STTR Phase I: Design a Just-In-Time Formative Assessment Algorithm for an Adaptive Education Platform

Award ID: 2451599

The broader/commercial impact of this Small Business Technology Transfer (STTR) Phase I project will be achieved by developing and validating cutting edge assessment technology to address two critical problems in math education: a) student learning deficiencies and b) teacher overload and attrition. At the high school level, the Program for International Student Assessment (PISA) reported that the math scores of U.S. students in 2022 ranked 28th among 37 participating countries, posing substantial risk to the nations competitiveness in STEM fields. This occurs while educators burnout and attrition is at an all time high. Such significant learning deficiencies in math will cost an estimated $1.1T in GDP due to the loss of workforce productivity and innovation. Meanwhile, available tools and innovations for high school math are drastically low, in comparison to tools available for their K-8 counterparts. In response, this STTR project will develop a web-based system providing highly efficient, personalized, formative assessments that are easily customizable by teachers themselves. By addressing critical classroom and market needs, the project will help improve student math learning, cultivate a competitive and diverse STEM workforce, and contribute to high-tech innovation in a Federal Opportunity Zone in the heart of Midwest. This Small Business Technology Transfer (STTR) Phase I project will develop a web-based formative assessment system providing highly efficient and personalized assessments that are easily customizable by teachers themselves, empowering teachers to do their work more effectively and efficiently. Unlike traditional adaptive assessment systems that often reduce the teachers role, this platform leverages cognitive diagnostic modeling to identify students strengths and weaknesses in high school math in real time, both individually and collectively. An innovative machine learning algorithm clusters students for targeted instruction based on their mathematical competencies and current understanding, while also tracking their progress to enable timely interventions. Teachers can regularly and flexibly regroup students based on updated assessments, ensuring that instruction remains tailored to each classs needs. Additionally, advancements in large language models (LLMs) will be utilized to expand the item bank, supporting the platforms scalability and meeting ongoing assessment demands in diverse classroom environments. The systems usability and effectiveness will be validated through a comprehensive pilot study, demonstrating its potential to enhance educational outcomes and streamline teaching processes. 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.

NSF Program Director: Lindsay Portnoy