Application & Evaluation


Our current work on application and evaluation revolves around providing teachers with automated feedback and rigorously assessing whether and how this feedback can improve mathematics teaching and student learning outcomes.

Educators in brick and mortar K-12 classrooms, online settings, and tutoring programs have used the M-Powering Teachers automated feedback tool. By incorporating suggestions from these educators, as well as by utilizing randomized controlled trials to examine how various aspects of the automated feedback impact teachers’ instruction and students’ outcomes, we continue to develop our tool so that it can be as helpful to teachers as possible. For more information on the M-Powering Teachers app, click here.

Through studies funded by the National Science Foundation and the Overdeck Family Foundation, we are also working with instructional coaches to determine the best ways to integrate automated feedback with coaching. We see an opportunity to combine coaching and automated feedback in ways that extend the power of each. Instructional coaching is widely regarded as one of the most promising forms of professional development (Kraft et al., 2018). However, few educators receive such effective feedback on a regular basis. This is because generating formative feedback tends to be resource-intensive (Kraft & Gilmour, 2016), limiting the number of teachers an individual coach may serve. Coaches also face challenges engaging teachers who are reluctant to deprivatize their practice, limiting their reach in many schools. Frequently, coaches are assigned more teachers than they can reach on a regular basis and as a result, they typically engage in a very limited number of coaching cycles with any one teacher. Embedding automated feedback in teacher coaching can solve these complementary challenges and enhance the effectiveness of teacher professional learning. Given that teachers are the single most influential school-based factor in student success, the M-Powering Teachers tool brings the potential to greatly reduce inequities by improving instructional quality.

If you’re interested in learning more about our application and evaluation work or participating in a related study, please contact Research Project Manager Hannah Rosenstein at hrosenst@umd.edu.

Highlighted Publications

The Promises and Pitfalls of Using Language Models to Measure Instruction Quality in Education


Paiheng Xu, Jing Liu, Nathan Jones, Julie Cohen, Wei Ai

arXiv, NAACL 2024, 2024


Empowering educators via language technology


Dorottya (Dora) Demszky, Jeffrey B. Bush, Sidney K. D’Mello, Jennifer Jacobs, Isabelle Hau, Heather Hill, Jing Liu, Susanna Loeb, Bethanie Maples, Kylie Peppler, Rhea Pokorny, Matthew Rascoff, Jenny Robinson, David Yeager, Laura Wentworth

2023


Improving Teachers’ Questioning Quality through Automated Feedback: A Mixed-Methods Randomized Controlled Trial in Brick-and-Mortar Classrooms


Dorottya Demszky, Jing Liu, Heather C. Hill, Shyamoli Sanghi, Ariel Chung

EdWorkingPaper, 2023, pp. 23-875


Resources for Instructional Coaches

The protocols below were developed as part of our NSF sponsored project, to support coaches and teachers with using automated feedback as part of their coaching sessions. They are meant to guide the coaching session by providing with suggested questions and topics for discussion based on the automated feedback. 
Share



Follow this website


You need to create an Owlstown account to follow this website.


Sign up

Already an Owlstown member?

Log in