Collaborative research: phase i camel-CN: from errors to outcomes – indicators of procedural and conceptual understanding to improve student learning

The National Science Foundation Division of Behavioral and Cognitive Sciences awarded the Strategic Education Research Partnership Institute $316,500 on July 1, 2026, under the Social, Behavioral, and Economic Sciences program (CFDA 47.075) to compile a comprehensive database of student learning and outcome data for mathematical content in grades 4–10.

The project develops AI-enabled assessment tools to analyze student handwritten responses to math problems, automatically identify problem-solving strategies, and diagnose errors. The research leverages previously collected data on student learning from over 3,500 students and 700,000 student work samples covering arithmetic computation, algebraic equation-solving, word problems, and other foundational mathematical skills. The project team aims to link error analysis to learning and motivation outcome data and develop user-friendly tools for mathematics teachers and researchers to better understand mathematics learning and education. Research products are made available in the Open Science Framework Data Repository to facilitate data re-use by other researchers. The award is structured as a Project Grant with no sub-awards planned. Performance runs through June 30, 2028, with place of performance in Washington, DC.

Abstract

This project compiles a comprehensive database of student learning and outcome data for mathematical content in grades 4-10. These data come from projects that implemented successful interventions to support student math learning and include fine-grained data on accuracy and strategy errors over a wide range of math concepts and grade levels. Project outcomes are intended to support new discoveries about student learning and the design and testing of new instructional innovations for teaching mathematics and supporting student long-term success. Research products made available in the open science framework data repository can facilitate data RE-use by other researchers. The project leverages previously collected data on student learning of a broad range of foundational mathematical skills, including arithmetic computation, algebraic equation-solving, word problems, and many other skills from over 3,500 students and 700,000 student work samples. The research team aims to (1) develop AI-enabled assessment tools to analyze student handwritten responses to math problems, automatically identify the problem solving strategy, and diagnose errors made in the process; (2) link to learning and motivation outcome data; and (3) develop user-friendly tools designed to support mathematics teachers in real-world classrooms and researchers working to better understand mathematics learning and education. 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.- subawards are not planned for this award.