Researchers followed more than 71,000 people and found that the most optimistic were up to 70% more likely to live past 85 than the least optimistic — an association that remained even after accounting for exercise, diet, smoking, illness and depression.

Seventy per cent is the kind of number that can turn a subtle epidemiological result into a slogan.

In this case, the underlying finding was genuinely striking. Across two long-running American cohorts, people with the highest optimism scores lived longer on average and had better odds of reaching 85 than people with the lowest scores. The association persisted after researchers accounted for depression, existing illness and a long list of health behaviours.

But the shortest version of the result also blurs three details that matter. The paper reported odds, not a direct increase in probability. The 70 per cent estimate was reduced when lifestyle variables were added. And although the full study followed more than 71,000 people, the specific age-85 calculation used the smaller group who were old enough to reach that birthday during follow-up.

None of those qualifications erases the result. They tell us what the result actually was.

The researchers joined two very different cohorts

The 2019 study, published in PNAS, analysed 69,744 women from the Nurses’ Health Study and 1,429 men from the Veterans Affairs Normative Aging Study. Together, the cohorts contained 71,173 participants.

The women had been followed with regular questionnaires since 1976. They completed a six-item optimism assessment in 2004, at an average age of 70, and their deaths were tracked until 2014. The men had entered the veterans study beginning in 1961. Their optimism scores came from a personality inventory completed in 1986, at an average age of 62, and they were followed through 2016.

That gave the researchers ten years of follow-up for women and 30 years for men. It also gave them replication across separate groups using different measures of optimism. In both cohorts, greater optimism at baseline was associated with longer subsequent survival.

The differences between the cohorts impose limits. The women were nurses; the men belonged to a veterans ageing study. Both samples were predominantly white and had higher socioeconomic status than the general population. Because sex, occupation, cohort and measurement method all changed together, apparent differences between women and men cannot be treated as clean sex effects.

Only participants old enough to reach 85 entered that analysis

The full 71,173-person sample was used to model lifespan. Researchers compared survival time across levels of optimism while adjusting progressively for demographic factors, health conditions and behaviour.

The question of “exceptional longevity” required another step. A person born too late to turn 85 before the study ended could not fairly be classified as having failed to reach 85. The team therefore restricted that analysis to participants who had enough potential follow-up.

The age-85 calculation included 13,045 women and 1,117 men, a total of 14,162. Many popular accounts, including the Boston University summary, accurately describe the study as involving 69,744 women and 1,429 men. It is still important to distinguish the size of the two cohorts from the size of the subset used for this particular result.

The endpoint itself also needs context. The authors called survival to 85 “exceptional longevity” because it was well beyond average life expectancy for people born early in the twentieth century. Eighty-five was a research threshold, not a claim that 84 represents ordinary life while 85 is biologically exceptional.

The 70 per cent figure was one stage in the modelling

Researchers built the adjustment in layers. In models accounting for demographics, baseline health conditions and depression, women in the highest optimism quartile had 50 per cent greater odds of reaching 85 than women in the lowest quartile. Men in the highest optimism group had 70 per cent greater odds than those in the lowest group.

That is the origin of the widely reported 50-to-70 per cent range.

When the researchers additionally adjusted for smoking, alcohol use, physical activity, diet, body mass and primary-care visits, the estimates were attenuated. The highest-versus-lowest comparison became an odds ratio of about 1.2 for women and 1.6 for men, or roughly 20 per cent and 60 per cent greater odds. Confidence intervals were wide at their lower boundaries, particularly in the much smaller male sample, but the overall trend across optimism levels remained statistically significant.

The lifespan models told a related story. Before lifestyle adjustment, the most optimistic women had an estimated lifespan 14.9 per cent longer than the least optimistic women; the corresponding estimate for men was 10.9 per cent. After health behaviours were included, those estimates fell to 8.7 and 9.8 per cent.

So the association did remain after exercise, diet, smoking and other behaviours were considered. It did not remain unchanged.

Greater odds do not mean the same thing as greater probability

“Seventy per cent more likely” is familiar language, but the statistical result was an odds ratio of 1.7. Odds and probability are related but not interchangeable.

If an outcome has a probability of 50 per cent, its odds are one to one. Multiplying those odds by 1.7 produces a probability of about 63 per cent, not 85 per cent. If the starting probability is 10 per cent, the same odds ratio produces a probability of roughly 16 per cent. The absolute difference depends on the baseline.

This matters because odds ratios can sound larger when translated casually into percentage likelihood. The study supports the conclusion that the most optimistic groups had meaningfully higher odds of reaching 85. It does not mean optimism added 70 percentage points to anyone’s chance of doing so.

It is also worth noting that optimism was not a momentary report of feeling cheerful. The questionnaires measured dispositional optimism, broadly the expectation that good things will happen or that the future will be favourable. A person could experience grief, anger or a terrible week and still score relatively high on that longer-term orientation.

Adjustment cannot tell us which pathway is causal

More optimistic participants differed from less optimistic participants in ways that could affect longevity. They were generally more physically active, and the groups differed on education, income, diabetes, depression, alcohol use and other characteristics. Statistical adjustment was designed to compare people who were more similar on those measured factors.

Health behaviour creates a particularly interesting problem. Exercise, diet and smoking may confound the association if they independently influence both optimism and longevity. They may also be mediators if optimism helps someone persist with exercise, resist an unhealthy impulse or recover from a setback. Adjusting them away can therefore remove both bias and part of a possible mechanism.

The authors proposed several other routes worth testing. Optimistic people may regulate emotion more effectively under stress, reframe difficulty as a challenge and adjust goals when an original plan becomes impossible. Optimism has also been associated elsewhere with cardiovascular, metabolic, immune and lung-function markers. This study did not establish that any one of those pathways caused its survival pattern.

That distinction resembles the one behind our reports on running and premature mortality and book reading and longer survival. Longitudinal associations can persist after careful adjustment and still stop short of proving that changing the exposure will change lifespan.

Observational evidence is not a prescription to think positively

Participants were not randomly assigned to optimism. The researchers measured an existing psychological trait and watched what happened later. Better underlying health could make the future seem brighter, while social advantage could support both optimistic expectations and longer life.

The team took reverse causality seriously. It adjusted for major illness and depression, excluded people who died early in follow-up and repeated analyses after removing those with major chronic diseases at baseline. The pattern remained. Those checks make the simplest “healthy people merely feel more optimistic” explanation less convincing, but they cannot remove every unmeasured difference.

The finding also sits within a broader literature. A systematic review and meta-analysis of 15 studies, covering 229,391 participants, found optimism associated with lower cardiovascular risk and lower all-cause mortality. For all-cause mortality, the pooled relative risk was 0.86. The studies showed substantial heterogeneity, and the authors detected evidence of publication bias, so even the wider literature is not perfectly tidy.

Brief psychological exercises can change reported optimism, but no trial has shown that teaching optimism makes people live to 85. That causal question would require different evidence and an exceptionally long horizon.

The safest interpretation is neither cynical nor magical. Optimism may be a genuine psychological resource, partly expressed through behaviour and partly connected to how people respond to stress. It may also reflect circumstances that the questionnaires could not fully capture.

What the study found is an association strong enough to investigate. What it did not find is that illness reveals a failure to think positively, or that optimism is a longevity guarantee. Hope may travel with a longer life. This study cannot yet tell us how much of the journey it drives.