AIMS Mini-Grants
AIMS Collaboratory teams who wish to collaborate can apply for mini-grants to explore promising projects. Mini-grants are intended to increase cross-team collaborations to create new research and/or development insights. These funds are awarded if proposals will have a clear benefit for the Collaboratory and for all those engaged in R&D to improve math teaching and learning more broadly.
Explore the AIMS mini-grant projects below, along with their collaborating partners and deliverables. Projects still underway are marked In Progress.
How We Work Together
Secure Data Enclave-In-A-Box Adoption Needs Assessment
This project explored how K-12 schools and districts might adopt “Enclave-In-A-Box,” an extension of the SafeInsights secure data enclave infrastructure that allows researchers to study learner data without the data ever leaving a school's chosen edtech platform. Working with the AIMS Data Infrastructure Working Group, the team engaged a diverse cross-section of the AIMS community—researchers, educators, learners, edtech providers, school districts, and policymakers—to surface the appeals, barriers, and open questions around adopting this privacy-preserving approach to education data access.
Through small-group sessions and follow-up interviews, the project built a shared understanding of what secure enclaves offer, cataloged existing solutions and lessons learned around data privacy in research, and laid the groundwork for broader community adoption of privacy-preserving research infrastructure.
Enhancing Math Instruction Through AI-Driven Resources for Teachers
This project centered teacher voice in the design of “next steps” instructional resources—a set of AI-driven, actionable guidance teachers receive to respond to student thinking in math class. Through empathy interviews with teachers across partner districts, the team explored what motivates consistent, effective use of such resources, what makes them feel actionable in the classroom, and how context (such as scope, sequence, and ease of integration) shapes what teachers need.
The findings inform concrete recommendations for making instructional resources more usable and impactful, with an eye toward advancing equitable, responsive instruction for priority student populations across the AIMS Collaboratory and the broader field.
Enhancing Research and Development Through Professional Learning
This project developed case studies examining how two AIMS Collaboratory projects integrate professional learning (PL) into their research and development work, drawing on the experiences of the East Baton Rouge Parish School System (EBRPSS) and the NYC Partnership for Math Equity. The work grounds these examples in the current evidence base for effective, equity-focused PL grounded in high-quality instructional materials (HQIM).
Through project team meetings and a cross-network roundtable, the case studies surface recommendations for how AIMS projects—and classroom R&D efforts more broadly—can better connect PL best practices with the sustained use of technology and AI tools in classrooms, and issue a call to action for tool developers, PL providers, and researchers to align their work around teacher and student needs.
Facilitating Research Partnerships through a Common Education Data Sharing Agreement
This project tested the viability of a common, legally reviewed data sharing agreement that would make it easier for districts, edtech solution providers, and researchers to share student data for research purposes. Working with two AIMS development partners (the Lastinger Center and Carnegie Learning), the team piloted a streamlined agreement with multiple schools and districts, engaging privacy law experts and state-level education agency attorneys in at least two states to refine the agreement's language.
The pilot tracked how the agreement moved through approval processes in different district contexts, measuring the time, cost, and effort involved—with the goal of creating a reusable model that can save future partners weeks or months of legal review and accelerate responsible data sharing across the education research field.
Finding Belonging: The Hunt for and Refinement of Measures of Belonging in Math Among Youth
Belonging is increasingly recognized as a key predictor of student success in math, but researchers across the field have lacked a shared, trustworthy way to measure it. This project built a catalog of existing measures of belonging in math, evaluated them for face validity with a diverse review team, and identified gaps in the existing item pool—strengthening the item set so that each sub-scale is balanced and includes appropriately reverse-worded items.
The resulting integrated, multidimensional index of math belonging is designed to be usable across AIMS teams and beyond, reducing the time other researchers spend searching for instruments of uncertain reliability and creating shared language for the field to discuss belonging in math learning.
Setting the Stage to Capture Learning-Related Engagement Behavior on Construct-Specific Instrumentation
Digital learning platforms generate rich clickstream data as students interact with lessons, but it's an open question whether that data reliably signals cognitive engagement—a key predictor of math learning. This project analyzed data from lesson activities focused on double number lines and proportional reasoning across three partner platforms, examining whether patterns of student activity (including whether students provided explanations, regardless of correctness) predict later performance on related assessments.
By focusing on a narrow, highly comparable set of activities across partners, the team was able to rigorously test whether a meaningful engagement signal exists in the data and to lay the groundwork for future instrumentation that could give teachers actionable insight into student engagement during math instruction.
Accelerating Scalable Coaching Models for Mathematics: Automated Feedback Measures for Instructional Routines
This project developed and validated AI-powered automated feedback measures for teachers. The feedback to teachers is aligned with a specific high-quality instructional materials (HQIM) curriculum, in support of scalable, personalized instructional coaching. Building on earlier work co-developing codebooks and annotating classroom recordings, the team tested automated detection of core instructional routines such as “Notice & Wonder” and “Think-Pair-Share.”
The project refined its measures, conducted qualitative validation with educators, and synthesized findings into openly shared resources—drawing on cross-team expertise in educational technology, curriculum design, learning sciences, and discourse analysis to ground the work in educator voice and lay the foundation for scalable, curriculum-aligned coaching tools.
Visualizing Eureka Math² (EM2) Implementation Health Data with the RPPL “Visualizer”
Teachers' success with curriculum depends on much more than the materials themselves—it also depends on factors like school culture, leadership support, and access to aligned professional learning. This project tested whether sharing these systemic “enablers of and barriers to implementation” of Great Minds' Eureka Math² (EM2) curriculum with district and school leaders and teachers, visualized through an Implementation Health Dashboard, can help improve the systems that shape classroom instruction.
The team piloted RPPL's existing Visualizer tool—originally built for RPPL's own English language arts measures—as the platform for this dashboard, integrating EM2-specific survey, observation, and platform-use data from Riverside Unified School District. This marks the Visualizer's first use outside the RPPL network, testing whether it can serve as shared infrastructure for implementation research across different curricula, products, and districts.
Beyond the Numbers: Rethinking Classroom Evidence with Multimodal Data
Traditional measures of teaching and learning—standardized tests and classroom observation rubrics—are resource-intensive and often miss a holistic picture of classroom practice. In this district-led project, a school district partnered with two product organizations to co-implement an AI-enabled assessment tool and an AI-enabled teacher feedback tool in 6th-9th grade math classrooms.
Together, the partners co-designed a prototype tool that generates richer, more holistic measures of teaching and learning; an implementation model for scaled district use; and a multimodal dataset combining district records, student work, and classroom audio. The project demonstrated how two complementary AI tools can save teacher time, produce more actionable and equitable measures of student learning, and inform district-level resource decisions.
Windows into Practice: A Video-Based Professional Learning Library for Elementary Mathematics*
(In Progress)
Classroom video can serve as both a “window” into strong math discourse and a “mirror” for teacher reflection, but high-quality, publicly available videos of elementary math conversations grounded in widely used curricula remain scarce. This project brings together Teachley and the National Training Network (NTN) with elementary teachers and coaches from High Tech High Schools (HTH)—a network serving predominantly students farthest from opportunity—to fill that gap using Illustrative Mathematics (IM) lessons.
Building on an existing partnership that created a public database of tagged student problem-solving work from IM lessons, the team is filming roughly 15 problem-solving lessons across grades 2-5 and producing a tagged video library showing how skilled teachers apply the “5 Practices for Orchestrating Productive Mathematics Discussions” to turn student thinking into rich classroom discourse. NTN is leading the design of companion professional learning tools so districts and researchers anywhere can use the library without specialized training.
Professional Learning at Point-of-Need*
(In Progress)
This project tests whether embedding professional learning materials—specifically Math Language Routines (MLRs)—directly within instructional materials, rather than delivering them separately, changes how teachers use them. The partnership pairs Kiddom's learning platform and Florida Math curriculum with MLR content from the University of Florida's Lastinger Center for Learning, embedding 16 structured routines (short videos, handouts, and facilitation guides) at teachers' exact point of lesson planning and delivery.
Roughly 12 K-8 teachers across two conditions—professional learning delivered through a separate system versus professional learning embedded at point-of-need—are being studied using platform analytics, surveys, and focus groups to see which approach leads to greater awareness, earlier and more consistent use, and better integration of MLR supports into everyday planning.