Administrative data is an important data source for answering policy questions in early care and education
Administrative data is a powerful and often underutilized source for answering important policy-related early care and education questions such as:
- How available are high-quality ECE programs in communities where many young children are covered by the Children’s Health Insurance Program? Are there enough high-quality ECE programs to meet the need?
- How many high-quality ECE programs that participate in the child care subsidy program are located near homeless shelters that serve families with young children?
- What are the most frequent child care licensing violations for each type of ECE program (e.g., centers, family child care homes) within a state? Do violations vary for programs that do and do not participate in a quality rating and improvement system?
- How many children with disabilities who receive child care subsidies are served by high-quality ECE programs?
A new resource from the Child Care Administrative Data Analysis Center (CCADAC) highlights administrative data sources that can answer these and other policy questions. For the last few years, Child Trends has led the CCADAC—funded by the Office of Planning, Research, and Evaluation—and its efforts to increase the use of administrative data to address policy-relevant questions in ECE. CCADAC defines administrative data as the data about services, providers, families, and children that are regularly collected as part of operating a program. They are sometimes the best—or only—data that can answer policy-relevant ECE questions like those noted above. These data are also relatively low-cost because they have already been collected, allowing researchers to find answers to questions without collecting any new data.
Even though a preponderance of administrative data are available, they are not always analyzed to address pressing questions. In part, this is because of the challenges in accessing the data and developing a team of people who, collectively, understand the program, the data, and the methodological and analytic techniques to use them (no one person is likely to hold all this expertise). CCADAC has developed resources, presented at conferences, and hosted an online discussion forum to support peer-to-peer learning about using administrative data in ECE research.
Through this work, we’ve learned a few things:
- There is a strong interest among researchers and state administrators in using administrative data to answer important questions. We’ve seen this in the 125+ people who have joined the administrative data online discussion forum, as well as the increased number and array of resources about working with administrative data.
- There are challenges in using administrative data. It can be difficult to get data from older systems: The quality of the data varies, and there may not be written documentation about what is included in a dataset. It also can be time-consuming to develop data sharing agreements to allow non-program staff to see and use these data (although some helpful tips and examples do exist).
- Research using administrative data is most successful when conducted in the context of a partnership between researchers and the agency that owns the data. Sometimes agencies have access to in-house researchers, while at other times they must partner with external researchers. It’s critical, though, to include people who know the program and data, as well as people who have the research expertise to analyze the data. Working together, teams of researchers can identify the most pressing questions and the best analysis plan, and develop appropriate interpretations of the findings. Across the country, there are examples of such partnerships in ECE, including child care research partnership grants funded by the Office of Planning, Research, and Evaluation that have supported research collaborations using administrative data.
- There are growing opportunities to answer important questions using administrative data. More complex questions, like the previous examples, require combining multiple types of administrative data (e.g., data about child care, pre-K, health services). An integrated early childhood data system makes it easier to answer more complex questions because the data have already been combined. A 2018 report identified 22 of 50 states as having an early childhood integrated data system (ECIDS) that combines program site-level data, with another 16 states in the process of developing one. Recent efforts to add more data to ECIDS, such as home visiting data, will allow teams to use linked data to address an even broader set of questions.
In the future, we hope that ECE policymakers and researchers will work together to use administrative data to answer pressing policy questions. We’d like local, state, and federal policymakers to consider how they can use administrative data—the data they already have—to answer key questions and implement strategies to maximize the usefulness of these data (e.g., develop instructions for entering data, minimize missing data or blank data fields). We also want researchers to consider administrative data as an important source of data to include in their work—and to partner with policymakers to co-create ideas and design studies to address pressing, policy-relevant questions that will advance the field’s knowledge of ECE.