Since Accelerate launched in 2022, we have funded more than 100 studies, including nearly 40 randomized controlled trials (RCTs), considered the gold standard of research design. Each new set of studies reflects our research agenda, which focuses on gaps in our current understanding of what interventions work best for students, particularly those who are furthest behind academically. Across all these studies, our approach has evolved, reflecting lessons learned in how to manage research, who we partner with, and how we track progress and set expectations for findings.
Accelerate’s first two grantee cohorts, in 2022-23 and 2023-24, represented an array of tutoring providers including traditional nonprofits, districts, community-based organizations, and for-profit entities. Each provider selected their own research partner, a decision that resulted in a number of unexpected challenges.
Our first year of grants launched 13 randomized controlled trials and quasi-experimental studies. Two years later, we had only eight reports in hand.
This underwhelming result prompted us to explore a new direction to approach research. Three challenges in particular informed our thinking:
- Asking tutoring providers to manage research projects. The onus of managing and keeping track of the research project was put on each grantee, many of whom had never managed a research partnership before and were not organized to do this work well. In cases where providers tapped into existing relationships, the researchers were accountable to the provider, creating an unintentional bias towards reporting out on positive results.
- Setting unclear or incomplete expectations. From the beginning, there was often a lack of clear expectations about data collection that frequently delayed projects and caused problems for subsequent analyses. Progress monitoring was also inconsistent over the course of the various studies. Furthermore, there was a missed opportunity for researchers to collaborate and share with each other along the way.
- Under-communicating results. All too often negative and even null results are not publicly shared, are underemphasized, or are anonymized to protect the provider. While there may be motivation from the adults delivering services to pursue one of those paths, districts, schools, and ultimately students need not only thoughtfully designed and executed studies but also transparency around results.
Challenge: Asking tutoring providers to manage research projects
Our Solution: Selecting research partners and matching them with grantees
Beginning in 2024-25, we released a Request for Information (RFI) to a select group of researchers with the experience and reputation to produce objective and thoughtful results. Accelerate matches each grantee with a research partner, who manages the impact evaluation studies and provides technical assistance to grantees. Rather than requiring grantees to manage research (potentially distracting from their core work of delivering tutoring to students), Accelerate contracts with the researcher and holds responsibility for making sure studies are successfully planned and executed.
Running an RFI process and managing these contracts with intentionality also brought costs down since working with Accelerate essentially guarantees the researchers a certain number of studies. To date, our formal partners have included Abt Global, Harvard University, Johns Hopkins University, Mathematica, MC2, SETA-ED, the University of Maryland, the University of Michigan, and WestEd.
An ancillary but notable benefit is that matching grantees with vetted researchers also keeps the evaluation independent of the organization being evaluated. That independence works in providers’ favor. An organization that commissions its own study gets a result it will have to defend. An organization evaluated by a partner it did not select gets something more useful: a credible read on whether the program works, and one that districts, states, and other funders have reason to trust.
Challenge: Setting unclear or incomplete expectations
Our Solution: Pre-analysis plans, progress monitoring, and opportunities for collaboration
Formal partnerships let us create clarity and set terms at the outset: submitting a pre-analysis plan, scheduling regular check-ins, providing ongoing updates, and sharing interim and final reports. Accelerate maintains strategic oversight of research design and implementation throughout the grant period, until the final report is delivered.
The pre-analysis plan lays out the research questions and methodology before data collection starts. Developing the plan is an iterative process, with Accelerate offering input, feedback, and final approval. Each study is then pre-registered on the Open Science Framework, with the registration page public for the life of the project. This guards against a real risk in a field where everyone involved wants the intervention to work: adjusting the analysis until the findings look better than they are.
When applicable, randomization does not happen until the rest of the setup is in place. Research partners secure IRB approval and data-sharing agreements with the participating district first, both as dated deliverables rather than assumptions. District conditions are a recurring reason studies stall, so we would rather find out early.
We also developed the Data Alignment and Tutoring Assessment Standards, an open reporting framework that defines a common language for tutoring records at the session and student level, so that data arriving from different providers can actually be compared.
In addition to monthly meetings with the Accelerate team, our research partners connect with each other quarterly through the Research Learning Community, alongside representatives from Arnold Ventures, the Overdeck Family Foundation, the University of Chicago, and the Walton Family Foundation. These touchpoints keep individual projects on track and build the knowledge base around tutoring more broadly. This approach has fostered collaboration among researchers, facilitated the exchange of expert feedback on original reports and tools, and yielded improvements to the research process itself, including a common cost tool.
Across the randomized controlled trials launched under the 2024-25 cohort, every interim and final report was submitted on or before the agreed-upon deadline.
Challenge: Under-communicating results
Our Solution: Setting expectations for findings
We believe that research should be shared transparently and accurately, because it’s the right thing to do and because the findings matter for our kids. With improved student outcomes as the north star, it is our responsibility and obligation to provide decision makers with unbiased information, even if the results of a study are negative or null.
From the outset of our research projects, we aim to set clear expectations. Payment is structured around research deliverables: an approved pre-analysis plan and submission of the pre-registration link; an interim report; and a final report.
This year, we require research partners to post the final report in an open-access format by a set date, regardless of whether the findings are positive, negative, mixed, or null, and to name the provider, program, or product being studied. We will continue sharing the findings publicly so that decision makers can make informed choices about how to use scarce resources for students most in need of support.
Ensuring transparency continues to be a work in progress. Even with Accelerate’s increased efforts to incentivize unfiltered results, the broader field still has a long way to go.
What Comes Next
Our practices have improved, and we still face challenges. Sometimes, even with the best laid plans and tight management, studies fail for a variety of reasons. A handful of grantees in each cycle have requested no-cost extensions because they couldn’t secure a district partner or establish the conditions for a reliable study. Managing research this closely takes real capacity from our team. We’ll continue improving and aspire to 100% of studies being rigorous and relevant. But we know if we’re being properly ambitious, sometimes there will be misses.
The same questions now apply to AI-powered and educational technology tools, an area where we’ve been making grants since the 2025-26 school year. The fundamentals of our process hold: independent evaluators, studies in real classrooms rather than pilot settings, and sample and evidence standards set in advance. Because AI- and tech-enabled tools change rapidly, decision makers need evidence they can trust, not just hype and sales. That need guides our work to build the evidence base on whether AI-powered personalized learning tools improve student outcomes, for which students, and under what conditions, just as we have done for tutoring.
Kimberly Ueyama is Managing Director of Advancement at Accelerate.
