Families of places

For deaths in 2010 to 2015, life expectancy at birthLife expectancy at birthThe average number of years a newborn would live if death rates at every age stayed as they were in the period measured. It summarizes the death rates of a period and is not a forecast for any child. was 81.781.7 years of life expectancy at birthEstimandpopulation-weighted mean of the tract life expectancy over the tracts of the family, bootstrap interval over countiesInterval95% CI 81.5 to 81.8n16,221 tractsWeightsACS 2019-2023 tract populationPopulationresidents of tracts in the high income familyPerioddeaths in 2010 to 2015SourceUSALEEP life expectancy years in the census tracts of this chapter's high-income family and 74.774.7 years of life expectancy at birthEstimandpopulation-weighted mean of the tract life expectancy over the tracts of the family, bootstrap interval over countiesInterval95% CI 74.5 to 74.9n12,903 tractsWeightsACS 2019-2023 tract populationPopulationresidents of tracts in the high poverty familyPerioddeaths in 2010 to 2015SourceUSALEEP life expectancy years in the tracts of its high-poverty family. The sort that drew those families never used life expectancy. K-meansK-meansA method that sorts places into a chosen number of groups so that places in a group resemble one another on the measures used. The analyst sets the number of groups. placed 75,71775,717 tractsEstimandcount of tracts clusteredWeightsnone (full count)Populationcensus tracts in the frame (2020 tracts with a published ADI rank, residents and all 15 measures)Periodinputs from ACS 2019-2023SourceAmerican Community Survey tracts into 44 familiesEstimanddefinition: the number of families the chapter usesWeightsnone (definition)Populationk-means fits comparedPeriodACS 2019-2023SourceAmerican Community Survey families on 1515 measuresEstimandcount of measures (definition)Weightsnone (definition)Populationdeprivation measures given to k-meansPeriodACS 2019-2023SourceAmerican Community Survey of the census measures of advantage and hardship that the Area Deprivation IndexArea Deprivation Index (ADI)A score that ranks neighborhoods by income, education, employment and housing conditions drawn from the American Community Survey. In this book a higher national percentile means a more deprived neighborhood. (ADI) combines. Life expectancy, the adult income rank of children raised in low-income families and several health measures were kept out of the sort and compared with the families afterward.

Families sorted without outcomes differ in life expectancy

The chapter names the four families for the measures that set each apart. They are the high-income family, the crowded renters, the mid-range family and the high-poverty family, numbered by the mean ADI rank of their tracts from the least to the most deprived. The names say nothing about who lives in the tracts, and no measure in the sort is a race, an ethnicity or a place of birth. Life expectancy and adult income rank, both held out of the sort, are highest in the high-income family and lowest in the high-poverty family. Life expectancy and adult income rank are older than the measures, since the families are drawn on conditions in 2019 to 2023 and adult income rank describes children born in 1978 to 1983. The comparison shows present conditions lining up with past outcomes, and it does not show that the families produce them.

Figure 1

Life expectancy runs from 81.7 years in the high-income family to 74.7 years in the high-poverty family, and the crowded-renter and mid-range families do not differ detectably

Population-weighted mean of tract values by family with 95% intervals from a bootstrap over counties; none of these measures was used to form the families

Measured outcomesModeled health measures

Life expectancy and adult income rank are highest in the high-income family and lowest in the high-poverty family

Life expectancy at birth for deaths in 2010 to 2015 and adult income rank of children born 1978 to 1983, population-weighted family means with 95% intervals

Life expectancy at birth (years)

Adult income rank, children from low-income families (percentile)

Population-weighted mean of tract values (units in each panel title)

The order of the families is the same on both measures except that the crowded renters are below the mid-range family on income rank and not detectably different on life expectancy.

Show data

Data for Life expectancy runs from 81.7 years in the high-income family to 74.7 years in the high-poverty family, and the crowded-renter and mid-range families do not differ detectably, Measured outcomes

conf_high	conf_low	estimate	family	label	n	panel
81.81	81.50	81.66	High income	High income	16,221	Life expectancy at birth (years)
78.99	78.09	78.58	Crowded renters	Crowded renters	6,587	Life expectancy at birth (years)
78.67	78.50	78.59	Mid-range	Mid-range	36,163	Life expectancy at birth (years)
74.94	74.55	74.74	High poverty	High poverty	12,903	Life expectancy at birth (years)
48.57	47.71	48.13	High income	High income	17,256	Adult income rank, children from low-income families (percentile)
42.07	40.51	41.30	Crowded renters	Crowded renters	6,869	Adult income rank, children from low-income families (percentile)
43.41	42.98	43.20	Mid-range	Mid-range	37,728	Adult income rank, children from low-income families (percentile)
37.58	36.90	37.23	High poverty	High poverty	13,814	Adult income rank, children from low-income families (percentile)

The two middle families do not follow the ADI. The index scores the crowded renters as less deprived than the mid-range family by 16.3 (95% CI 8.4 to 22.6)16.3 points of the ADI national rankEstimanddifference between the population-weighted mean ADI national rank of the mid-range and the crowded-renter family, bootstrap interval over counties with shared drawsInterval95% CI 8.4 to 22.6n75,717 tractsWeightsACS 2019-2023 tract populationPopulationresidents of tracts in the mid-range and crowded-renter familiesPeriod2023 release on ACS 2019-2023SourceArea Deprivation Index points on its national rank. Life expectancy is 78.678.6 years of life expectancy at birthEstimandpopulation-weighted mean of the tract life expectancy over the tracts of the family, bootstrap interval over countiesInterval95% CI 78.1 to 79.0n6,587 tractsWeightsACS 2019-2023 tract populationPopulationresidents of tracts in the crowded renters familyPerioddeaths in 2010 to 2015SourceUSALEEP life expectancy years in the crowded-renter family and 78.678.6 years of life expectancy at birthEstimandpopulation-weighted mean of the tract life expectancy over the tracts of the family, bootstrap interval over countiesInterval95% CI 78.5 to 78.7n36,163 tractsWeightsACS 2019-2023 tract populationPopulationresidents of tracts in the mid-range familyPerioddeaths in 2010 to 2015SourceUSALEEP life expectancy years in the mid-range family, with no detectable difference. The crowded renters also do no better on adult income rankIncome rank (kfr)The position that adult children reach in the national distribution of household income, from the lowest rank to the highest, averaged over children who grew up in a place with parents at a stated income level. The Opportunity Atlas labels this measure kfr. or on modeled estimates of diabetes, smoking and self-rated health. The ADI follows home value closely, as Chapter 1.1, What the deprivation index measures shows, and the crowded renters live in tracts with higher home values than the mid-range family does.

How the families were drawn

The frame is the 2020 census tracts that have a published ADI rank, residents and every measure, taken from the American Community Survey (ACS) for 2019 to 2023. Puerto Rico's tracts are outside it. Each measure is standardized so that a dollar amount and a proportion count equally, and k-means places the family centers so that tracts lie as close as they can to their nearest center.[1] Each tract counts once, so a family is a family of places and not of people, and residents return as weights in every family mean. The ADI is not an input, but it is built from the same ACS measures. Chapter 9.2, Eight types of neighborhood sorts tracts on income, poverty, education, family structure, tenure, age and density, with no outcome among its inputs, so its types separate places by how they are settled as well as by their resources, while the families here use deprivation measures alone. Life expectancy comes from death records and adult income rank from tax records, both interpolated from 2010 tracts to 2020 tracts by area. The health measures are modeledModeled estimateA figure produced by a statistical model instead of a direct count or measurement, usually for small areas where direct data are thin, borrowing strength from surveys, neighboring areas or other data. Neighboring areas can share errors, so single extremes call for caution. by PLACES from each survey respondent's age, sex, race and education and the poverty rate of the respondent's county. The model applies the result to the age, race and sex counts of each tract, and no tract-level input of the families enters it. The intervals resample whole counties given the families, and they do not carry the uncertainty about the families themselves. The report gives every estimate with its interval and sample size.

Four families are less stable than three or five

K-means returns a partition for any number of groups, and the data do not say how many there are. The chapter therefore fits the same k-means to simulated clouds that keep each measure's distribution and the correlations among measures but have no clusters. On the silhouetteSilhouetteA score for how well a place fits the group it was assigned in a clustering. Higher values mean the place sits close to its own group and far from the others., a score of how well tracts fit their own group against the nearest other, the tract data depart most from such a cloud at four groups.[2] That is the rule by which the chapter chooses four families. There the silhouette is 0.210.21 silhouette (-1 to 1)Estimandmean silhouette of the four families, interval from the sampling error of the sampleInterval95% CI 0.21 to 0.21n10,000 tractsWeightsnone (each tract counts once)Populationa seeded random sample of tracts in the framePeriodACS 2019-2023SourceAmerican Community Survey for the tract data against 0.120.12 silhouette (-1 to 1)Estimandmean silhouette of the 4-group k-means fit to simulated data with no clusters (mean of 3 draws)n10,000 tractsWeightsnone (each tract counts once)Populationa seeded random sample of tracts in the framePeriodACS 2019-2023SourceAmerican Community Survey for the clouds. The departure is modest, and the families are regions of a continuum and not separate clusters. The choice draws lines on that continuum, and it does not count the kinds of neighborhood that exist.

Refits on resamples that draw whole counties test whether the families come back.[3] At four groups, 17.5% (95% CI 12.9% to 23.4%)17.5% percent of refitsEstimandshare of refits whose agreement with the families is below the cutoff, Wilson interval that treats the refits as independent drawsInterval95% CI 12.9% to 23.4%n200 refitsWeightsnone (each refit counts once)Populationrefits on resamples of counties at four groupsPeriodACS 2019-2023SourceAmerican Community Survey of the refits find a different arrangement, against 0.0% (95% CI 0.0% to 1.9%)0.0% percent of refitsEstimandshare of refits whose agreement with the families is below the cutoff, Wilson interval that treats the refits as independentInterval95% CI 0.0% to 1.9%n200 refitsWeightsnone (each refit counts once)Populationrefits on resamples of counties at 3 groupsPeriodACS 2019-2023SourceAmerican Community Survey at three groups and 2.0% (95% CI 0.8% to 5.0%)2.0% percent of refitsEstimandshare of refits whose agreement with the families is below the cutoff, Wilson interval that treats the refits as independentInterval95% CI 0.8% to 5.0%n200 refitsWeightsnone (each refit counts once)Populationrefits on resamples of counties at 5 groupsPeriodACS 2019-2023SourceAmerican Community Survey at five. The other arrangement also turns up in single-start fits to all tracts, with only 1.1%1.1% percent more within-group varianceEstimandwithin-group sum of squares of the best different arrangement relative to the fit that defines the familiesn75,717 tractsWeightsnone (each tract counts once)Populationcensus tracts in the framePeriodACS 2019-2023SourceAmerican Community Survey more within-group variance than the families. It puts 76.9%76.9% percent of the family's tractsEstimandshare of the family's tracts that fall in one group of the best different arrangementn6,870 tractsWeightsnone (each tract counts once)Populationtracts in the crowded-renter familyPeriodACS 2019-2023SourceAmerican Community Survey of the crowded-renter tracts in one group with 47.6%47.6% percent of the family's tractsEstimandshare of the family's tracts that fall in the same group of the best different arrangement as most crowded-renter tractsn13,817 tractsWeightsnone (each tract counts once)Populationtracts in the high-poverty familyPeriodACS 2019-2023SourceAmerican Community Survey of the high-poverty tracts. Weighting tracts by population, or taking the logs of the dollar amounts, also changes the partition, and the crowded-renter family depends most on those choices.

Figure 2

The tract data beat a cloud with no clusters by 0.09 in silhouette at four families, more than at any other number of families

Mean silhouette of k-means fits to the 15 standardized deprivation measures, tract data and simulated data with no clusters, and agreement of refits on resamples of counties

SilhouetteAgreement of refits

At four groups 17.5% of refits on resampled counties find a different arrangement, more than at three or five groups

Adjusted Rand index between the fit on all tracts and refits on resamples of counties, by number of families, median and tenth to ninetieth percentile

Agreement with the fit on all tracts (adjusted Rand index)

Median agreement is high up to five families, lower at six and lowest at seven. The band reaches low values at four families because some refits find a different arrangement of the four groups, which is rarer at three and five.

Show data

Data for The tract data beat a cloud with no clusters by 0.09 in silhouette at four families, more than at any other number of families, Agreement of refits

k	median	p10	p90	refits
2	0.884	0.738	0.978	200
3	0.919	0.855	0.961	200
4	0.882	0.352	0.953	200
5	0.881	0.753	0.947	200
6	0.798	0.645	0.922	200
7	0.543	0.467	0.777	200
8	0.780	0.677	0.917	200

Deciles of the ADI mix the families

A ranking puts tracts in order, and a decile of the ranking holds whatever kinds of tract land there. No decile of the ADI national rank is a single family, and the crowded renters appear in every decile. In the most deprived decile, residents of mid-range tracts live 2.4 (95% CI 2.1 to 2.7)2.4 years of life expectancy at birthEstimanddifference between the population-weighted mean life expectancy of the mid-range and the high-poverty family within the most deprived decile of ADI ranks, bootstrap interval over counties with shared drawsInterval95% CI 2.1 to 2.7n75,717 tractsWeightsACS 2019-2023 tract populationPopulationresidents of tracts in the most deprived decile of tracts that belong to the mid-range and high-poverty familiesPerioddeaths in 2010 to 2015SourceUSALEEP life expectancy years longer than residents of high-poverty tracts, for deaths in 2010 to 2015. Across the families, present conditions line up with past life expectancy, and for the two middle families the outcomes do not follow the ADI. At county scale the families blur. Each county can be labeled with the family whose center lies nearest the county's own measures, leaving out the single-parent share, which the county tables define differently. Those labels agree with the tract families at 0.220.22 adjusted Rand index (1 = identical partitions, 0 = chance agreement)Estimandagreement between the tract families and the county labels, each tract taking its county's labeln75,717 tractsWeightsnone (each tract counts once)Populationcensus tracts in the framePeriodACS 2019-2023SourceAmerican Community Survey, on a scale where one means identical partitions and zero means chance. A county label does not recover the crowded-renter family. Chapter 1.2, What a county average leaves out makes the same comparison for the index itself.

Figure 3

In the most deprived decile of tracts, 79.2% of residents live in high-poverty tracts and 16.1% in mid-range tracts

Share of residents in each decile of the ADI national rank by family of their tract, percent of residents, tracts of the frame, ACS 2019 to 2023

• High income
• Crowded renters
• Mid-range
• High poverty

Decile of the ADI national rank (1 = least deprived, 10 = most deprived)

Show data

Data for In the most deprived decile of tracts, 79.2% of residents live in high-poverty tracts and 16.1% in mid-range tracts

decile	family	share	tracts
1	High income	88.30	6606
1	Crowded renters	11.09	901
1	Mid-range	0.378	39
1	High poverty	0.234	15
2	High income	70.94	5256
2	Crowded renters	19.28	1508
2	Mid-range	9.17	770
2	High poverty	0.617	54
3	High income	50.70	3698
3	Crowded renters	12.78	960
3	Mid-range	35.16	2708
3	High poverty	1.36	123
4	High income	18.56	1418
4	Crowded renters	10.33	795
4	Mid-range	68.03	5269
4	High poverty	3.07	282
5	High income	3.32	256
5	Crowded renters	8.23	589
5	Mid-range	82.74	6171
5	High poverty	5.72	477
6	High income	0.440	42
6	Crowded renters	8.14	559
6	Mid-range	82.69	6223
6	High poverty	8.73	699
7	High income	0.048	6
7	Crowded renters	7.61	516
7	Mid-range	77.17	5966
7	High poverty	15.18	1177
8	High income	0.015	1
8	Crowded renters	6.78	430
8	Mid-range	68.51	5193
8	High poverty	24.70	1806
9	High income	0.006	1
9	Crowded renters	5.13	331
9	Mid-range	52.11	4127
9	High poverty	42.76	3176
10	High income	0	0
10	Crowded renters	4.71	281
10	Mid-range	16.14	1280
10	High poverty	79.16	6008

What these families cannot tell you

The families describe tracts, and a family's life expectancy is an average over tracts that says little about any one resident (ecological inferenceEcological inferenceDrawing conclusions about individuals from data about places or groups. A pattern across counties need not hold for the people inside them.). Nothing here shows that living in a family's tracts changes a person's outcome, and the outcomes that line up with the families reflect selection, history, prices and policy that these data do not separate. The frame leaves out tracts for which the ACS withholds a median rent, home value or mortgage cost, and it leaves out Puerto Rico. The measures are tract estimates without margins of error, and the USALEEPUSALEEPThe U.S. Small-area Life Expectancy Estimates Project, a National Center for Health Statistics release of life expectancy at birth for most census tracts. It covers deaths from 2010 to 2015 only and has not been updated. life expectancy estimates suit maps and gradients and not single tracts.

---

Go deeper

• The report: every estimate with its interval and sample size.
• Tables: every figure of the report as a table.

Reading paths that include this chapter:

• Methods first

Data sources

• American Community SurveyACS 5-year estimates 2019-2023 for 2020 tracts and counties: the 15 deprivation measures, from the staged tract file and the Census API pull of housing, occupation, income-distribution and schooling tablesYears: 2019-2023 (5-year) · Weights: ACS 2019-2023 tract population · License: Public domain (U.S. Census Bureau)
• Area Deprivation IndexArea Deprivation Index 2023 (v4.0.1) national rank, tract values averaged from block groups, used to number the families and to form decilesYears: 2023 release (version 4.0.1; built on ACS 2019-2023) · Weights: ACS 2019-2023 tract population · License: Free for non-commercial use; cite the Neighborhood Atlas and Kind and Buckingham
• USALEEP life expectancyNCHS small-area life expectancy at birth, deaths in 2010 to 2015, 2020 tracts (area-weighted), institutional tracts left outYears: 2010-2015 · Weights: ACS 2019-2023 tract population · License: Public domain (NCHS, CDC)
• Opportunity AtlasAdult household income rank of children born 1978 to 1983 raised in families at the 25th income percentile, 2020 tractsYears: 1978-1983 birth cohorts (2018 release) · Weights: ACS 2019-2023 tract population · License: Public use (Opportunity Insights and the U.S. Census Bureau)
• CDC PLACESCDC PLACES 2024 release (BRFSS 2022): diabetes, current smoking and fair or poor health among adults, modeled prevalenceYears: 2024 release (tracts); 2025 release (ZCTAs, 2023 data year) · Weights: ACS 2019-2023 tract population · License: Public domain (CDC, U.S. Department of Health and Human Services)

All data sources

References

1. MacQueen, J. (1967). Some methods for classification and analysis of multivariate observations. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Statistics, 281-297.
2. Rousseeuw, Peter J. (1987). Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics 20: 53-65. doi:10.1016/0377-0427(87)90125-7
3. Hennig, Christian (2007). Cluster-wise assessment of cluster stability. Computational Statistics & Data Analysis 52(1): 258-271. doi:10.1016/j.csda.2006.11.025

How to cite

@misc{singh2026familiesofplaces,
  author = {Singh, Vishal},
  title = {Families of places},
  year = {2026},
  url = {https://vishalsingh.org/zipcode/families-of-places},
  note = {In ZIP Code Destiny. Published 9 October 2026.}
}