A State Scan of Early Learning Assessments and Data Systems

Early learning assessments are a critical source of information. Educators can use assessment data to tailor instruction, families to better understand their children’s development, and state and local leaders to make policy and resource allocation decisions. We interviewed 53 state partners to gather information about their assessment and data practices. The information highlighted in this overview is not exhaustive or necessarily meant to serve as an exemplar, but rather is meant to provide examples of work states are doing in a few fundamental areas related to pre-k assessments and data systems. We hope this resource can assist states and spark ideas about possibilities for continuing to improve these systems.

How are early learning assessments selected?

Note: Hover over each state in the map above for a detailed description.

• In Maryland, assessments are locally determined. Use of the Kindergarten Readiness Assessment and the Early Learning Assessment, both developed by the state, is encouraged, but providers determine if they want to administer the tools and which components of the tools to use.
• In Wisconsin, assessments are locally determined. District and community-based providers must demonstrate that the tools are aligned with the early learning standards framework.
• In Michigan, comprehensive, authentic assessments are selected from an approved list. The curriculum review committee regularly assesses potential child assessment tools. Approval decisions are made using a rubric, which is tightly connected to the state's early learning standards. Those standards hit all of the academic and non-academic domains for children, as well as standards for the classroom level.
• In Minnesota, statute requires that Voluntary Pre-Kindergarten programs use Kindergarten Entry Profile (KEP) approved assessments, which must be aligned with the state’s early learning standards. School Readiness Plus programs are required to use KEP-approved assessments as well. Both programs are required to collect entry and exit data, at a minimum. Assessment requirements vary by early learning program.
• In North Carolina, assessments are selected from an approved list. Providers must use tools approved by the North Carolina Child Care Commission, which must be aligned with the state’s early learning standards.

Does the state require specific assessments for providers that participate in the Quality Rating and Improvement System (QRIS)?

• Ohio’s QRIS has five levels. At level two, programs identify a comprehensive child-level assessment they plan to implement and complete any required training to administer it. At level three and above, providers must administer a comprehensive child-level assessment to all enrolled children.
• Rhode Island providers must demonstrate they administer a child assessment if they reach level three on the QRIS. Standards for use of child assessment become more rigorous as providers move up QRIS levels.

Are any specific assessments required by the state?

• Tennessee developed student growth portfolios to serve as a growth measure for children in grades and subjects that are not included in the Tennessee Value-Added Assessment System. All districts or charters serving children through Tennessee Voluntary Pre-K must use the pre-k/kindergarten student growth portfolio or an alternative growth measure (a state board approved universal reading screener). Data from the portfolios can be included in VPK teachers' annual evaluations.
• Indiana requires that all On My Way Pre-K providers administer the Kindergarten Readiness Indicator.
• Virginia requires, as of 2021, that all publicly supported pre-K classrooms participate in the Virginia Kindergarten Readiness Program (VKRP). VKRP includes the Phonological Awareness Literacy Screening, Early Mathematics Assessment System, and Child Behavior Rating Scale assessments.

Does the state have any specific considerations for assessments for dual language learners?

• California’s Desired Results Developmental Profile (DRDP) assessment is designed so that educators can observe children’s progress across multiple domains in both their home language and English. Because children learn English in different ways and at different rates, the four English Language Development (ELD) measures are related to the child’s experiences with English and not the child’s age. This approach allows educators to develop a comprehensive profile of each child’s skills and behaviors. California originally developed the DRDP, but it is now used in other states, such as Minnesota.

Does the state require or offer any training on early learning assessments? If so, what is the structure of the training?

• Indiana requires training and certification for Kindergarten Readiness Assessment (KRA) assessors. The process is free and offered through the Indiana-Licensing and Education Access Depot (I-LEAD). Training is approximately 90 minutes and is available asynchronously or live. The certification process is approximately 30 to 45 minutes and conducted through a one-on-one virtual appointment where the assessor-in-training will use the KRA to assess the meeting host.
• Mississippi requires the administrator of each school district pre-K program to complete the BRIGANCE training webinar. BRIGANCE is a screening tool used to identify children for delays and giftedness. The 45-minute webinar includes an overview of the Online Management System and information on how to prepare to screen, score a screener, interpret results, and identify the next steps.
• New Jersey offers in-person and online training courses on using curriculum and assessments. This includes courses on Teaching Strategies GOLD, Classroom Assessment Scoring System, and Child Observation Record (COR) Advantage Assessment.
• Ohio requires training for Early Learning Assessment (ELA) assessors. The three trainings are only open to educators employed in licensed early childhood programs. The ELA initial training is designed for those who will be observing, assessing, and scoring the ELA. This 10-hour training requires participants to attend three virtual sessions, complete on-demand modules, and pass a simulator and content knowledge test. Two additional optional trainings are available. The ELA training for support teachers is designed for assistant teachers, paraprofessionals, and educational aides. This training gives an overview of the ELA and how to document observations. The ELA technology training is for individuals who have been trained to use the ELA but need to be trained to use the technology supports in the Ready for Kindergarten Online system. Individuals will be required to pass the simulator and content knowledge test during this training.
• The Virginia Kindergarten Readiness Program (VKRP) team provides online and in-person training on how to use and navigate VKRP and how to interpret VKRP data. There is also a Train-the-Trainer program that produces VKRP affiliate trainers who can teach others about the VKRP system.

Does the state have a data system that houses early learning data? What type of data are in the system and which agencies’ data feed into the system?

• Georgia’s Cross Agency Child Data System (CACDS) stores data from programs and services for children birth to five and their families. The system serves as a single repository for early childhood data from Georgia’s Department of Early Care and Learning, Department of Education, Department of Public Health, Department of Human Services, Division of Family and Children Services, and Head Start Association. CACDS provides information to help state leaders and researchers better understand how well programs are serving children and families. Programs contributing data to CACDS include Early Head Start, Head Start, Preschool Special Education Individuals with Disabilities Education Act (IDEA Part B), Childcare and Parent Services, Georgia’s Pre-K, Rising Kindergarten Summer Transition program, Babies Can’t Wait (IDEA Part C), Children 1st, and Georgia Home Visiting. Pre-K child-level and classroom participation data as well as de-identified child-level data and provider-level data are stored in CACDS.
• The New Jersey Enterprise Analysis System for Early Learning (NJ-EASEL) is the state’s early childhood integrated data system (ECID). Planning for NJ-EASEL, a cross-agency collaboration designed to collect and integrate early childhood data to measure the impact of early care and education programs on the children and families they serve, began in 2014, and production began in November 2018. When fully operational, the system will store information from 15 data systems regarding child, family, classroom, program, and workforce characteristics. The complete system will draw on data from the state Department of Education, Department of Human Services, Department of Children and Families, and Department of Health. It will provide longitudinal data to assess outcomes and the effectiveness of early learning programs. It will also help the state make informed business, fiscal, and policy decisions to provide better program delivery, and access to, early care and education.
• Pennsylvania’s Enterprise to Link Information for Children Across Networks (PELICAN), combines the state’s early learning programs into a single information management system. PELICAN includes data on provider access to and participation in a range of social services. The systems that make up PELICAN include: child care provider licensing, Child Care Works (low-income subsidized child care), Keys to Quality (quality rating and improvement), Pre-K Counts (state-funded pre-K and state-funded Head Start Supplemental Assistance Program), Early Learning Network, Early Intervention, Provider Management, and Provider Self-Service. Data include early learning provider/program information as well as parent/caretaker and/or child information for the early learning programs previously mentioned. The Pennsylvania Office of Child Development and Early Learning (OCDEL), a joint project of the Departments of Education and Human Services, is currently underway to link general information about children participating in OCDEL’s home visiting/family support programs to the PELICAN data.
• Wisconsin’s Early Childhood Integrated Data System links data from the state Department of Children and Families (DCF), Department of Health Services (DHS), and Department of Public Instruction (DPI). DCF shares child welfare, child care subsidy, TANF, and child support data. DHS shares birth record, maternal and child health, immunization, early hearing detection and intervention, lead screening and remediation, and birth to three developmental delay data. DPI shares student demographic data, attendance, enrollment, retention, suspension, English proficiency, and third-grade math and reading assessment data.

What type of early learning data does the state collect?

• Texas's early childhood data system houses both kindergarten and prekindergarten student information (student demographics, classroom information, and assessment data). Data reporting is mandatory for public pre-K and kindergarten programs. These data are used to provide information about the effectiveness of early childhood education programs to families, school administrators, community stakeholders, and policymakers.
• Ohio collects Early Learning Assessment (ELA) data from preschool special education programs and state-funded pre-K programs. The ELA data is used to monitor trends, understand children’s growth, and measure progress for two literacy grants.
• Louisiana collects Teaching Strategies GOLD data three times a year and holds public pre-K programs accountable for completing the assessment and uploading the data into the performance dashboard.
• South Carolina’s First Steps 4K program collects Teaching Strategies GOLD data. The data is used to evaluate programs, prevent duplicative services, and assign unique identifiers to link children to their longitudinal data records. Legislation requires First Steps 4K to include how children are doing on the Teaching Strategies GOLD assessment in the annual reports.

Does the state use a standard student identifier?

• Alabama generates a unique identifier for children receiving child care subsidies and for children receiving Part C/Part B 619 services.
• Florida assigns a unique identifier to children in Voluntary Pre-K, a state-funded program. This allows the Office of Early Learning at the Department of Education to access an unduplicated count of children and better understand factors that affect early childhood education participation.
• Louisiana contracted with a vendor to create and maintain the statewide unique identification system, eScholar. Students served, and staff members working, in K-12 schools that serve children from birth through grade 12 are assigned a unique identifier. The unique identifier system allows school systems to assign identifiers and also acts as a central matching and validation hub for the data system. The identifier remains unchanged across time, location, and source. The Department is exploring ways to expand the unique identifiers to children served in mixed delivery early childhood settings (independent Head Start Centers and/or publicly funded seats in child care sites participating in the state accountability system).
• New Mexico uses the eScholar Uniq-ID as the foundation for its Early Childhood Integrated Data System. The identifier tracks children’s progress over time and as they move among programs, schools, and locations.
• South Carolina is required by law to create a student unique identification number for children in the state-funded pre-K program, which is operated by the Department of Education and First Steps. The state is working on getting identifiers for children served in other programs, such as home visiting and parenting groups.

How does the state fund its early learning data system?

• Georgia began its development of an integrated child care data system in 2010, using a grant from the State Advisory Council. In 2013, Georgia used the Race to the Top–Early Learning Challenge grant to create the infrastructure needed to link data across state agencies. The CACDS Policy Manual outlines the system’s goals, contents, governance, rules, and structure. Georgia also received a grant from ECDataWorks at the University of Pennsylvania to design a data visualization tool to improve the delivery of early childhood data.
• Rhode Island has used multiple federal grant opportunities for the Executive Office of Health and Human Services Data Ecosystem. The Ecosystem was initially funded by a State Innovation Models (SIM) Test Grant in 2017. In 2019, the Ecosystem began receiving a 90/10 match for all costs from the Centers for Medicare and Medicaid Services (CMS). Since then, the state has used a variety of funding sources to cover the 10 percent match, including the Preschool Development Grant B–5 (PDG B–5) and COVID-19 (CARES) stimulus funds. In particular, the state used PDG B–5 funds to expand early care and education data assets, develop data dashboards, and track children’s participation in state-funded pre-K, Head Start, Early Intervention, and other subsidized child care programs. The Ecosystem is currently funded by CMS, state general revenue appropriation, the sale of data products, and a Centers for Disease Control and Prevention (CDC) health disparities grant.
• South Carolina is using the Preschool Development Grant B–5 and Statewide Longitudinal Data Systems Grant to build its early childhood integrated data system. The five goals of the system are to inform policies, create standard definitions and reports, identify and resolve disparities, connect early childhood data to longitudinal data systems, and understand the impacts of investment from birth through workforce.

What type of technology does the state use to maintain and hold its early learning data?

• Iowa transmits data from partners through Cybox, Iowa State University’s Box.com service, which is a secure, cloud-based file storage system. The data are then automatically transferred to secure internal hosting in the data management platform.
• Rhode Island hosts data in a cloud-based data warehouse. Data are linked using a custom-built matching algorithm and master person index. The data team uses R, Python, and other analytic resources to conduct exploratory work.

Perspectives from state agency staff

Thank you to the state and local leaders who contributed valuable insights and information to this project. We appreciate Abby Thorman, Alicia Miao, Amanda Williford, Amy Poirier, Beth Meloy, Bridget Cullen, Charlie Greir, Chelsea Richard, Crys O'Grady, Dan Harris, Deb Wise, Erika Anderson, Ginny Vitiello, Herman Knopf, Jan Hume, Janice Cole, Jeannie Allen, Jessica Whittaker, Jon Vaupel, Judy Walker, Karen Powell, Karissa Palmer, Kayla Rosen, Keli Houston, Leslie Doyle, Lillie Moffett, Lisa Brochard, Lisa Hildebrand, Lisa Howard, Marina Merrill, Martha Stickland, Michael Bock, Michelle Levy, Morgan Orr, Nicole Norvell, Olivia DeMarais, Pamela Truelove-Walker, Pat Sargent, Richard Lower, Sarah Nardolillo, Shelly Jackson, Sophia Hubbell, Steven Hicks, Susan Adams, Tiffany Wynn, Tonya Coston, Tonya Williams, and Wendy Uptain for sharing their expertise. Any errors are the authors’ alone. Thank you to Sabrina Detlef for editorial insight. Thank you to Julie Brosnan, Fabio Murgia, and Riker Pasterkiewicz for their help in the production of this brief and to Aaron Loewenberg and Cara Sklar for managing this project.

This brief is based on research funded in part by the Bill & Melinda Gates Foundation. The findings and conclusions contained within are those of the authors and do not necessarily reflect positions or policies of the Bill & Melinda Gates Foundation.