems, which offer choice, withhold evaluation, make room for varied teacher practices, and promote a greater locus of control. The project has several goals towards testing this hypothesis, including: (1) to work with a diverse panel of teachers to design and refine automated feedback systems; (2) to enhance the robustness and fairness of computational models that underlie automated feedback; and finally, (3) to test fundamental design principles of professional feedback. The project will begin by leveraging TeachFX's corpus of instructional observations from approximately 5,000 educators to develop automated, robust, accurate, unbiased, generalizable, and interpretable feedback models. Next, working with teacher participants, the feedback interfaces will be co-designed and iteratively refined. Further, a variety of observational and survey-based measures will be used to assess teacher responsiveness to feedback. The project will culminate in a longitudinal, experimental study contrasting the effects of evaluative- with non-evaluative feedback on teacher learning, empowerment, and student achievement outcomes with a sample of 300 teachers. The study will create a blueprint for effective and efficient professional observation and feedback, and working systems to implement that feedback, driving the next generation of advancement in the sciences, technology, engineering, and mathematics.This project is supported by two programs at NSF: Primary support comes from the Future of Work at the Human-Technology Frontier program which supports multi-disciplinary research to sustain economic competitiveness, promote worker well-being, lifelong and pervasive learning, and quality of life, and illuminate the emerging social and economic context and drivers of innovations that are shaping the future of jobs and work. Additional support is from the Discovery Research preK-12 program (DRK-12) which seeks to significantly enhance the learning and teaching of science, technology, engineering, and mathematics (STEM) by preK-12 students and teachers, through research and development of innovative resources, models, and tools.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.