Connecticut by planning region: Census 2020–2024
Every adult in Connecticut, as the Census counts them, by planning region (the state's county equivalents since 2022) and grouped by metro area. Each person is a real combination of traits from the Census's anonymised person records, weighted so every region matches its published tables, with party lean and 2024 turnout modelled and labelled so.
9 places, by metro area / county · Team plan
Every plan can look; creating a population from a state by county or by município needs Team.
What each person has
- Age
- 18-2412%
- 25-3416%
- 35-4416%
- 45-5416%
- 55-6418%
- 65-7414%
- and 1 more
- Sex
- Man48%
- Woman52%
- Race and ethnicity
- White65%
- Black10%
- Hispanic16%
- Asian5%
- Other or more than one race4%
- Education
- No high school diploma9%
- High school diploma or GED26%
- Some college or associate's degree26%
- Bachelor's degree23%
- Graduate degree17%
- Household income
- Under $25,0008%
- $25,000–49,99910%
- $50,000–74,99911%
- $75,000–99,99911%
- $100,000–149,99918%
- $150,000 or more36%
- and 1 more
- Work
- Working62%
- Unemployed, looking for work4%
- Not in the labor force34%
- Home
- Lives in an owned home68%
- Lives in a rented home28%
- Lives in group quarters4%
- Marital status
- Married49%
- Widowed6%
- Divorced11%
- Separated1%
- Never married33%
- Children at home
- No children at home70%
- Children at home30%
- Veteran
- Not a veteran95%
- Veteran5%
- Born
- Born in the US81%
- Born abroad19%
- Health coverage
- Has health coverage94%
- No health coverage6%
- Disability
- No disability85%
- Has a disability15%
- Citizenship
- US citizen92%
- Not a US citizen8%
- Party lean (modelled)
- Democratic-leaning49%
- Republican-leaning32%
- Independent or no lean18%
- Voted in 2024 (modelled)
- Voted61%
- Did not vote31%
- Not eligible (not a citizen)8%
County by county
Adults in each county: darker has more. Point at a county for its numbers.
Error report
How close these people come to the published figures, and where they fall short.
2,892,922 adults · 146,626 Census records · 9 counties in 6 areas (metro areas and the rest of the state)
Built 2026-10-10 · records connecticut-2024-v1
How close the people land on the Census
Each county's records are weighted until they match its American Community Survey tables. The figure is the worst county's largest gap on each table, in percentage points.
| Table | Worst county |
|---|---|
| Sex by age | 0.2 |
| Race and ethnicity | 0.3 |
| Education | 0.3 |
| Household income | 0.4 |
| Work | 0.1 |
| Owned or rented home | 0.1 |
| Citizenship | 0.0 |
A check left out of the fit: veterans
Veteran status was not used in weighting, so it shows how well the weighted people carry traits nobody fitted. Per county, the share of veterans among civilian adults lands a median 0.1 points from the Census's own figure; 95% of counties within 0.2, the worst 0.2.
What drawing a population adds
A population is a sample of these weighted people, so small counties vary from draw to draw. Largest gap per county between the people drawn and the county's records, counties with at least 30 people.
| Trait | 10,000 people | 100,000 people | ||
|---|---|---|---|---|
| Median | 95% of counties within | Median | 95% of counties within | |
| Race and ethnicity | 0.1 | 0.2 | 0.0 | 0.0 |
| Education | 0.3 | 1.3 | 0.0 | 0.1 |
| Household income | 0.2 | 0.8 | 0.0 | 0.1 |
Party lean and 2024 turnout (modelled)
No Census record says how anyone votes, so these two traits are modelled. Each person starts from what Cooperative Election Study 2024 respondents like them (race, age, education and sex, in the Census Northeast) said and did, with turnout checked against the voter file. Each county is then calibrated so its people's turnout and Democratic-minus-Republican margin match its 2024 presidential result (MIT Election Lab). Those results count votes for president, so turnout runs a point or so under ballots cast.
Connecticut's 2024 returns are reported by its eight old counties, which the Census replaced with planning regions in 2022, and the two do not line up. Each old county's votes are shared among the regions by the population they have in common, as if each old county voted the same throughout, so each region's result is itself an estimate.
| Check | Margin | Turnout |
|---|---|---|
| Each county against its 2024 result (calibrated) | median 0.0, worst 0.0 | median 0.0, worst 0.0 |
| Demographics alone, given only the statewide result | median 3.3, worst 13.2 | median 1.0, worst 5.6 |
| The same, with priors that never saw anyone from Connecticut | median 3.2, worst 13.1 | median 1.0, worst 5.6 |
Leans against Connecticut's own survey respondents
Nothing was fitted to these: the share of each lean among Connecticut's adults, as modelled and as the CES's Connecticut respondents gave it.
| Democratic | Republican | Neither | Turnout | |
|---|---|---|---|---|
| Model | 49% | 32% | 18% | |
| CES Connecticut (n=589) | 46% | 28% | 25% | |
| Priors that never saw anyone from Connecticut, applied to its 574 citizen respondents | 49% | 33% | 17% | 61% |
| CES Connecticut, citizens (n=574) | 47% | 29% | 25% | 56% |
The county totals are right by construction: that is what calibration does. The honest test is the people alone, in the table above: given only the statewide result, how far each county's people land from its margin. Counties that vote very differently from people like them elsewhere in the region land furthest; the map shows where.
Within a county, calibration moves everyone's lean by the same factor, so lean by race inside one county is the weakest part. In the smallest counties the Census's five-year estimates carry wide margins, and where more presidential votes were cast than the estimate of citizens, turnout is held at 98%.
Use these traits to include or compare groups. They are not anyone's stated party, and results are never polls of real people.
Every county (9)
| County | Area | Adults | Census gap | Veterans (Census / people) | Turnout 2024 | Margin: result / demographics alone |
|---|---|---|---|---|---|---|
| Capitol Planning Region | Hartford-West Hartford-East Hartford | 780,266 | 0.2 | 5% / 5% | 65% | +20 / +15 |
| Greater Bridgeport Planning Region | Bridgeport-Stamford-Danbury | 256,480 | 0.3 | 3% / 3% | 69% | +20 / +16 |
| Lower Connecticut River Valley Planning Region | Hartford-West Hartford-East Hartford | 146,455 | 0.2 | 6% / 6% | 73% | +13 / +12 |
| Naugatuck Valley Planning Region | Waterbury-Shelton | 362,719 | 0.3 | 5% / 5% | 64% | +11 / +12 |
| Northeastern Connecticut Planning Region | Small towns and rural Connecticut | 78,558 | 0.4 | 8% / 8% | 61% | −9 / +4 |
| Northwest Hills Planning Region | Small towns and rural Connecticut | 93,041 | 0.2 | 6% / 6% | 71% | −5 / +8 |
| South Central Connecticut Planning Region | New Haven | 458,975 | 0.2 | 4% / 4% | 61% | +12 / +16 |
| Southeastern Connecticut Planning Region | Norwich-New London-Willimantic | 225,847 | 0.3 | 8% / 8% | 66% | +11 / +10 |
| Western Connecticut Planning Region | Bridgeport-Stamford-Danbury | 490,581 | 0.4 | 3% / 3% | 70% | +18 / +19 |
Sources
- Records
- ACS PUMS 2020–2024 5-year, Connecticut, adults 18+
- Tables
- ACS 2020–2024 5-year tables B01001, B01001B/D/H/I, B15001, B19001, B23025, B25008, B26001; out of sample: B21001
- Geography
- 2020 Census tract to 2020 PUMA relationship file; tract population B01003; metro areas, OMB delineation of July 2023
- Citizenship
- ACS 2020–2024 5-year table B05003 (raked)
- Party lean
- Modelled. Priors: Cooperative Election Study 2024 common content (Harvard Dataverse doi:10.7910/DVN/X11EP6, CC0), citizen respondents in the Census Northeast, validated 2024 vote and 7-point party ID with leaners as partisans. Calibration: 2024 county presidential returns, MIT Election Data and Science Lab (Harvard Dataverse doi:10.7910/DVN/VOQCHQ, CC0), presidential votes standing in for ballots cast.