Methodology

Single-parent household share

How the figure for each Maryland county is derived from the U.S. Census Bureau, why this definition was chosen, and how accurate it is.

What we’re measuring

For each Maryland county, the share of households with own children under 18that are headed by a single parent — that is, a male or female householder with no spouse present, raising their own children. This is sometimes shortened to “single-parent households” on the charts, but the denominator is specifically households with children, not all households.

The formula
single-parent % = (S1101_C03_005E + S1101_C04_005E) / S1101_C01_005E
  • C01_005ETotal households with own children of the householder under 18 years (the denominator).
  • C03_005EMale householder, no spouse present, family household, with own children under 18.
  • C04_005EFemale householder, no spouse present, family household, with own children under 18.

All three columns are pulled from American Community Survey Table S1101 (Households and Families), 2024 5-Year Estimates. Grab the raw data fresh from data.census.gov.

Why this definition

The Census Bureau doesn’t publish “single-parent household share” as a single field. It can be derived from S1101 in several ways, each measuring something slightly different:

  • Single-parent / households-with-kids (what we use) — among households actually raising children, what fraction are headed by a single parent. This is the figure most directly relevant to a school district, because the denominator is the families whose children attend those schools.
  • Single-parent / all households — includes households without children (single retirees, empty-nesters, etc.). This dilutes the signal because two counties with identical family structures could differ wildly on this measure just because one has more retirees.
  • Children in single-parent homes / all children — counts children, not households. A close proxy but typically a few points higher than the household-share figure because single-parent households tend to be smaller on average.
Data vintage

The figures are from the 2024 ACS 5-Year Estimates, released in December 2025. A 5-Year Estimate is a rolling average of survey responses collected over 2020–2024.

The Census Bureau publishes 5-Year Estimates rather than 1-Year Estimates for all geographies (including small counties) because individual year samples for places like Kent or Somerset don’t have enough respondents to produce statistically reliable estimates. The trade-off: any single year’s real-world shift takes a few years to fully appear in the rolling average.

How accurate is it?

Every ACS estimate ships with a 90% margin of error. For a derived ratio like this one, the margin of error of the percentage is propagated from the margins of error on the numerator and denominator using the Census Bureau’s standard ratio-MoE formula:

MoE(X/Y) = (1/Y) × √( MoE(X)² − (X/Y)² × MoE(Y)² )

Interpretation: a county listed as 20.0% ± 2.4 ppmeans we’re 90% confident the true value lies between 17.6% and 22.4%. For the smallest counties (Kent, Somerset), the margin is wide enough that the reported value should be treated as approximate; for larger counties (Montgomery, Prince George’s) it is tight.

The table below shows the 90% margin of error for every Maryland county. Sorted by margin of error ascending so the most reliable estimates are at the top.

CountySingle-parent %90% MoE90% CISample (HH w/ kids)
Montgomery County23.5%±1.1 pp22.424.6%119,879
Anne Arundel County26.6%±1.6 pp25.028.2%65,202
Howard County20.8%±1.7 pp19.122.5%40,923
Prince George's County41.6%±1.8 pp39.843.4%86,728
Baltimore County34.9%±1.8 pp33.136.7%85,108
Frederick County24.3%±2.2 pp22.126.5%34,602
Carroll County20.0%±2.4 pp17.622.4%19,680
Harford County24.9%±2.5 pp22.427.4%29,020
Baltimore City57.7%±2.7 pp55.060.4%47,986
Calvert County18.3%±2.8 pp15.521.1%10,981
Charles County28.2%±3.3 pp24.931.5%19,692
Washington County39.3%±3.6 pp35.742.9%16,001
St. Mary's County28.7%±3.8 pp24.932.5%13,714
Queen Anne's County24.7%±4.5 pp20.229.2%5,517
Worcester County35.8%±4.9 pp30.940.7%4,636
Cecil County33.1%±5.1 pp28.038.2%10,903
Wicomico County42.6%±5.4 pp37.248.0%10,009
Allegany County41.5%±5.6 pp35.947.1%5,906
Caroline County44.7%±6.3 pp38.451.0%3,447
Garrett County30.5%±6.5 pp24.037.0%2,970
Talbot County37.2%±6.6 pp30.643.8%3,406
Dorchester County50.3%±7.5 pp42.857.8%2,731
Somerset County32.8%±10.6 pp22.243.4%1,880
Kent County41.9%±12.6 pp29.354.5%1,478
Caveats & limitations
  • Cohabiting unmarried partners are excluded from the numerator. S1101 separates them as their own household type. A child living with two unmarried cohabiting biological parents is not counted as “single-parent” here, even though they are not in a married household.
  • Multi-generational homes count by householder. A child living with a single mother and her own parents contributes to the numerator if the mother is the householder; if the grandparent is the householder, the family-type classification can differ.
  • Small-county estimates are noisy. Kent (±12.6 pp) and Somerset (±10.6 pp) have 90% confidence intervals roughly 25 points wide. Treat their exact rankings with caution; their order can shift on the next ACS release.
  • The Census definition of “household” is residence-based. Children who split time between two parents’ homes are counted at the residence the Census respondent listed as primary. Two homes, one child — one row in the count.
Files for this metric

The derived dataset behind this page. The underlying Census counts and 90% margin of error are in every row so anyone can verify the calculation without leaving the spreadsheet.

CSV
Single-parent household share (derived, with MoE)
24 counties × {single-parent %, 90% MoE, single-dad count, single-mom count, total HH with kids, dataset, vintage, table, formula}.
4 KB
Reproduce it yourself

The numbers on this site are not an opinion. They are arithmetic applied to public data. To verify them:

  1. Open S1101 on data.census.gov filtered to Maryland counties. Click Download to grab the raw CSV.
  2. For each Maryland county row, pull the values for C01_005E, C03_005E, and C04_005E.
  3. Compute (C03_005E + C04_005E) / C01_005E × 100. That’s the figure.
  4. Cross-check against the derived CSV above — it carries the underlying counts and the margin of error for every county.