Gallup US Daily × county mortality · 2008–2012
Deaths of despair gave America a mortality map. Buried in 1,769,079 Gallup interviews is the question that map raises: which feeling lives in the places where people die? Not sadness. Not worry. Pain — and a future that looks shorter than the present.
When Anne Case and Angus Deaton named the rise in midlife deaths from suicide, alcohol and drugs “deaths of despair,” they handed American geography a grim new layer: a county map of self-destruction. But mortality is the end state. The despair itself — whatever it is — had to be felt first, by the living, in those same places. This file holds the feelings. Between 2008 and 2012, Gallup’s daily poll asked 1,769,079 Americans whether yesterday brought physical pain, sadness, or worry; where their life stands on a zero-to-ten ladder; and where they expect it to stand in five years. Each interview carries a county code, and 89 percent of them land in one of the 1,341 larger counties with a published deaths-of-despair mortality rate.
So we can finally ask the question the mortality map begs: which self-report actually tracks the geography of despair death? If despair were ordinary unhappiness, the sadness and worry maps should match the death map best. They don’t — not even close. Across 1,027 counties with enough interviews to measure, the share of residents reporting physical pain a lot of the day yesterday is the strongest correlate of despair mortality (Spearman ρ = 0.55). Worry, the most common negative feeling in the country, is the weakest (ρ = 0.21). And among the ladder questions, where people think their life is going (ρ = −0.43) tracks the death map better than where they say it stands today (ρ = −0.36).
That ranking is Case and Deaton’s account, written in survey data by the living: despair is not sadness. It is pain plus the death of expectation.
One map, two despairs
Every county is colored on two axes at once: deaths-of-despair mortality (purple, left to right) and a reported feeling (teal, bottom to top of the legend). The dark diagonal is concordance — places where despair is both felt and fatal. The saturated corners are the mismatches. Swap the felt measure and watch the map recompose.
Most of the map sits near the diagonal: where despair kills, people report more pain, lower ladders, bleaker futures. The interest is in the 53 counties where mortality runs in the top quintile of mapped counties (above 65.7 deaths per 100k) while reported feeling sits below the median — call them fatal, not felt — and the 49 counties of the reverse, felt, not fatal.
The quiet-fatal list is strikingly Western. Fremont County, Wyoming — home to the Wind River Reservation — has the second-highest despair mortality on the map, 157.7 per 100k, yet its residents’ reported feelings land below the national middle. Carson City, Nevada (121 per 100k), Eddy County, New Mexico (117), San Juan County, New Mexico (104), Coconino County, Arizona, and Natrona County, Wyoming follow the same pattern: extraordinary death rates, unremarkable self-reports. These are sparse, heavily rural or reservation-adjacent places where the survey may simply be reaching different people than the ones dying — a mortality rate describes a county’s dead over years; a phone poll samples its living adults.
The loud-but-living list is Appalachian and Southern: Greene County, Tennessee tops it, with despair mortality of just 39.4 per 100k but the highest felt-despair composite on the map — one in three adults there reported physical pain yesterday. Henry County, Virginia (36.5 per 100k) is next. Places like these report the feelings of despair at full volume without (yet, or ever) the matching death rate — a reminder that the two measures capture genuinely different things, and that misery is not a death sentence.
Which feeling tracks death?
Spearman rank correlation between each county-level measure and deaths-of-despair mortality, across counties with at least 300 interviews for that measure. Whiskers: 95% interval from a thousand bootstrap resamples of counties.
The ordering is not subtle. Pain leads at 0.55, with the composite (0.48) and diagnosed depression (0.46) behind it, then the future ladder (−0.43), sadness (0.36), the current ladder (−0.36), and finally worry (0.21). The bootstrap whiskers barely graze each other at the top and bottom: pain’s interval (0.50 to 0.59) sits entirely above every other measure’s.
Note what kind of measures these are. Pain, sadness and worry are experiential — did you feel this a lot of the day yesterday, yes or no. The ladders are evaluative — a considered judgment of your life now, and in five years. Despair mortality recruits from both columns, but selectively: the body (pain) and the horizon (the future ladder), more than the mood (sadness, worry). Worry is nearly useless as a marker of fatal geography — 29 percent of Americans reported it in any given year of this window, and it concentrates in striving places as much as suffering ones.
The pattern holds, and sharpens, within kinds of places. Among metro counties the pain correlation is 0.58; among nonmetro counties it is 0.36 — thinner samples and a compressed range, but the ranking survives. Split counties by median household income and pain leads in every third: 0.43 in the poorest, 0.28 in the middle, 0.61 in the richest — where the future ladder also bites hardest (−0.45 in metro counties, −0.12 outside them). Geography of despair is not just poverty in disguise; within each income band, the places that hurt are the places that die.
The ranking survives inside bands
Spearman correlation with despair mortality, computed separately within urbanicity and county-income bands. Larger teal dots: physical pain.
County correlations are about places. A harder question: does living in a high-despair county go with feeling worse even for observably similar people? Sort the 1,341 mortality-covered counties into quintiles and compare individuals — not county averages — after standardizing every quintile to the same mix of age and education.
It does. In the bottom mortality quintile, 21.4 percent of age-and-education-standardized respondents reported pain yesterday; in the top quintile, 26.4 percent — a gap of 4.9 points that composition cannot explain (the raw gap is 7.5). The share placing their life in five years at 4 or below on the ladder — a foreclosed future — rises from 9.5 to 11.9 percent, standardized, top against bottom. Composition matters: adjustment takes roughly a third off the raw pain gap and nearly half off the future-ladder gap. High-despair counties are older and less schooled, and that explains some of the feeling. But not most of it. Something about the places, or about who stays in them, remains.
Living in a dying county feels different
Individual-level shares by county despair-mortality quintile, 2008–2012, unweighted. Solid: standardized to a fixed age × education mix. Dashed: raw. Vertical axes magnified to show the gradient; they start above zero.
Everything here is ecological in at least one direction. These are places where people report more pain and places where despair kills more — never the same people, and never the people who die. Gallup interviews living adults with phones; mortality counts the dead. A county can be high on both without any overlap between its sufferers and its casualties, and the discordant-county lists above show how far apart the two can drift.
The mortality figure is a single ~2018-vintage rate joined to all five survey years — context, not a time series; nothing here speaks to trends. The Gallup extract carries no survey weight, so every number is an unweighted sample estimate: good for contrasts and rankings, less reliable for levels. County codes come from ZIP codes and are approximate where ZIPs cross county lines. The mortality file covers only 1,341 larger counties — most of the population, a minority of the land — and thin counties are suppressed, so rural America is underrepresented on the map and the quiet-fatal pattern out West may be even stronger than it looks. And a correlation of 0.55 is a strong claim about geography, not destiny: it leaves most county-to-county variation unexplained.
What survives all of that is the ranking. Ask the living of a dying place how they feel, and the answer that maps onto the dying is not that they are sad, and not that they worry. It is that they hurt, and that they expect less. Case and Deaton inferred it from death certificates; almost two million interviews say the same thing from the other side.