The Long Retreat of Trust

Since 1972, the General Social Survey has put the same blunt question to Americans: generally speaking, would you say that most people can be trusted, or that you can't be too careful in dealing with people? In that first year, 46.6 percent — very nearly half — chose trust. In 2024, 24.8 percent did. Across fifty-two years and roughly 43,000 respondents, the trusting share of the country has been cut almost in half, and the line that traces the decline is remarkable mostly for how little drama it contains. There is no single cliff, no one scandal or crash you can pin it on. It is a long, patient retreat, about four percentage points a decade — roughly half a point a year — with brief rallies that never hold.

That slow erosion would be worth charting on its own. What gives it weight is a second pattern running quietly underneath it. On the GSS's three-step happiness scale — not too happy (1), pretty happy (2), very happy (3) — people who say most others can be trusted average 2.31. People who say you can't be too careful average 2.12. Put differently: 38 percent of trusters call themselves very happy, against 29 percent of the wary. And that gap has not narrowed as trust collapsed. If anything it has widened a touch — a 0.18-point premium in the 1972–1989 surveys, 0.23 points in 2015–2024. The trusting side of the line has stayed the happier side, even as fewer and fewer Americans live on it.

Fifty years, one direction

The chart’s honesty requires admitting its gaps. The GSS did not ask the trust question every year — there is no reading for 1974, 1982, or most strikingly 2020, when the pandemic scrapped in-person fieldwork entirely. The dashed segments on the chart span its two longest silences — 1980–83 and the pandemic gap of 2018–22 — connecting real observations while asserting nothing about what happened in between. And the two most recent points, 2022 and 2024, come after the survey’s shift toward web-based interviewing — a mode change that can move answers on its own, independent of any real change in the public mood. The 2022 reading of 26 percent and the 2024 reading of 25 percent are the lowest in the series, but some unknown slice of that final step down may belong to the questionnaire rather than the country.

Even with those caveats, the trajectory is unambiguous, because it does not depend on any single point. Trust ran around or above 40 percent through much of the 1970s and 1980s, touched 49 percent in 1984, slid through the 30s across the 1990s and 2000s, and has not touched 35 percent since 2000 — nor 34 since 2018. Pew Research, using its own surveys, reported the same downslope in 2025; Robert Putnam documented its first half in Bowling Alone a quarter century ago, tying it to the thinning of clubs, leagues, congregations, and unions where strangers once became familiar. The GSS simply lets us watch the whole arc in one consistent instrument.

Part of the mechanism is generational replacement. Over five decades, the high-trust cohorts born before mid-century have aged out of the sample, and the cohorts replacing them came of age warier and stayed that way. That matters for interpretation: a 50-year drift in a repeated cross-section is not 50 years of individuals changing their minds. It is, in large part, one America being slowly swapped for another.

The premium that wouldn’t die

Now the second construct, and it is worth being pedantic about the difference. Trust is a social measure: a belief about other people. Happiness, as the GSS asks it, is an evaluative one: a judgment about your own life. They are correlated, not interchangeable, and nothing in this essay treats a trusting person as a happy one by definition.

What the data show is an association of unusual stamina. In every era the GSS has covered, people who say most others can be trusted report higher happiness than people who say you can’t be too careful. The gap is not large in absolute terms — about a fifth of a point on a three-point scale — and an honest reading keeps that modesty in view. But small and stubborn is its own kind of finding. The premium was 0.18 points in the 1972–1989 surveys, 0.17 in 1990–2004, 0.20 in 2005–2014, and 0.23 in 2015–2024. Five decades, four eras, one war on terror, one financial crisis, one pandemic, and the gap never closed. It grew.

Three stories, one correlation

None of this establishes that trust causes happiness, and the arrow could plausibly run every direction at once. Story one: trust pays. Trusting people cooperate more easily, build wider networks, extract more warmth from daily life, and the happiness follows. Story two: happiness pays. People whose lives are going well — secure jobs, stable marriages, good health — find the world has given them little reason for suspicion, and generalize accordingly. Story three: something upstream depresses both. Economic insecurity, inequality, segregated and frayed neighborhoods, the migration of social life onto adversarial platforms — any of these could push trust and happiness down together without either causing the other.

The political scientist Eric Uslaner, who spent a career on this question, argued that generalized trust is less a running tally of recent experiences than a moral disposition laid down early — absorbed from parents and stable childhoods more than from yesterday’s interactions. If he is right, the durable premium looks less like a reward trust pays out and more like two readings from the same deep gauge: people equipped early with optimism about others tend to score higher on both dials for life. A cross-sectional survey cannot adjudicate between these stories. What it can do is weigh against one comforting possibility — that the trust decline is cosmetic, a change in how people talk rather than in how life sorts them. A gap this consistent, in a sample this large, over a span this long, is not noise.

Trust is graded like a diploma

If the trusting are the happier group, it matters enormously who gets to be in it — and here the GSS delivers its sharpest verdict. Pooling the full 1972–2024 file, 23 percent of Americans without a high-school diploma say most people can be trusted. Among high-school graduates it is 35 percent; among bachelor’s degree holders, 52 percent; among those with graduate degrees, 61 percent. From the bottom of the education ladder to the top, the trust share rises about 2.6-fold.

The gradient is best read sociologically, not psychometrically. A graduate degree is a bundle: higher income, more autonomy at work, safer neighborhoods, thicker professional networks, fewer encounters where misplaced trust is ruinous. Uslaner’s blunter formulation was that trust flourishes among the optimistic and the secure — and education, in America, is the master variable of security. The wary are not making an error; many are reporting, accurately, what their corner of the country has taught them. Distrust can be a rational tax levied by circumstance, paid in the currency of well-being.

Living on the wary side

Assemble the three panels and the shape of the story emerges. The trusting share of America has nearly halved. The happiness premium attached to trust has quietly persisted, and lately widened. And membership on the trusting side is increasingly rationed by education — which is to say, by class. A half-century ago, the happier, higher-trust side of the line held nearly half the country and drew from across it. Today it holds a quarter, disproportionately the degreed and the comfortable. Whatever the causal traffic between trust and happiness, the population sorted onto the favorable side of both has become smaller and more select.

That reframes what the decline of trust costs. The usual accounting is civic — weaker institutions, harder collective action, pricier contracts in a society where handshakes need lawyers. The GSS suggests a private ledger too, denominated in the survey’s simplest question: taken all together, how would you say things are these days? The fifth of a point separating trusters from the wary is small on any one life. Spread across the tens of millions of people who crossed from one column to the other over fifty years, it describes a quiet, regressive redistribution of well-being — away from the many, toward a shrinking and credentialed few. The retreat of trust was never just a story about strangers. It was always also a story about how it feels to be home.