Ask what makes people happy and you get a familiar list: money, marriage, health, faith, friends, the politics of the moment. Every item on it is real. But the list quietly assumes a single answer-key for a whole life — as if the things that lift a twenty-two-year-old are the same ones that hold up a seventy-eight-year-old. They are not. The happiness an adult reports is one outcome with a changing cause. Split the American lifespan into six stages, run the very same model inside each, and the predictors don't merely grow or shrink. They take turns leading.
This is the book's engine pointed at the one axis everyone travels. For six measures of well-being — how people rate their lives Evaluative, and how their day actually felt Experiential, plus satisfaction with their standard of living Domain — we estimate, separately for each life stage, the mutually-adjusted effect of twenty predictors, all put on one comparable scale. The result is a cast of characters that rotates as you age, with exactly one constant.
The numbers below are standardized effects: how much a one-step change in a predictor moves an outcome, measured in standard deviations of that outcome, so a coefficient of 0.20 means roughly the same thing whether it sits under "income" or "health" or "married." They are honest about size — most are modest, and relationships and health routinely beat money. And because this is the Gallup US Daily poll, which carries no survey weight, they are unweighted contrasts: excellent for patterns, not for headline national levels.
One outcome, a changing cause
Read the grid down a single column and you see what governs a life at one age; read across a single row and you watch a predictor's power rise, crest, or collapse over a lifetime. Six rows tell the whole story. Money's grip swells in the middle of life and slackens at both ends. Faith's tie to a good day climbs steadily with age. The wound of a lost partner is devastating to the young and ordinary to the old. Race crosses over. Politics opens late. And health holds the line throughout. Pulled out of the grid and placed side by side, those are the six arcs of an adult life.
The grip of money is the clearest bulge. Household income's effect on standard-of-living satisfaction is its weakest at the start of adult life — 0.14 — climbs to 0.31 in the forties, then falls back to 0.12 among the over-75s, the steepest age-decline of any predictor here. The pattern on the life rating is the same hump: 0.096 young, 0.23 at midlife, 0.113 old. Money matters most precisely in the years of mortgages, dependents and peak earning, and buys conspicuously less once those years are behind you.
Faith runs the opposite way, and splits in two. Saying religion is important to your daily life does almost nothing for a young person's mood — its tie to felt happiness yesterday is just 0.04 at 18–25 — but rises at every stage to 0.19 by old age, a roughly 3.5-fold strengthening once other things are held equal. Yet for the young its payoff is the highest of any age on the life rating (0.14): in youth religion buys meaning and evaluation; in age it buys the day itself.
The mark of a lost partner heals. Being separated, divorced or widowed instead of never-married costs an emerging adult 0.19 of a standard deviation on the life rating — the heaviest relational penalty anywhere in the data — and that scar shrinks monotonically to essentially zero (+0.00, not significant) by 75+. The fade is mostly widowhood becoming statistically ordinary; divorce and separation keep a real sting later in life. Politics, by contrast, switches on late: identifying Republican rather than Democrat is, if anything, a small advantage in youth and only becomes an evaluative penalty after 40, deepening to its largest drag near retirement (−0.135 on the ladder at 65–74) — and, strikingly, it never touches day-to-day mood at all.
That last one is the constant. Self-rated health is the strongest single predictor in nearly every stage and the most stable, its pull on the life rating drifting gently up from 0.23 in youth to 0.28 in old age — never collapsing, never handing off the lead. (It is also the most proximal of these measures: how you rate your health and how you rate your life are both subjective verdicts, so read its size as an upper bound.) Everything else specializes. Health endures.
Six ages, six different engines
Set the leaderboard below to any stage and it ranks what most lifts (teal) and most lowers (rust) well-being there, for whichever outcome you choose. Walk it forward through life and watch the cast turn over. One thread runs through all six — a gender paradox in which women rate their lives higher than men at every age yet feel the day harder, and the size of that emotional gap is itself age-shaped: present in youth, smallest in midlife, widest near retirement.
The classic adult drivers of happiness are still asleep. Money is at its lifetime weakest — a one-step rise in household income lifts standard-of-living satisfaction just 0.14 here, less than half its midlife force, and the life rating only 0.10. The partisan gap hasn't opened: being Republican rather than Democrat is, if anything, a slight plus on the ladder (+0.07) and flat on Thriving. Marriage barely registers (+0.08 on the ladder, against +0.19 a decade later). What rules instead is the body and the breakup. A health problem that limits daily activity inflicts its lifetime-maximum emotional toll — worry +0.34 and stress +0.30, both higher than at any later age — and pulls standard-of-living satisfaction down 0.25. And a serious relationship that has ended is more scarring now than it will ever be again: separated, divorced or widowed costs −0.19 on the life rating, −0.20 on Thriving, and +0.18 on stress.
Two things set this stage apart from every later one. Religion buys its highest life-rating payoff of any age (+0.14) yet does almost nothing for the day itself (daily happiness +0.04, its lifetime low): in youth faith supplies meaning, not mood — the reverse of what it becomes in old age. And Black emerging adults sit just before the well-being crossover. Alone among the six stages, they rate their standard of living (−0.20) and daily happiness (−0.15) below otherwise-similar white peers — the exact opposite of the clear Black advantage that appears after 40. The one familiar anchor already in place is health, lifting the ladder 0.23.
Women & men. The happiness paradox is there from the start: young women rate their lives higher than young men (+0.13 on the ladder, +0.12 on Thriving) yet report a notably harder day — the largest youthful stress gap (+0.14) and more worry (+0.09).
This is the household-formation stage, and the relationship numbers peak with it. The marriage premium reaches its lifetime high — +0.19 on the life rating, +0.13 on Thriving, +0.11 on daily happiness — exactly when most people are forming unions; nowhere later does a partnership matter this much. The breakup penalty is still near its own maximum (−0.13 on the ladder, −0.20 on standard-of-living satisfaction). And income is climbing fast — 0.26 on standard-of-living satisfaction, nearly double its 0.14 in the youngest stage — but has not yet crested.
What is most distinctive here is that minority status looks its best in young adulthood, not later. The Hispanic edge peaks: the lowest stress of any stage (−0.24) and the highest life rating (+0.21). Young Asian adults post their largest calm too (stress −0.20). Yet the forces that will dominate after 40 are still only half-grown. Black Americans are essentially level with white peers on the life rating (+0.05) and Thriving (+0.04) — the midlife advantage hasn't formed — and the Republican penalty is mild (−0.085 on Thriving) where it will later reach −0.21. Health stays the broad backbone (+0.22 on the ladder; a health limit still drives worry +0.32).
Women & men. This is the most favourable balance women will see all life: the female evaluative edge is at its lifetime peak (+0.17 on the ladder, +0.15 on Thriving) while the emotional cost is muted (stress +0.07, worry +0.06).
Now money takes the wheel. Household income's grip is at its across-stage peak on both standard-of-living satisfaction (0.31) and the life rating (0.23), and it is the most protective against worry it will ever be (−0.12) — the squeeze years of earnings, mortgages and dependents, when an extra dollar moves a life the most.
Two forces switch on at exactly this point. The Black advantage in day-to-day feeling reaches its deepest protective trough — Black Americans report markedly less stress (−0.31) and worry (−0.28) than otherwise-similar white Americans, the widest such gap of the whole life course — and the life-rating gap that was zero in youth has opened into a clear Black advantage (+0.20 on the ladder, +0.18 on Thriving). The partisan evaluative penalty also crystallizes here for the first time: Republicans −0.17 on Thriving and −0.09 on the ladder, after being flat or positive before 40. Meanwhile a force that ruled youth has quietly retired — a health limit, emotionally devastating to the young, now barely dents the life rating (−0.09, against −0.18 at 18–25) even as it still drives worry (+0.28). Health remains the universal backbone (+0.25 on the ladder).
Women & men. Women still rate life clearly higher (+0.15 on the ladder), and the emotional tax is at its midlife lull — the smallest female stress (+0.07) and worry (+0.04) gaps of adulthood.
The conditional Black life-rating advantage reaches its widest point of all — +0.21 on the ladder, +0.23 on Thriving — a full reversal of the youth-era disadvantage, with the felt-stress gap still deep (−0.28). At the same time the partisan penalty climbs toward its own peak: Republicans −0.21 on Thriving, Independents −0.14 on the ladder. Income is still near its high (0.29 on standard-of-living satisfaction) and health near its strongest (+0.27 on the ladder).
The mark of this stage is what has gone quiet. Education — never a large independent force — is now effectively nil: +0.02 on the ladder, and statistically indistinguishable from zero on both daily happiness and standard-of-living satisfaction. The breakup scar has half-healed (−0.06 on the ladder, a third of its youth size). And a health limit's grip on the life rating is near its lifetime minimum (−0.07): by their late fifties people seem to fold physical limitation into expectations, even as it still drives worry hard (+0.26). The body leads here more through how it feels day to day than through how it colours the verdict on a life.
Women & men. The bill begins to come due: the female stress gap climbs back to +0.13 (worry +0.08) even as women's Thriving edge narrows to +0.08.
The body leads — self-rated health (+0.26 on the ladder, near its lifetime high) and functional limits move every outcome — but the surprise of this stage is political. Identifying Republican or Independent rather than Democrat carries the largest evaluative drag of the whole life course: Republicans −0.135 on the ladder and −0.21 on Thriving, Independents −0.15 on the ladder. And yet it leaves day-to-day affect untouched — worry, stress and daily happiness all statistically flat for Republicans — so partisanship colours how older adults judge their lives without changing how they feel hour to hour.
Two long-running gaps hit turning points here. The female stress-and-worry penalty is at its widest of any age (stress +0.17, worry +0.13), even as women still rate their lives higher than men (+0.13 on the ladder). And the Hispanic "happiness paradox" of youth has reversed: the early-adult calm (stress −0.24 at 26–39) is gone, and Hispanic young-old adults now report more worry than white peers, not less. Money's grip, dominant at midlife, has eased markedly (0.19 on standard-of-living satisfaction, down from 0.31), and the breakup scar is nearly closed (−0.07 on the ladder).
Women & men. The negative-affect gap is now at its widest of the whole life course — women report markedly more stress (+0.17) and worry (+0.13) than men — yet still rate their lives higher (+0.13). The paradox is most stretched here.
At the very end, two forces carry well-being. Self-rated health's pull on the life rating is at its lifetime peak (+0.28), and religion's dividend on daily happiness reaches its maximum (+0.19) — the culmination of a steady climb from just +0.04 in youth. Faith has fully traded places with itself: meaning in the young, mood in the old.
The mirror image is how much has fallen away. Money's grip collapses to its lifetime minimum (0.12 on standard-of-living satisfaction, down from a midlife 0.31 — the steepest age-decline of any predictor here), and on daily mood it is at its weakest of adulthood (worry −0.04). The penalty for having lost a partner disappears entirely (+0.00 on the ladder, not significant): when widowhood is near-universal, it stops marking anyone as worse off. And the emotional sting of a failing body is at its gentlest — a health limit lifts worry just 0.17, half its youthful 0.34. The oldest adults look emotionally adapted, reacting less to material shortfall, lost spouses and physical limits than people half their age — though the partisan drag, while easing from its young-old peak, is still present (Republicans −0.13 on Thriving).
Women & men. The paradox holds to the very end: women rate life notably higher than men (+0.14 on the ladder) but carry a large daily-feeling tax (stress +0.16, worry +0.12) — the evaluative advantage steady, the emotional cost still high.
The gender paradox, at every age
One pattern is so consistent across the six stages that it deserves its own card. At every age, women rate their lives higher than men — and at every age they also report more stress and worry. Both halves of the paradox hold across the whole adult lifespan. What changes is the size of the emotional tax: it is heavy in young adulthood, falls to its lightest in the busy midlife years, then widens to its lifetime maximum around retirement before easing only a little. The evaluative edge, by contrast, barely moves.
Coda: a different engine for men and women?
The card above tracks the gap between women and men. A harder question is whether the whole machine runs differently — whether money, health, marriage and the rest carry different weight for each sex. Re-fit separately for men and women, the answer is mostly no: the predictors point the same way and sit close to the same size for both. The one place the engine clearly splits is connection in old age.
For older men, living alone predicts a lower life rating (−0.06 at 75+) and a touch more worry (+0.02) and stress (+0.02); never marrying leaves them no better off (married vs never-married is −0.02 on the ladder and −0.01 on living-standard satisfaction). For older women, the very same circumstances are neutral or even easier — living alone goes with a slightly higher rating (+0.03) and less worry (−0.06) and stress (−0.06), and marriage still pays (+0.10 on the ladder, +0.14 on living-standard satisfaction). It is the familiar portrait of the isolated older man and the more socially resilient older woman, surfacing in the coefficients.
Three fainter differences round it out. Women's daily mood tracks their health more tightly — the protection good health gives against worry is larger for them at every older stage (men −0.12 vs women −0.18 at 65–74). Income's pull on the life rating is marginally a men's story (men +0.25 vs women +0.21 in midlife, and similarly into old age). And the conservative life-rating penalty is concentrated in men (Republican men −0.12 vs women −0.07 in midlife). None of these is large — most are under a tenth of a standard deviation — so the honest summary is one engine, tuned slightly differently, with its sharpest sex difference reserved for who is wounded by solitude late in life.
What this does and doesn't show
The shape is sturdy — every stage holds tens of thousands of interviews, the directions all check out, and a leaner universal-coverage model reproduces them — but the reading needs three guardrails.
It is a snapshot of ages, not a film of ageing. Gallup's poll is a repeated cross-section, so each "stage" is a different group of people, not the same people followed through life. An age pattern blends growing older with the different histories of different birth cohorts, and with who survives to be interviewed at 75+. Treat these as how well-being is distributed across the ages at one moment — the social reality each life stage presents — not as a guaranteed personal trajectory. Nothing here is causal.
Two of the arcs are conditional, and differ from the raw gaps. The Black–White life-rating crossover appears only after holding income and marital status constant: in raw averages, Black Americans rate their lives at or below white Americans until old age, because their lower average income masks an underlying within-circumstance advantage. The honest statement is "among people of the same income and family situation, Black Americans from midlife on rate their lives higher" — and, separately, that the Black advantage in felt stress and worry is robust even in raw means. The partisan gap is likewise net of religion and income, so it is not the raw Republican–Democrat happiness difference.
Health is special, and so is its measure. Self-rated health wins partly because it is the most proximal thing on the list — a subjective verdict that shares a vantage point with the life rating itself, open to reverse causation. Its lead is real and stable, but its size should be read as a ceiling. The deeper point survives all of this: the question "what makes people happy?" has no single answer because it has no single asker. It depends on when in a life you ask.