Improving Data Improving Outcomes How Can Partnerships with

“Improving Data, Improving Outcomes” How Can Partnerships with Higher Education Help Your State Agency Use Early Childhood Data for Decision-Making? Robert L. Fischer, Ph. D. , Claudia J. Coulton, Ph. D. , & Seok-Joo Kim, Ph. D. Center on Urban Poverty & Community Development Jack, Joseph and Morton Mandel School of Applied Social Sciences Case Western Reserve University Cleveland, Ohio September 16, 2013; Washington, DC

Overview • State-wide resource in Ohio (Ohio Educational Research Center) • Local data system in Cuyahoga County (Cleveland) • Leveraging existing data to answer new questions • Recommendations for pursuing this kind of work 2

Overview Educational Data Projects from State to Local. Level Area Project Implementation • Education projects • Collaboration with partners State Ohio OERC County Cuyahoga CHILD system • Database for children • Geographic analyses Projects (examples) I. Health care II. Homeless family III. 3 rd Grade reading* *OERC project Local Cleveland Researcher 3

State: The OERC The Ohio Education Research Center (OERC), is a network of Ohio-based researchers and research institutions, that develops and implements a statewide, preschool-through-workforce research agenda to address critical issues of education practice and policy. • Provide timely and high quality evaluation & research products • Maintain a research data base • Bridge needs, research, practice & policy • Bring together resources to improve access to knowledge 4

State: The OERC Current Projects Standards / Assessments Teachers & Leaders Future. Ready Students STEM Education Initiatives Ohio Education Research Center Improvement & Innovation Improving with Data State Success Factors Early Childhood Education Investigating the pathway to proficiency from Birth Cleveland, OH through 3 rd grade Cleveland, Ohio 5

County: CHILD system The Need for Integrated Data. • Data helps inform our understanding of the early childhood system • Individuals and families interact with multiple systems and services, so integrated data offers a more complete view of reality [“Big Data”] • Understanding of how systems work and how to better meet existing needs can be informed by integrated data • Service models emphasize long term and collective impact, so data needed across services and over time 6

County: CHILD system Concept. th Bir t. Cer • Teen births • Low weight birth • • Public Assists Ch Me ild di Dat cal a ID 1 Medicaid Food Stamp TANF Child care voucher ID 5 Ser • Home visiting • Special needs child care • Early childhood mental health • Universal pre-k ID 2 Child. Hood Common Integrated ID Data Longitudinal (CHILD) System ID 6 vic es • Infant mortality • Elevated Blood Lead ID 3 ID 4 Public School • • • Attendance KRA-L Proficiency test Graduation test Disability ld Chlitreat Mament • Abuse/neglect reports • Involvement with ongoing services 7

County: CHILD system Structure. REPORTS Data files-Births, Home Visiting, DCFS, UPK, KRA-L, Medicaid, etc. Geocode & Standardize Geographic E. g. % LBW births receiving ongoing home visits by neighborhood Longitudinal Master Files for Each Data Source IDS Registerincludes ID#’s, names, addresses, DOB, etc. Match New Records to IDS Register Time Trends e. g. Total Children Served by birth cohort Updated IDS Register-includes ID#’s, names, addresses, DOB, etc. Profiles E. g. Birth characteristics & service use for children entering kindergarten Outcomes E. g. Kindergarten Readiness Scores among children in UPK program 8

Geographic Analyses County District 2 (2008) County District 8 (2008) Cuyahoga County (2008) Births 1, 443 1, 877 16, 246 # Teen Births, mother’s age 10 -14 (per 1, 000) 2 (1) 12 (2) 42 (1) # Teen Births, mother’s age 15 -19 (per 1, 000) 124 (39) 358 (79) 2, 031 (41) % Mothers without High School diploma 14% 32% 19% % Low Birth Weight 9% 14% 10% % Premature Low Weight Births 6% 9% 7% % Mothers w/adequate prenatal care 52% 42% 53% % Mothers w/out prenatal care 1% 2% 1% % Healthy Births 53% 36% 49% 10 (7) 29 (15) 164 (10) Indicators # Infant Death (per 1, 000 births) 9

Cleveland Metropolitan School District Profile Kindergarten 2008 -9 Cleveland Cuyahoga County % Teen Births, mother’s age 10 -14 <1 <1 <1 % Teen Births, mother’s age 15 -19 22. 4 16. 7 9. 8 % Mothers without High School diploma 41. 7 30. 2 15. 9 % Low Birth Weight 12. 6 11. 6 9. 4 % Premature Low Weight Births 8. 7 8. 2 6. 7 % Mothers w/adequate prenatal care (Kessner Index) 63. 1 69. 4 81. 3 % Mothers w/out prenatal care 1. 9 . 9 % Health Births 56. 4 61. 5 70. 9 % Children with a substantiated or indicated report of abuse/neglect by age 4 12. 1 9. 6 5. 1 % Children referred to ongoing services with Child & Family Services by age 4 19. 8 14. 7 7. 6 % Children with any report of abuse/neglect by age 4, including substantiated and unsubstantiated 35. 2 26. 7 14. 7 % Children in households receiving Food Stamps in 2008 76. 9 51. 1 28. 8 % Children in households receiving Cash Assistance in 2008 19. 0 11. 3 6. 1 Indicators 10

Data Influence Examples 1) More children have access to health care via public insurance, but are they using it? 2) How are homeless families involved with child welfare services? 3) What children will be most impacted by the State’s 3 rd Grade reading Guarantee? 11

Local Example I: Child Health Summary. • Dramatic increase in health insurance coverage for children ages 0 -6 in the county: Hooray! • But only 43% of children get all the recommended well-child visits in the first year of life: Oh no! • Data show that 49% of these families were involved with supportive services close to birth, so we can use that connection to reach families: Hooray! • But wait, due to data lags and coordination issues, outreach would happen too late to have an effect: Oh, no! • A preventive approach could be adopted by having dedicated staff at clinics reach out to families… • Result o Medical Home Pilot launched at two health clinics; 86% of families completed scheduled well-child visits, double the rate for children born on Medicaid in Cuyahoga County; one clinic has integrated the model into care with 9 patient advocates serving the needs of families with infants 12

Local Example II: Homeless Families Summary. • County undertaking social impact bond approach to social services o Fund preventive services that pay for themselves through lower use of later high-cost services • Focus on homeless families who are also involved with child welfare services o High-costs associated with of out-of-home placements and shelter stays • Found that 30% of women in shelter had children involved with welfare agency o 52% of these women had no children with them in shelter o 25% of their children were in a foster care placement • County developing strategies to intervene with mothers before they become homeless and to intervene when mothers enter shelters 13

Example III: 3 rd Grade Reading Study Significance. • Importance of early childhood exposures o Early exposure to stressful circumstances, environmental hazards, and less than optimal early learning environments negatively and persistently affect early development. • Usefulness of longitudinal data • State adopted ‘ 3 rd Grade reading Guarantee’ to ensure that students pass reading proficiency test before advancing beyond 3 rd grade • Districts can project how many of their students will be held back when the policy is implemented • What is less understood is o What early childhood factors best predict the students who will be impacted by this policy? o What early childhood interventions appear to lessen the odds a child will not attain third grade reading proficiency? 14

Example III: 3 rd Grade Reading Cohort Design. Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Cohort 1 B K 3 rd Collected Cohort 2 Recently collected Will be collected B Cohort 3 K B Cohort 4 3 rd K B 3 rd K 3 rd 15

Example III: 3 rd Grade Reading Conceptual model • Abuse/Neglect • Out-of-home placement • Access to well-child care Child Welfare Medical • Newborn home visit • Help Me Grow • Out-of-home • Mom’s First child care Home Visits Child Care K-3 Outcomes K Birth • Birth weight • Maternal risk • Housing distress Family Economic • Cash assist/ Poverty • Food insecurity Pre-K • Public preschool • Universal Pre-K Pilot Nhood / Residence • Nhood condition • Housing distress • Residential instability 1 st 3 rd • KRA-L • STAR Early Literacy • NWEA MAP • OAA • Benchmark Assessments 16

Example III: 3 rd Grade Reading Current Process • Sample (N=3, 679): Children who took KRA-L in 2007 & 2008 and 3 rd grade proficiency test in 2010 & 2011 in Cleveland Metropolitan School District, OH. • Sample and variables will be updated. Educational Information % Demographic / Welfare / Neighborhood % Pass of 3 rd grade readting test 55. 7 Girl 49. 7 KRA-L band 1 (Score 1 -13) KRA-L band 2 (Score 14 -23) KRA-L band 3 (Score 24 -29) 38. 1 44. 6 17. 3 Hispanic African-American Other race White 10. 6 69. 3 4. 3 15. 8 Below 11% of attendance at Kindergarten 29. 7 TANF + (Medicaid or SNAP) at Kindergarten Medicaid or SNAP at Kindergarten No assistance at Kindergarten 17. 3 67. 4 15. 3 Living a census tract with poverty rate above 30% at Kindergarten 49. 4 (Substantiated or indicated) maltreatment before Kindergarten 17. 5 Reported disability before 3 rd grade 14. 5 17

Example III: 3 rd Grade Reading Implications. • Collaboration with Cleveland Metropolitan School District o Data Sharing o Uses - Building profiles - Community collaborative planning - Risk factor reduction • Helpful to establish educational planning; especially schools with large numbers of disadvantaged students • Understand challenges for 3 rd grade guarantee 19

Discussion Data into Practice Observations… • Data don’t make policy… People with data make policy • Policy shapes research • Everyone wants outcomes… few want to pay for them (or pay very much) • Great divides need to be bridged in terms of institutional practice and philosophy 20

Discussion Ongoing Challenges for Integrated Data. • Data inclusion decisions o Relevance o Continuity o Correct geography • Data usage issues o Data access o Data quality o Data linkage 21

Discussion Recommendations. • Identify what data exist and in what form it exists; consider partnering with universities in this work • Become familiar with relevant federal and state laws and policies regarding data sharing/use • Convene interested parties – data holders and data users – to discuss the opportunities to learn from integrated data • Pilot data matching procedures to demonstrate how specific questions can be answered 22

Discussion Funding Prospects. • Institute of Education Sciences has funding work to integrate data related to young children • US Department of Education Race to the Top funds can be used for longitudinal data systems using integrated data • Various federal funding opportunities exist for studies that could develop and draw on integrated data systems • Mac. Arthur Foundation very interested in use of integrated data 23

Thank you! Q/A State County Local Contact Information: Robert Fischer, Ph. D. (fischer@case. edu) Resources • Ohio Education Research Center: http: //oerc. osu. edu/ • Center on Urban Poverty & Community Development: http: //povertycenter. case. edu/ • NEO CANDO: http: //neocando. case. edu/ 24