State of Childcare Index — Methodology | Beverly Research

State of Childcare Index — Methodology

Beverly Research · Version 1 · April 2026 · 250 cities, 50 states

Beverly Research — 2026 State of Childcare Report
STATE OF CHILDCARE INDEXMethodology

What the index measures

The index is a 0-100 composite score that captures how well childcare works for families in a given US geography. Higher is better. A score of 100 would mean childcare is widely available, affordable, supports decent worker pay, doesn't crush family finances, and is backed by strong public investment. A score of 0 would mean none of those things are true.

The score is comparable across cities, across states, and across years (annual cadence: May 2026, May 2027, etc.).


The 5 dimensions and weights

# Dimension Weight What it captures
1 Affordability 30 pts How much of a typical family's income childcare consumes
2 Supply 25 pts Whether there are enough licensed slots for the children who need them
3 Workforce Health 15 pts Whether childcare workers can earn a living wage
4 Family Strain 15 pts The downstream burden on parents (mothers' workforce participation, single-parent share)
5 Policy Support 15 pts State-level public investment (pre-K access, subsidy reach, paid leave)

Weight rationale: Affordability and Supply are weighted highest because (a) they're the two pain points families and journalists cite most consistently, and (b) they're the dimensions with the cleanest, most defensible data. Workforce, Strain, and Policy are equal supporting dimensions. The 30/25/15/15/15 split is also clean enough to explain in a single sentence to a journalist.


Metric spine — what feeds each dimension

Dimension 1 — Affordability (30 pts)

Metric Source Granularity available Notes
Center-based infant care annual price as % of median household income DOL Women's Bureau NDCP (county) ÷ ACS B19013 (city) County-level price → city-level income The single most-cited childcare statistic.
Center-based toddler care annual price as % of median household income Same Same Captures cost differential by age.
Childcare-to-rent ratio (annual center infant care vs. annual median gross rent) NDCP ÷ ACS B25064 Same Captures the "more than rent" framing without leading on it.

Dimension 2 — Supply (25 pts)

Metric Source Granularity available Notes
Licensed childcare slots per 100 children under 5 with all parents working childcaregap.org (state/county, Sept 2025 release) ÷ ACS B23008 County-level The "childcare gap" headline metric; the Buffett/BPC/CCAoA standard.
Number of licensed childcare establishments per 1,000 children under 5 BLS QCEW (NAICS 624410) ÷ ACS B01001 County-level Independent supply measure.

Dimension 3 — Workforce Health (15 pts)

Metric Source Granularity available Notes
Childcare worker median hourly wage (BLS SOC 39-9011) as % of MIT/EPI living wage for a single adult, no children BLS OEWS (MSA) ÷ EPI Family Budget Calculator MSA-level → city via MSA mapping A city scores better when local wages can actually sustain the workers.
Estimated share of childcare workers below the federal poverty line ACS B24010 + occupation crosswalk County-level Cross-check on wage adequacy.

Dimension 4 — Family Strain (15 pts)

Metric Source Granularity available Notes
Mothers' labor force participation rate (women 25-44 with kids under 6) ACS B23008 City-level Lower LFP = childcare may be pushing mothers out.
Single-parent share of households with kids under 18 ACS B11003 City-level Single parents bear disproportionate childcare burden.

(Parental stress / mental health proxies are not city-granular enough — addressed in the narrative chapter, not the index.)

Dimension 5 — Policy Support (15 pts) — STATE-LEVEL, INHERITED BY CITIES

Metric Source Granularity available Notes
State pre-K access score (% of 4-year-olds enrolled in state pre-K) NIEER State of Preschool Yearbook 2024 State-level Inherited by cities in that state.
CCDF subsidy reach: % of eligible children receiving subsidies HHS/ACF State-level Inherited.
Paid family leave (weeks of paid leave for new parents) DOL + Bipartisan Policy Center State-level Inherited. 0 if no state program.

For city reports, the Policy Support score = the city's state's score, with a footnote: "Policy Support is measured at state level. [City]'s policy environment reflects [State]'s public investment in childcare."


Scoring methodology

For every metric, every geography (city or state) gets a percentile rank within its cohort: - 250 cities form one cohort → city percentile rank - 50 states + DC form one cohort → state percentile rank - National = absolute value (no rank; described in narrative)

For metrics where higher is BETTER (e.g. supply, wages, LFP): percentile_score = (rank ÷ cohort_size) × 100

For metrics where lower is BETTER (e.g. cost as % of income, single-parent share): percentile_score = ((cohort_size - rank + 1) ÷ cohort_size) × 100

So all metric scores are 0-100 with 100 = best.

Within each dimension, the metric scores are simple-averaged → dimension score (0-100).

The index = weighted average of the 5 dimension scores, using the weights above.

Dimension and scores are rounded to whole numbers (presented as "73" not "72.6") for press readability.


Tier labels for press use

Tier score range Plain-English label
Excellent 80-100 "Childcare works here."
Strong 65-79 "Above-average childcare access."
Moderate 50-64 "Mixed picture — some dimensions strong, others strained."
Strained 35-49 "Families face significant childcare pressure."
Crisis 0-34 "Childcare is functionally broken for most families."

These tiers give journalists a clean shareable phrase. Any city, state, or national rating can be expressed in one sentence: "Phoenix scored 38/100 — Strained — in the 2026 State of Childcare Index, ranking #198 of 250 US cities."


Ranking conventions

Each city report shows two ranks: - National rank — out of all 250 cities - State rank — out of all cities in this state with a score (often just 1-15 per state)

Each state report shows: - National rank — out of 51 (50 states + DC)

The national report shows aggregate score for the US plus dimension breakdowns (no rank — there's only one US).


What's deliberately NOT in the index (and why)


Implementation notes