Methodology

Data Sources

• U.S. Census ACS 5-Year Estimates (2022) — Population by sex, age, marital status, race/ethnicity, and employment at the ZIP Code Tabulation Area (ZCTA) level (~33,000 areas).
• Census ACS PUMS — Public Use Microdata Sample providing cross-tabulated employment and income rates by sex, age, and marital status at the PUMA level (~2,500 areas).
• CDC NHANES 2017-2020 — National height distributions by sex, race/ethnicity, and age group from the National Health and Nutrition Examination Survey.
• ARDA / RCMS 2020 — Religious Congregations and Membership Study providing county-level Christian adherent counts by denomination.
• Pew Religious Landscape Study (2014) — Weekly worship attendance rates by denomination, used to estimate "practicing Christian" populations.
• HUD Income Limits (FY2025) — Area median income thresholds for the "family-sustaining career" employment filter.

How the Estimate Works

We apply sequential filters to the population within your search radius. Each filter reduces the pool by a percentage based on real demographic data:

1. Sex — Direct count from Census ACS per ZCTA.
2. Age range — Census age buckets with fractional overlap for ranges that don't align with bucket boundaries.
3. Single status — Census marital status data (never married + separated + widowed + divorced). Note: this includes people in committed unmarried relationships, which Census does not distinguish.
4. Employment — PUMA-level cross-tabulated rates (% of singles who are employed or earning above area median income) applied to ZCTA populations.
5. Height — NHANES height distributions modeled as normal curves, weighted by local race/ethnicity composition from Census ACS.
6. Faith — County-level rates from RCMS, with "practicing Christian" estimated using denomination-specific weekly attendance rates from Pew Research Center (e.g., 55% for evangelical, 37% for Catholic, 34% for mainline Protestant).

Key Assumptions and Limitations

• Filters are applied independently. The funnel multiplies filter rates together, which assumes that characteristics like employment, height, and faith are independent of each other given sex and age. In reality, these are correlated — for example, taller people tend to earn more, and religious people tend to marry earlier. This means estimates with multiple active filters likely overcount the actual pool.
• "Single" is broadly defined. Census data counts all unmarried people, including those in committed relationships. Roughly 18% of unmarried adults are cohabiting (Pew Research).
• Faith data uses county-level averages. Religious adherence rates from RCMS are available only at the county level and are applied uniformly to all ZIP codes within a county. Actual rates may vary significantly within a county.
• Employment data is at the PUMA level. Employment rates from PUMS cross-tabs cover areas of ~100,000-200,000 people, which may mask variation between specific neighborhoods within that area.
• Height distributions are national. NHANES provides height data at the national level only, stratified by sex, race/ethnicity, and age. We adjust for local demographics, but regional variation in height is not captured.
• Data has a time lag. ACS 5-year estimates average 2018-2022. RCMS religion data is from 2020. NHANES height data covers 2017-2020.

Sources

• U.S. Census Bureau, American Community Survey 5-Year Estimates (2022)
• U.S. Census Bureau, American Community Survey Public Use Microdata Sample (2022)
• Centers for Disease Control and Prevention, NHANES 2017-March 2020 Pre-Pandemic
• Association of Religion Data Archives, Religious Congregations & Membership Study (2020)
• Pew Research Center, Religious Landscape Study (2014)
• U.S. Department of Housing and Urban Development, Income Limits (FY2025)