Definitional and coverage incompatibilities encountered while building the baseline. Per §2's method, this list is itself an output.
| # | Incompatibility | Consequence | Disposition |
|---|---|---|---|
| 1 | Census PEP state file (SC-EST2024) covers 50 states + DC only — Puerto Rico and territories are separate files with different structures | The committed denominator understates the §-scope population; territories were flagged in the protocol scope line and §6 | Add PR/territory pull in next increment or descope explicitly per §6 |
| 2 | Protocol §1.4 bands are deliberately fuzzy (0–~3, ~3–5, ~5–12); the pipeline cuts at 0–2 / 3–4 / 5–12 | Age-3 children (3.69M) sit in "preschool" here but are infant/toddler in some programs (EHS runs to 3; many pre-K programs start at 4) | Fixed cut documented; sensitivity check when program-specific analysis needs it |
| 3 | PEP civilian estimates vs. ACS resident population differ slightly in level and reference date | Sub-1% differences at these ages, but mixing sources across tables would introduce silent inconsistency | PEP is the single denominator source for all baseline tables; ACS used only for characteristics (income, employment) with the difference noted |
| 4 | NHES-ECPP "nonparental care" ≠ "paid care"; published toplines do not isolate payment | H2.1 cannot be adjudicated from toplines | SIPP/NSECE microdata pass scheduled |
| 5 | Licensed FCC home counts available for only 39 states (2019–2022 comparison) | National FCC supply is a floor, not a count; trend direction reliable, level is not | Reported as "39-state" figures, never as national totals |
| 6 | NIEER counts state-funded pre-K only — district- and city-funded programs (e.g., locally funded seats) fall outside | State pre-K spending/enrollment is a floor on public preschool activity | Note carried on every NIEER-sourced figure |
| 7 | Spending streams have mixed vintages (FY2025 appropriations vs. SY2023–24 NIEER) | Summing across streams mixes years | Single-vintage normalization scheduled for the complete inventory |
| 8 | ECPP arrangement categories overlap (children in multiple arrangements); shares sum >100% | "Slots ≠ children" — §2's multiple-arrangement warning confirmed in the data structure | All arrangement figures reported as participation rates, never summed |