MDRC Center for Data Insights: Improving Programs and Systems with Actionable Data Science

Overview

The MDRC Center for Data Insights (CDI) empowers partner organizations to unlock actionable insights from their data. CDI strengthens data infrastructure and combines deep policy expertise with AI and advanced analytics to create meaningful data products. We help partners apply these insights to enhance their programs, processes, and services while building internal capacity to sustain evidence-driven practices in the long term.

To bridge the gap between data and practice, CDI takes an interdisciplinary and design-centered approach, facilitating collaboration between data specialists and program staff to ensure that each challenge is fully understood and addressed with the appropriate set of methods. The result? Accurate and perceptive data insights that can be used to improve services effectively.

Our Services

CDI is staffed by professionals in data science, human-centered design, quantitative methods, and data visualization, as well as by subject matter experts on state and local strategy, housing and financial stability, workforce, education, and criminal justice policy. They help organizations use data and evidence to improve programs, policies, and decision-making through services such as:

Help partners better use their data for program improvement

Services

Offer training, mentorship, and resources to help partners sustainably manage and use their data.
Improve data access, infrastructure, and governance practices.
Improve data literacy and analytic capacity among staff.

Projects

Temporary Assistance for Needy Families (TANF) Data Collaborative (TDC) helps state and local TANF agencies expand the use and analysis of administrative data via training, technical assistance, and hands-on support to strengthen data-informed decision-making and improve program administration.
The State Impact Collaborative brings together state agency staff members, MDRC researchers, and Coleridge Initiative data scientists to collaborate on high-quality tests of the effectiveness of programs in education, criminal justice, income support, and employment.
STRIVE Forwardis a partnership to build data capacity and support the expansion of STRIVE’s workforce programs, Career Path and Fresh Start. The project guides program growth, ensures quality implementation, and generates insights to help STRIVE better serve people facing barriers to employment nationwide. A central element of the project is the development of an employer dashboard to help track the quality of job placements.
From Theory to Practice (T2P) is focused on expanding use of administrative data for studying long-term outcomes of employment and human services programs. Building on prior work, T2P developed infrastructure and methods to link evaluation records with administrative data, advancing evidence-based policymaking.

Help partners leverage AI and machine learning to understand how their programs and services work – and translate what they learn into meaningful action

Services

Extract meaningful insights from data and translate them into concrete actions that improve daily practice.
Identify knowledge gaps and translate data into actionable insights for program improvement using tools like funnel analyses, theories of change, and logic models.
Use diagnostics to foster transparency about service disparities and potential bias
Employ iterative learning cycles to drive continuous program improvement

Projects

SkillUp NavGPTtests a generative AI–powered career navigation tool that provides personalized guidance to workers, particularly those with low incomes or without college degrees. The project assesses how well the tool engages users, improves access to training and jobs, and strengthens career outcomes, offering some of the first rigorous evidence on the role of generative AI in workforce development.
Improving a Chatbot to Increase Financial Aid Applications (OTTERS) MDRC partnered with the Washington Student Achievement Council to study and improve OtterBot, an AI-informed chatbot that helps high school students in Washington navigate financial aid and college applications.
Partnership with the Center for Employment Opportunities (CEO) for Predictive Analytics is harnessing granular, longitudinal administrative data to build a system for ongoing, advanced analytics that support CEO’s continuous improvement process.

Help partners prepare for rigorous impact and program evaluation

Services

Optimize existing programs to build readiness for an evaluation.
Support or implement rigorous impact or quasi-experimental design studies.
Support sequence analysis, social network analysis, and place-based studies.
Develop effective data-driven visualizations and dissemination materials.

Past Projects

Using Generative Artificial Intelligence for Reading (U-GAIN) R&D Center, led by Digital Promise, explores how generative AI can improve elementary reading instruction for English learners. The project combines AI-driven content, interaction, and culturally responsive approaches with research-based tutoring tools to enhance reading fluency and comprehension, with MDRC assessing learner outcomes to guide development and measure impact.
New Jersey Criminal Justice Reform Advancing Racial Equity (NJ CARE) evaluated New Jersey’s 2017 pretrial justice reforms, which replaced money bail with a risk-based system. The mixed-methods study used an interrupted time series design to assess the reforms’ impacts on racial equity across pretrial practices, explore local factors shaping outcomes, and incorporate the voices of people with lived justice system experience to understand broader implications for health and well-being.
The Washington College Grant Impact Evaluation assesses the implementation and effects of Washington state’s need-based financial aid program on postsecondary access and outcomes for students from low-income backgrounds. Using state longitudinal data, the study examines program uptake and student outcomes while building the Washington Student Achievement Council’s capacity for ongoing, data-driven evaluation.

Help partners innovate, expand their technical capacities, and implement advanced analytics

Services

Scale interventions using evidence.
Make sense of large volumes of complex data.
Harness data science, predictive analytics, machine learning, and generative AI.

Projects

New Visions for Public Schools (NVPS) Researcher-Practitioner Partnership for Predictive Analytics is a partnership in which MDRC researchers developed and implemented a comprehensive predictive modeling framework that allows for rapid and iterative estimation of a continuous measure of risk for each student.
Using Machine Learning to Predict Third-Grade Reading Proficiency is a project with the Wake County Public School System. MDRC researchers are applying a machine learning approach to predicting third-grade reading proficiency, using DIBELS reading tests and subtests.
Sequences to Success is an MDRC-led project, in partnership with the Washington Student Achievement Council and the Washington Education Research & Data Center, that analyzes the postsecondary and workforce pathways of young adults in Washington state. Using state longitudinal data and advanced statistical methods, the project maps common trajectories—including dual-enrollment and nondegree pathways—and examines their labor market outcomes.
Subprime Lending Data Exploration Project, a “big data” project funded by the MetLife Foundation, was designed to produce policy-relevant insights using an administrative data set that includes nearly 50 million individuals who have applied for or used subprime credit.

Build capacity across the data analytics field by contributing to the knowledge base

Many of CDI’s projects pilot proof-of-concept approaches that offer real-world models for federal agencies, policymakers, foundations, and other funders interested in investing in efficient, long-lasting approaches to improving data analytics capabilities. CDI shares lessons from its work on the MDRC website.

Projects

Machine Learning for Investigating Variation (MLIV) explores the use of machine learning to identify heterogeneous program effects across participant subgroups in education interventions. By analyzing multiple MDRC randomized trials, the project aims to uncover complex patterns of impact, evaluate the effectiveness of machine learning methods, and produce tools and guidance for researchers studying subgroup variation.
The Long-Term Outcomes Project is an effort to produce new findings from older studies using new matching techniques and approaches. The project is assessing the feasibility of linking administrative data sets to program evaluation records, a promising and potentially low-cost means of tracking the long-term impacts of social interventions.
Job Search Support in Sector Programs is an MDRC partnership with Per Scholas to explore how sector training programs can better support participants in finding employment. By studying alumni experiences and labor market shifts in the technology sector, the project aims to generate insights for strengthening job search services during and after training.

How We Work

CDI provides cutting-edge data science tools, technical assistance, and research. CDI employs human-centered and interdisciplinary data analytics approaches to help organizations unlock insights hidden in existing data. In other words, the CDI’s data analytics approaches are complemented by participant perspectives and other modes of inquiry, such as qualitative research, policy analysis, and behavioral science. CDI also uses an agile approach to project management to ensure that our work is aligned with our client’s needs, quickly adapts to changes in project requirements, and delivers high-quality results. CDI builds capacity for organizations seeking to make better use of their data; creates replicable prototype code and materials that many organizations can use; acts as an innovation hub; and disseminates our work to inform the larger field. With our commitment to excellence, CDI strives to empower organizations to make data-driven decisions that benefit low-income individuals, families, and communities.

CDI offers three collaboration models:

Consultation collaborations: partners define the focus and scope of the research agenda and reflect on the results of the data work produced by the CDI team.
Cocreation collaborations: program staff members share leadership roles with the CDI team by learning data coding and documentation best practices on the job.
Coaching collaborations: the partner organization leads the work, while the CDI team provides training, mentorship, and resources to promote the sustainable management of data projects

Who We Help

CDI is dedicated to supporting government agencies, institutions, nonprofits, and other organizations in serving individuals, families, and communities with low incomes.