The Education Policy Initiative connects researchers and policymakers—helping make good use of their one-of-a-kind secure data clearinghouse
By Daniel Rivkin
A recent research project examined the ties between school funding and crime in Michigan. The study showed a correlation between educational inputs like changes in school funding, outcomes like elementary reading proficiency, and the likelihood that young people would become "court-involved."
Elucidating those relationships required a complex process of matching different kinds of databases to each other: state and local budgets, state-level education attainment scores, and criminal records. How could researchers do it?
That's the type of challenge the Ford School's Education Policy Initiative (EPI) tackles on a regular basis.
Scholars from around the U.S. tap EPI for assistance with data gathering on a broad range of educational issues, thanks to its management of the Michigan Education Data Center (MEDC), a secure data clearinghouse that includes 19 different datasets tracking school enrollment, funding, assessments, coursework, graduation, and disciplinary records. Among the most comprehensive and unique databases are those reflecting postsecondary and early childhood information.
MEDC represents one of the few "P-20" longitudinal data systems in the country built to support education research. "P-20" refers to Preschool to Postsecondary (grade 20). While many states have data spanning P-20, MEDC goes far beyond this by consolidating the data into a single data warehouse with common student identifiers, strict variablenaming practices, and intentionally designed populations.
Kyle Kwaiser, EPI data architect, notes, "This removes many of the barriers researchers customarily face when linking administrative data across time and educational systems."
Other recent research topics using MEDC data and services have given deep insight into some of the most pressing policy issues. What are the educational implications of homelessness? How did the Flint water crisis affect educational outcomes? What are the disparities in access to career and technical training? How has Michigan's HAIL free tuition program performed in its target population? And significantly, what are the effects of the COVID-19 pandemic on school enrollment? These studies have real policy implications.
EPI is in position to facilitate for investigators because of its deep relationships with education experts in Michigan across an amalgam of acronyms: MEDC is part of the Michigan Education Research Institute (MERI), which is a collaboration between U-M, Michigan State University, Michigan Department of Education (MDE), and the Center for Educational Performance and Information (CEPI), often with involvement of the Department of Technology, Management and Budget (DTMB).
The EPI data team helps researchers frame their questions, points them to the appropriate MDE program office, and aligns them with the type of information they will need. Once a data use agreement has been established, access to that data needs to be secure, and the MDE and DTMB review the final output to ensure no personally identifiable information has been exposed. In that way, EPI acts as a service provider for both the research community and the MDE.
In this way, EPI helps promote research that aligns with the MDE's Top 10 Strategic Education Plan, and provides metrics that help evaluate the plan's goals, including: expand early childhood learning opportunities; improve early literacy achievement; improve the health, safety, and wellness of all learners; expand secondary learning opportunities for all students; increase the percentage of all students who graduate from high school; increase the percentage of adults with a post-secondary credential; increase the numbers of certified teachers in areas of shortage; and provide adequate and equitable school funding.
"We take a hands-on approach with researchers and actively help them navigate the process of requesting data from the state and making sure their projects are aligned with MDE's goals. Because we are researchers too, our goal is to facilitate a process that is friendly to researchers and our state agency partners as well," says Nicole Wagner (MPP '10), EPI's associate director.
Currently, MEDC is providing data to research teams at 22 institutions from across the U.S., who have already published 68 public facing research reports, with more on the way. When MERI was launched in 2018, then Deputy State Education Superintendent Venessa Keesler noted that, "Michigan has long been a leader in research practice partnerships. The MERI partnership launches us into the next generation of this work."
The work of the MEDC also helps fill in necessary gaps and avoid research duplication. Jasmina Camo Biogradlija (MPP ‘09), EPI project manager, says the partnership helps the policymakers at MDE. "We've gotten really good responses from program offices. They will say, ‘This study sounds really great, but it would be helpful if you can add one or two research questions.' And the vast majority of researchers have been open to being as useful as possible to reflect those MDE priorities."
One recent working paper produced by Ford School associate professor and EPI faculty affiliate Katherine Michelmore, with Peter Rich, assistant professor of sociology and demography at Cornell's Brooks School of Public Policy, examined how income and segregated schools drive Black-white education gaps. The study relied on MEDC and MERI data, which allowed them to follow nearly 130,000 Michigan students from kindergarten through college enrollment. It was an example of using student-level education data, and matching it with other factors, including eligibility for school lunch vouchers, neighborhood poverty rates, family background, and school segregation rates.
They conclude that given the same levels of family, school, and neighborhood hardship, Black students would be more likely than their white classmates to complete high school and attend college—reversing current disparities. The authors say their analysis suggests a need for complex policy solutions addressing systemic inequities.
"The MEDC data gave us a unique opportunity to investigate this question," says Michelmore. "It provides a lot of rich information about the contexts in which students grow up, including unique insights into how chronic disadvantage across three different domains (school, neighborhood, family socioeconomic status) is associated with education outcomes over the course of grade school, high school, and beyond."
"We use a slightly different matching algorithm than what is in place for reporting purposes by the State," says Camo-Biogradlija. "That means we can have more flexibility in terms of matching to records in the external datasets, which can be really helpful when researchers want to use data from, for example, the Michigan Department of Health and Human Services (MDHHS), Department of Treasury, or even the Office of Retirement Services."
Other projects have involved data from the Office of Retirement Services, or even voter records.
Importantly, each time MDE data is used, Personally Identifiable Information (PII) needs to be protected. Major security steps have been established in conjunction with U-M's Information and Technology Services' Advanced Research Computing department so that PII can be compared, and then scrubbed from the research findings.
For another current study looking to assess the impact of the state's Tuition Incentive Program, researchers will match PII from MDE, MDHHS, and the Department of Treasury.
"We know this research is getting out there. Our researchers get requests from the Michigan legislature to understand the findings more deeply, and from MDE and the governor's office. We are happy to provide the extra capacity that makes these projects happen," says Wagner.
Meet the experts
Katherine Michelmore
Focus
The social safety net, education policy, labor economics, and economic demography, in particular the efficacy of the Earned Income Tax Credit and its impact on children
Recent research
"Contextual Origins of BlackWhite Educational Disparities in the 21st Century: Evaluating Long-Term Disadvantage Across Three Domains," Social Forces, October 2022.
Kevin Stange, co-director, Education Policy Initiative
Focus
College affordability, enrollment, and persistence
Recent research
"The Pandemic's Effect on Demand for Public Schools, Homeschooling, and Private Schools," EPI working paper, September 2021.
Currently on leave for a one-year assignment with the U.S. Department of Education's Chief Economist.
Megan Tompkins-Stange
Focus
Education policy and philanthropy, focusing on the influence of private foundations on the politics of K-12 school reform
Recent research
"How the Political Economy of Knowledge Production Shapes Education Policy: The Case of Teacher Evaluation in Federal Policy Discourse," Educational Evaluation and Policy Analysis, April 2021.
Christina Weiland, co-director, Education Policy Initiative
Focus
Early childhood interventions and public policies on children's development
Recent research
"Historic Crisis, Historic Opportunity: Using Evidence to Mitigate the Effects of the COVID-19 Crisis on Young Children and Early Care and Education Programs," EPI report, June 2021.
Education Policy Initiative
Research and training to improve education outcomes
Since its founding in 2012, the Ford School's Education Policy Initiative (EPI) has generated and supported rigorous research to help policymakers make evidence-based investments that address institutional and systemic barriers and educational inequities.
An interdisciplinary team of educators, economists, sociologists, statisticians, and data scientists use causal research methods to identify the impact of specific policies, programs, and practices that improve educational outcomes—and determine which do not. They put the research to work in partnerships with community stakeholders, government agencies, and policymakers across the country, to make informed decisions and create better results for students.
Through a distinctive partnership with the State of Michigan, EPI operates the Michigan Education Data Center (MEDC), a secure data clearinghouse that houses over 7.5 million education records. This powerful resource helps researchers answer questions about early learning and literacy, school choice, P-20 pathways, college success and affordability, career and technical education, and much more.
On campus, students gain knowledge and experience through specialized courses, events, and the pre and post-doctoral fellowship programs. The Causal Inference in Education Research Seminar is a cornerstone of EPI's efforts to train the next generation of education policy researchers.
More in State & Hill
Below, find the full, formatted fall 2022 edition of State & Hill. Click here to return to the fall 2022 S&H homepage.