Oklahoma Child Wellbeing

Prevention-Focused Accountability & Early Warning Platform

Verified Public Aggregates & Separate Simulated Demonstration

Data Sources & Methods

Explaining how child well-being indicators would be collected, organized, protected, and used for coordinated action.

OK Child Wellbeing is an independent pilot. The verified public-data dashboard displays official published aggregates; the separately labeled demonstration dashboard uses simulated examples. A real implementation would require formal data partnerships, privacy review, data-sharing agreements, role-based access, and compliance with FERPA and other applicable laws.

Current Status: Verified Aggregates and Simulated Examples

The verified view displays cached public Census estimates, CDC adult community context, USDA classifications and published OSDE agency counts. These are dated releases, not live individual records or operational referral tracking.

Current Data Status

Official aggregates in the verified view retain their source periods, geographic definitions, uncertainty and suppression. Case examples and simulated school snapshots in the demonstration view remain fictional. Ten warning-sign domains are selectable: one has verified outcome evidence (child poverty); nine need further outcome integrations.

Verified statewide IDEA publications · Stage B2

Independently retrieved November 7, 2026 from official OSDE report indexes. Both PDFs identify Oklahoma, FFY 2024, and a February 2, 2026 due date. A due date is not a publication date. HTTP Last-Modified metadata was May 18, 2026 for Part B and June 4, 2026 for Part C; neither establishes first publication. Fixed publication records are reviewed releases, not a live feed. Verification means source and arithmetic checks, not an audit of child records.

Part B · Connected publication records

Part B FFY 2024, PDF pp. 56–57: Indicator 11 has 21,399 timely evaluations among 21,698 parental consents after federal exclusions; 21,399 ÷ 21,698 × 100 = 98.62% rounded. Oklahoma uses 45 school days from parental consent, not the default 60-day wording. There are 299 outside the timely numerator. This is not all referrals or first delivered services.

34 CFR §300.301(d) exceptions are removed from both numerator and denominator: repeated parental failure/refusal to produce the child or qualifying transfer between agencies before determination. No other state exceptions are allowed. Excluded-case counts are not given here, so a raw pre-exclusion rate is unavailable. LEAs certify EOY aggregates. The report describes the entire reporting year and school year 2024–2025; exact inclusive extraction dates, carryovers and deduplication are unconfirmed.

Indicator 12(a), PDF pp. 59–60: 1,638 children served in Part C and referred to Part B for eligibility determination. This is a count only, with no percentage denominator. It includes categories later excluded from compliance. It does not demonstrate completed transitions.

Part B Indicator 12 rate · Hold retained

The adjusted formula is c ÷ (a−b−d−e−f), excluding timely noneligible determinations, parental-refusal/federal exceptions, late Part C cases and the extended Part C option. The report’s category e changes from Part C eligibility in the instructions to Part C referral in the table. The published rate and its component outcome counts remain withheld from connected records pending authoritative reconciliation. No raw referral-completion percentage is substituted.

SoonerStart Part C Indicator 7 · Hold retained

Part C FFY 2024, PDF pp. 29–30: eligible infants/toddlers evaluated and assessed for whom an initial IFSP meeting was required; initial evaluation, assessment and IFSP meeting within 45 days from referral. The cohort is April 1–June 30, 2025, not the full July 1, 2024–June 30, 2025 collection year described on p. 2. Documented exceptional-family-circumstance delays are added to the numerator and retained in the denominator. That adjusted compliance is not raw timeliness.

Staff-delay incidents do not reconcile with the remaining cases implied by the table, and printed quarterly proportions do not sum to a complete year. No Part C Indicator 7 count or percentage is released. Existing Part C first-service and exit-total holds remain. No agency clarification or corrected report was obtained, and no agency was contacted.

Stage B2 accountability and future aggregate review

The developmental domain now separates nine pathway measures, statewide submeasures, source-native geographic availability and rural/urban comparison readiness. A verified compliance or referral submeasure does not establish a complete pathway. The aggregate import specification defines county FIPS, reporting periods, population, raw and adjusted counts/rates, suppression and provenance. Its review-only records cannot publish themselves or replace the existing pinned publications. No child-level upload or public import endpoint exists.

Geography, disclosure and missing outcomes

Only published statewide aggregates without suppression symbols are connected. No local suppression threshold is inferred from these PDFs. No county allocation, district conversion, complementary-cell reconstruction, pooled program rate or population-based denominator is used. County referral cohorts, referral-to-first-service completion, rural/urban completion, unverified eligibility and waiting-period outcomes remain DATA NOT YET CONNECTED. Existing Stage B1 participation and geography records keep their own definitions and dates.

Dashboard statewide IDEA section · Independent verification and delivery record

SoonerStart (IDEA Part C) public data · Stage B2 acquisition

Checked November 9, 2026. All 21 PDFs linked from the official SoonerStart Data and Reporting page were downloaded (HTTP 200) and hashed (SHA-256). Twenty were transcribed from their text layers; the 2024–2025 site determinations PDF is image-only and was not transcribed.

Reporting periods
SPP/APR FFY 2022–2024 (each July 1–June 30), with FFY 2017–2021 percentages republished in the FFY 2022 report. Indicators 1, 7 and 8 use an April–June sample of each year. Child counts are December 1 point-in-time counts (2021–2023). District profiles 2022–23 and 2023–24; site profiles 2021–22.
Geography
Oklahoma statewide; 12 SoonerStart service districts (2022 onward); 27 local sites (2021–22). No county-of-residence data are published. District results are never divided among counties.
Verification method
Each value was matched to its cited PDF page. Published percentages were recomputed from published counts, and totals were reconciled across reports (for example, district child counts against statewide Section 618 counts, Indicator 2 against the child count, and Indicator 3 exits against the exiting reports). Adjusted compliance (with documented family-circumstance delays added) is kept separate from the raw timely share.
Limitations and quality holds
These are compliance and results measures as published. They do not show referral-to-first-service completion. Profile percentages have no published counts. The following are withheld: FFY 2024 Indicator 7 (existing hold) and Indicators 8B/8C (conflicting denominators), FFY 2024 exits (totals do not reconcile), 2023–24 district birth-to-one counts, 2021–22 site Indicator 2, Outcome B2 and child counts, and profile rows whose printed "met target" flag contradicts their values. No suppression symbols appear in these publications; race-by-setting and race-by-reason cells were not imported.

Full acquisition and verification report · Dashboard section

Community Conditions · Phase 7 public data integration

Checked November 9, 2026. Nineteen official sources were retrieved (all HTTP 200) and hashed (SHA-256): OSDE state test and enrollment files, OESC and BLS labor-force data, and OSBI domestic abuse and crime reports. The oklaschools.com download tool requires a per-file reCAPTCHA check; it was inventoried but not automated, so graduation, chronic absenteeism, attendance, growth, dropout and report-card grades remain DATA NOT YET CONNECTED.

Reporting periods
Education: OSTP grades 3–8, school years 2022–23 to 2025–26; October 1 enrollment 2003–04 to 2025–26 (state) and 2022–23 to 2025–26 (district). Employment: annual averages 2016–2025 (2025 is an 11-month average excluding October); statewide monthly January 2016–September 2026, seasonally adjusted and not seasonally adjusted; county monthly July 2025–August 2026, not seasonally adjusted. Domestic abuse: calendar years 2021 (incident total only) to 2024. Crime: calendar years 2022–2024 (incident-based reporting only).
Geography
Statewide; all 77 counties (LAUS, place of residence); 4 MSAs; 561 OSDE school districts and, for 2025–26 test results, 1,229 schools; 467 OSBI reporting agencies and 77 OSBI county totals of agencies located in each county. District and school results are never assigned to county residents. OSBI county totals are sums of agency reports with OSBI’s covered population, not Census county estimates. Domestic abuse data are statewide only.
Verification method
Published values were located on their cited page or row. Employed + unemployed was checked against the labor force, rates were recomputed, OESC annual averages were checked against the BLS monthly series, and the county workbook against BLS county series. Test proficiency is Proficient + Advanced ÷ valid N from published counts and is cross-checked against the separate state results files. Crime category totals, offense groups, victim and arrestee breakdowns, agency-to-county sums and county-plus-statewide-agency totals were all reconciled with the state total. 2,972 automated checks passed; the 5 that did not are the documented discrepancies below.
Limitations, suppression and quality holds
Cells OSDE printed as *** are labelled SUPPRESSED BY SOURCE and never reconstructed. OSDE printed some charter district identifiers in scientific notation (e.g. 5.50E+04); those rows were excluded, not repaired. On quality hold: 2023–24 OSTP results (performance levels not comparable with adjacent years) and the OSBI live portal’s 2024 figures, which differ from the published 2024 report. Unemployment is not poverty or hardship. Reported crime and reported domestic abuse are not prevalence; arrestees are not convictions. Domestic abuse is community context, not child exposure. No rate per 100,000 is calculated because no matching population denominator is connected.

Full acquisition and verification report · Dashboard section

Developmental context · Source verification

Initial source verification: November 6, 2026 (Stage B1). The methods below describe that verification, not a new certification or a claim that these releases are current-day observations. Runtime retrieval and last-check dates appear in the feed responses.

USDA RUCC 2023 county classifications

Official source and publication

Reporting period
2023 release
Geography
All 77 Oklahoma counties
Verification method and date
November 6, 2026: Automated fixed-release parsing verified all 77 unique county FIPS identifiers and valid codes.
Limitations, suppression and quality holds
County metropolitan/nonmetropolitan context is not individual rural residence, tract RUCA or Census urban/rural residence. No child or service outcome is inferred. No small-cell suppression applies to classification codes; incomplete or invalid coverage blocks publication.

USDA RUCA 2020 census-tract classifications

Official source and publication

Reporting period
2020 release
Geography
1,205 Oklahoma 2020 census tracts across 77 counties
Verification method and date
November 6, 2026: Automated parsing checked primary and secondary codes and exact GEOID agreement with Census 2020 TIGERweb tract geometry.
Limitations, suppression and quality holds
Commuting context is not Census urban/rural residence or a service-access measure. Generalized geometry is for display only. No small-cell suppression applies to codes; missing, duplicate or mismatched tract records hold the layer unavailable.

Oklahoma SoonerStart Part C service areas

Official source and publication

Reporting period
2023–2024 Site Data Profiles
Geography
District 11: Tulsa; District 4: Craig, Delaware, Mayes, Nowata, Ottawa, Rogers, Wagoner and Washington
Verification method and date
November 6, 2026: Manual review of District 4, PDF pages 7–8, and District 11, pages 21–22. Source spelling “Deleware” was normalized to Delaware.
Limitations, suppression and quality holds
County memberships are not clinic locations, capacity or individual access. Shared service-area outcomes are not allocated to counties. Disputed FFY 2024 compliance denominators and exception adjustments, and inconsistent exit totals, remain on quality hold. No disputed percentages or exit-derived rates are published.

Oklahoma OSDE FY 2025 IDEA Part B agency counts

Official source and publication

Reporting period
FY 2025 / school year 2024–2025; October 1 snapshot per workbook notes
Geography
543 published education agencies; ages 3–21, not a county-of-residence population
Verification method and date
November 6, 2026: Automated workbook consistency checks reconciled visible Child Count by District names, native identifiers and totals against profile backing rows: 539 numeric totals and four suppressed totals. Population definitions and Notes were manually reviewed.
Limitations, suppression and quality holds
Published participation is not referral completion, preschool-only counts or an under-18 measure. Native CoDist IDs are not a verified NCES crosswalk. The source’s opaque Lakeside identifier remains unchanged. Preserve * for cells below 4 or complementary suppression; no reconstruction. Agency age splits, county rates and individual school measures remain unconnected.

Official Census 2020 tract geometry · OSDE population definitions · SoonerStart reports and reporting years

Cached feeds and manual verification

The existing /api/public-data feed serves saved Census ACS child-poverty and renter-housing estimates and CDC PLACES adult mental-distress context. Scheduled source checks validate definitions, geography and publication status. The browser reloads saved data every 15 minutes while visible; it does not refresh official sources. Permitted older snapshots may appear with stale warnings after ordinary failures; suppression, withdrawal or unexplained missing records trigger an unavailable hold.

/api/disability-context caches fixed-release USDA RUCC 2023, ACS 2024 five-year B09001 estimates and published 90% count margins of error, and OSDE FY 2025 Part B counts. /api/ruca-geography caches USDA RUCA 2020 classifications joined to Census 2020 tracts. Both use source snapshots in Netlify Blobs, isolated between production and previews. A cache miss or age above 24 hours triggers source retrieval and validation; failed rechecks hide expired context. This is read-through verification, not a daily publication or background schedule. New vintages require review.

SoonerStart service-area descriptions, cited PDF pages, program definitions and quality-hold interpretation are manually verified page content, not an automated SoonerStart outcome feed. Cached checks do not resolve disputed compliance or exit statistics. Referral, evaluation, eligibility, service initiation, transition and rural completion remain DATA NOT YET CONNECTED.

Census household age subsets are not all-resident denominators. Published margins of error remain count uncertainty, not percentage points. County populations are never assigned to education agencies, and no unmet-need subtraction, rural service rate or child-level inference is calculated. See the verified dashboard methodology for existing Census/CDC indicator definitions and safeguards.

Future Outcome Integrations May Include

1

Education Data

  • School attendance
  • Chronic absenteeism
  • Reading benchmark status
  • Math benchmark status
  • Graduation readiness
  • Credit risk
  • Behavioral referrals
  • Suspensions
  • School mobility / transfers
2

School Health and Well-Being Data

  • Nurse visits
  • Counselor referrals
  • Mental health referral patterns
  • Health-related barriers to attendance
  • Developmental or well-being concern trends
3

Family Stability Indicators

  • Housing instability
  • Homelessness / McKinney-Vento indicators
  • Transportation barriers
  • Food insecurity indicators
  • Family support needs
  • Service referral patterns
4

Safety and Protective Factors

  • Domestic violence exposure indicators
  • Protective order trends involving households with children
  • Mandated reporting patterns
  • Aggregate child safety concerns
  • Community violence exposure indicators
5

Community and Workforce Indicators

  • Poverty rates
  • Employment and workforce participation
  • Childcare access
  • Transportation access
  • Neighborhood-level resource availability
  • Career and technical pathway access
6

Partner Response Data

  • Assigned follow-up actions
  • Referral completion status
  • Partner action pending
  • Response timelines
  • Case management status
  • Outcome improvement tracking

Indicator Method

The pilot organizes indicators into three connected layers: child outcome indicators, risk driver indicators, and system response indicators. This structure is designed to show not only what is happening to children, but whether responsible systems are responding in a timely and coordinated way.

1

Child Outcome Indicators

Measures connected to child development, school success, health, safety, and long-term well-being.

  • Reading proficiency
  • Attendance
  • Behavioral well-being
  • Health concerns
  • Graduation readiness
2

Risk Driver Indicators

Conditions that may contribute to poor outcomes or signal that a child or family needs support.

  • Housing instability
  • School mobility
  • Domestic violence exposure
  • Poverty
  • Transportation barriers
  • Family stress
3

System Response Indicators

Measures showing whether the system acted after risk was identified.

  • Assigned lead partner
  • Family contact completed
  • Referral made
  • Support plan active
  • Follow-up completed
  • Outcome improved or unresolved

Privacy and Legal Safeguards

Public-facing dashboards should only show aggregate, de-identified information. Any internal case management or school-level implementation would require strict access controls, legal review, and formal agreements.

✓FERPA compliance
✓Role-based access
✓Data-sharing agreements
✓De-identification standards
✓Aggregate public reporting
✓No personally identifiable student or family information
✓Privacy review before implementation
✓Clear limits on who can view, edit, export, or act on data

Public vs. Internal Data

Public Dashboard

  • Aggregate trends
  • School or district-level summaries where authorized
  • De-identified indicators
  • Public accountability metrics
  • Policy and funding trends

Internal Authorized System

  • Role-based case coordination
  • Assigned response actions
  • Referral follow-up
  • Partner status tracking
  • Restricted school-level or student-level support workflows only if legally authorized

Methods Limitations

This pilot is not currently an official government, school district, healthcare, or child welfare data system. Demonstration data should not be used to make legal, clinical, educational, or policy conclusions. The purpose is to show the model, framework, and accountability structure that could be implemented through authorized partnerships.

Implementation Requirements for Real Deployment

This is currently a demonstration pilot. Moving from demonstration to real implementation would require formal governance, legal review, and institutional authorization. The following requirements would need to be met before individual-level operational data could be connected to this model.

✓Official data-sharing agreements with partner agencies and districts
✓FERPA and privacy review for all education-related data
✓Role-based access controls limiting who can view, edit, export, or act on data
✓Aggregate public reporting with no personally identifiable information
✓Formal data governance policies
✓Local partner approval from schools, agencies, and community organizations
✓Clear methodology documentation for each indicator
✓Defined update schedules for data refresh and reporting cycles
✓Protection against exposing personally identifiable student or family information

These requirements are standard practice for responsible data-driven accountability systems. They are listed here to demonstrate the pilot's commitment to credible, privacy-protective implementation should the project move beyond the demonstration phase.

Credibility & Governance

This demonstration pilot would require official partner approval, data-sharing agreements, privacy review, FERPA compliance, role-based access controls, data governance policies, clear indicator methodology, defined update schedules, and aggregate public reporting before any live implementation.

The verified dashboard does not collect or display individual child or family records. Published education-agency identifiers and aggregate counts are displayed separately, with source suppression preserved.

Why This Page Matters

Credible child well-being work requires more than compelling visuals. Policymakers, funders, schools, and families need to know where data comes from, how it is protected, what it can and cannot prove, and how it is connected to action. This page establishes the pilot's commitment to transparency, privacy, and responsible use of data.

Verified public aggregates and illustrative examples are labeled separately. Neither establishes individual needs or intervention effects.