Do Left-Handers Die 9 Years Earlier? Tracing the 9-Year Myth
A friend of mine is left-handed. At dinner a few weeks ago, someone at the table cheerfully told him that left-handed people die about nine years earlier than everyone else.
He laughed. Then he got quiet. Then he asked me, half-joking, "Wait, is that actually a thing?"
It is a thing. It's just not the thing people think it is.
The claim that left-handers have a shorter life expectancy has been circulating since 1991, and it still shows up in trivia posts, comment threads, and — this is the strange part — inside peer-reviewed papers. So I went looking for where the number actually came from.
What I found was three different studies, three different numbers, and one 2023 simulation that quietly took the whole thing apart.
Have you heard this one before? Let's trace it properly.
So do left-handed people really die 9 years earlier?
No. The mainstream position among handedness researchers is that the nine-year gap was a methodological artifact — a side effect of how the study picked its subjects, not a biological finding.
The short version: the original researchers surveyed people who had died in a single year, rather than following a group of people born in a single year. That one choice is enough to manufacture a nine-year gap out of thin air, which is exactly what a 2023 simulation demonstrated.
But "artifact" doesn't mean the whole literature is a flat zero, and I'll get to the leftovers honestly near the end. First, the origin story — because it's messier than you'd expect.
Where did the 9-year claim come from?
The nine-year figure comes from a single short item in the New England Journal of Medicine on April 4, 1991 — and it was one of three different numbers the same two researchers published in three years.
The pair is Diane F. Halpern and Stanley Coren. Here's what they actually published:
| Year | Where | What they looked at | The number |
|---|---|---|---|
| 1988 | Nature 333:213 (Scientific Correspondence) | Every player in The Baseball Encyclopedia with birth date, death date and throwing/batting hand | Left-handers died about 8 months younger |
| 1991 | NEJM 324(14):998, April 4 (correspondence / brief report) | 987 deaths from two Southern California counties in 1989; surveys mailed to next of kin | Mean age at death 66 vs 75 — a 9-year gap |
| 1991 | Psychological Bulletin 109(1):90–106 (full paper) | The long-form theoretical version | About 8 years for men, 5 years for women |
Read that table again, because it's the whole story in miniature.
The famous "nine years" did not come from the Nature paper. That one found eight months. The number inflated more than thirteenfold between 1988 and 1991 — and it wasn't because anyone discovered new biology. It was because the method changed.
One more detail that matters more than it sounds. The NEJM item was a short correspondence piece, not a peer-reviewed full research article. That's worth pausing on, because "published in the New England Journal of Medicine" is precisely what gave the claim its staying power in every retelling since. A one-page letter earned the authority of a landmark trial.
The full-length version in Psychological Bulletin did make a real argument. Titled "Left-handedness: A marker for decreased survival fitness," it proposed that left-handers face more accident risk from a right-handed environment, and that left-handedness itself marks birth stress, developmental irregularity, and immune-system deficiency from the prenatal hormonal environment.
So the mechanism was proposed. It just didn't survive contact with the data.
What if the real difference were zero? The 2023 simulation that changed everything
In July 2023, a team led by Juan Lavista Ferres published a simulation in Archives of Public Health that built a model of US deaths in 1989 — the exact year Halpern and Coren pulled their records — while assuming, by construction, that handedness had absolutely no effect on age at death. The simulation produced a 9.3-year gap anyway.
That's the moment the mystery cracks open. Let me spell out what they did.
They took US mortality for 1989. They assumed zero true difference between left- and right-handers. Then they applied historical left-handedness rates by birth cohort, estimated from Chris McManus's analysis of a survey of roughly 1.2 million people.
Result: an apparent gap of 9.3 years (95% CI 4.3–14.4) — statistically indistinguishable from the reported nine years, from a model with no real effect in it at all.
Then they ran it again with one thing changed. They held the left-handedness rate constant over time, as if the proportion of left-handers had never shifted across the 20th century.
The gap collapsed to 0.02 years.
Using estimated age of death per birth cohort instead, the mean difference came out at 0.43 years.
The authors' conclusion is about as blunt as journal writing gets: "The changing rate of left-handedness reporting over the years entirely explains the originally reported observation of nine-year difference in life expectancy."
Nine point three years, conjured from nothing. Two hundredths of a year once the bias is removed. That's not an argument about the finding — it's a reconstruction of it.
Why does the illusion happen? Because left-handedness wasn't always allowed
The gap appears because the share of people who identify as left-handed rose dramatically over the 20th century — so in any given year's deaths, the left-handers are mechanically drawn from younger birth cohorts.
Here's the hard data behind that.
In 1986, National Geographic ran a scratch-and-sniff smell survey as a magazine insert. The researchers, Gilbert and Wysocki, tucked in two questions about which hand people wrote and threw with, because they suspected a link between smell and cerebral dominance. Over 1.1–1.2 million usable responses came back — still the largest handedness dataset ever collected.
What it showed, by year of birth:
- Only about 3–4% of people born before roughly 1920 were left-handed.
- About 11–12% of those born after 1950 were — roughly a threefold difference.
- By age at the time of survey: about 13% left-handed among respondents under 40, falling to about 5% by age 80.
McManus's historical reconstruction fills in the curve: left-handedness fell through the Victorian era to a low of about 3% among people born around 1895, then climbed steeply, flattening out for those born after about 1945–1950 at roughly 11–12% of men and 9–10% of women in Western countries. Reported rates across 1900–1988 ranged from 2.5% to 12.6%.
Nobody thinks human biology changed that fast. What changed was school policy. From the late 19th century through much of the 20th, left-handed pupils across Europe, North America and Australia were routinely required to write with the right hand, sometimes enforced by striking the left hand. The practice persisted into the 1940s–60s and faded at different rates by region.
So picture a man who died in 1989 at age 85. He was born in 1904. If he was born left-handed, there's a good chance a teacher made sure his family never described him that way.
Now his next of kin fills out a survey asking which hand he wrote with. They write "right."
Multiply that by a whole generation, and the elderly end of your sample turns artificially right-handed. Left-handers look like they die young, because the old left-handers were edited out of the record decades earlier.
Chris McManus, the UCL psychologist who wrote Right Hand, Left Hand, puts the error plainly: the results were "entirely the result of a statistical artefact," caused by studying death cohorts rather than birth cohorts, because "handedness is in fact a marker of being born later in the 20th century."
He also offers my favorite sanity check. If a nine-year effect were real, it would be the single largest known influence on human lifespan in modern healthcare — roughly equivalent to smoking about 120 cigarettes a day. A hundred and twenty. Would we really have missed that?
And there's the analogy McManus and economist Alex Tabarrok both use: survey the recently deceased about whether they ever read Harry Potter. The readers will appear to die shockingly young. Not because the books are lethal — because readership is concentrated in younger birth cohorts.
What happened when researchers fixed the study design?
Every study built to avoid the cohort trap — prospective follow-up, twin designs, archives that include the still-living — has come back with no meaningful lifespan difference.
The objection landed almost immediately, by the way. Within weeks of the NEJM item, the Roanoke Times (March 18, 1991) quoted three handedness researchers pushing back: Paul Satz of UCLA urging caution, Alan Searleman of St. Lawrence University calling the findings "so outrageous" and arguing an effect that large would have been noticed long ago, and Philip Bryden of Waterloo challenging the statistics and sample bias. That article already named forced hand-switching as the alternative explanation.
Then came the formal work:
Salive, Guralnik & Glynn (1993), AJPH 83(2):265–267 — from the National Institute on Aging. Their simulation showed the lower mean age at death among left-handers "could be explained exclusively by the age distribution of laterality." Their 6-year prospective cohort of 3,774 older adults in East Boston found no association with mortality: relative odds 1.04 (95% CI 0.79–1.36).
Aggleton et al. (1994), BMJ — 5,960 first-class cricketers born 1840–1960 (3,387 deceased, 2,573 alive). No significant relation between handedness and mortality.
Basso et al. (2000), Epidemiology — Danish twins born 1900–1910. No evidence of differential survival. Among 118 opposite-handed twin pairs, the right-handed twin died first in 60% of dizygotic pairs and 50% of monozygotic pairs. Twins share a birth year, so the cohort artifact is eliminated by design.
Lauren J. Harris (1993), Psychological Bulletin 114(2):203–234 — a commentary reviewing both Coren and Halpern studies, concluding "the case for the 'decreased survival fitness' hypothesis cannot be sustained."
The baseball data didn't hold up either. Fudin and colleagues (1993) re-examined the same underlying records and found they "do not indicate a relationship"; Hicks and colleagues (1994) reached a similar negative conclusion.
One thing I want to be fair about: Halpern and Coren replied to Harris in Psychological Bulletin the same year. This was a genuine two-sided academic exchange over several years, not a single knockout punch.
Is being left-handed dangerous? And where did "2,500 lefties die every year" come from?
The accident-risk claim is much weaker than it sounds, and the widely shared "2,500 left-handers die each year using right-handed equipment" figure has no primary source at all.
Let's take them separately.
Coren's 1989 paper in the American Journal of Public Health surveyed 1,896 college students in British Columbia and reported elevated accident risk for left-handers — roughly +20% for sports injuries up to +80% for driving-related injuries. The 1991 Psychological Bulletin paper went further, reporting left-handers were more than five times as likely to die in industrial and auto accidents.
But a re-analysis pointed somewhere else. Voyer and Voyer (2015) found that weak cerebral lateralization — not left-handedness itself — correlates with accident proneness, and found no association between hand-preference scores and direct measures of hemispheric asymmetry. Their verdict: "the stereotype of the clumsy left-hander might be a myth." The elevated injury risk in Coren's data appeared to attach to mixed-handers, not left-handers.
Now the 2,500 number. You've probably seen it. It does not appear anywhere in Coren's 1989 paper, and no research supporting it has ever been located. Snopes rates it as unsupported.
It's an unsourced statistic that has been circulating online for decades, passed hand to hand, gathering authority purely from repetition. Which, honestly, is the same thing that happened to the nine years — just without even a letter to point to.
Honestly, it's not a perfect zero
"The nine-year gap was an artifact" is well supported. "Handedness has exactly zero relationship to mortality" is not the same statement, and I don't want to overclaim.
Here's what's genuinely left on the table.
A small non-zero result. Peter Rogerson at the University at Buffalo reworked baseball data in 1994 using 4,448 players born before 1920 and — crucially — including players still alive, which earlier work hadn't. Ages 25–65 showed little difference. Ages 65–85 did: 26% of right-handers versus 21% of left-handers survived to 85. Overall, right-handers lived "close to one year longer, plus or minus a little more than half a year." That's about one-ninth of the claimed effect, with wide uncertainty — but it isn't zero.
A study that found the opposite, in women. Ramadhani and colleagues (2007, Epidemiology) followed 12,178 middle-aged Dutch women for nearly 13 years. After adjusting for age, socioeconomic status, BMI and smoking, left-handed women had 1.36× the risk of all-cause death. Cause-specific hazards: any cancer 1.7 (95% CI 1.0–2.7), colorectal cancer 4.6 (1.5–14), breast cancer 2.0 (0.83–4.6), cerebrovascular disease 3.7.
Read those confidence intervals, though. Several touch or cross 1.0, and the colorectal interval of 1.5–14 is enormous, which means very few actual events. The authors themselves flagged the evidence as far from conclusive and possibly a chance finding — and the same issue of the journal carried a commentary titled "Right or wrong? On the difficult relationship between epidemiologists and handedness." The journal knew it was contested. Note also that this design controls for birth cohort, so it isn't vulnerable to the Halpern–Coren artifact — and it still found nothing remotely like nine years.
And the claim is still alive inside science. In December 2023, a study in Frontiers in Cardiovascular Medicine looked at 379 participants aged 18–50 (only 46 of them left-handed) and found lower flow-mediated dilation in left-handers (6.1% vs 7.6%, p = 0.043), which the authors calculated as roughly a 14% increased cardiovascular risk. They explicitly cited the "9-year decrease in lifespan" literature and suggested their findings might partly explain it — while acknowledging their design "cannot infer cause and effect."
So in the same year one team published a simulation showing the nine years was an artifact, another team was citing it as a live premise. That's how these numbers survive.
One correction while I'm here, because I've seen this cited wrong constantly: the famous "89 studies, 21,000 patients" figure from McManus and Bryden is real, but it is not about lifespan. That 1994 meta-analysis (Bryden, McManus & Bulman-Fleming, Brain and Cognition 26:103–167) covered 89 studies by 25 teams with over 21,000 patients and 34,000 controls, and its finding was that left-handers show no systematic tendency to suffer from immune disorders. It knocks down the mechanism Coren and Halpern proposed, not the mortality claim directly. If you see it described as "a meta-analysis finding no lifespan difference," two different things have been compressed into one.
Frequently asked questions
Q. Do left-handed people really die 9 years earlier?
A. No. The figure comes from a 1991 letter in the New England Journal of Medicine comparing mean age at death among 987 people who died in Southern California in 1989. A 2023 simulation in Archives of Public Health reproduced a 9.3-year gap while assuming zero real difference, and the gap fell to 0.02 years once historical changes in left-handedness rates were removed.
Q. Where did the 9-year claim come from?
A. Diane Halpern and Stanley Coren. Their 1988 Nature item on baseball players found an 8-month difference; the April 4, 1991 NEJM correspondence found 9 years; the 1991 Psychological Bulletin paper reported about 8 years for men and 5 for women. All three get quoted as if they were one finding.
Q. Is being left-handed dangerous?
A. Not in the way the old claim suggested. A 2015 re-analysis attributed elevated accident risk to mixed-handedness and weak cerebral lateralization rather than left-handedness itself. The often-repeated "2,500 lefties die each year using right-handed equipment" has no traceable primary source.
Q. What percentage of people are left-handed?
A. The best current estimate is 10.6%, from a 2020 meta-analysis in Psychological Bulletin, with a range of 9.3% under the strictest criterion to 18.1% under the most lenient. Men are about 25% more likely to be left-handed than women.
Q. Were left-handed children really forced to switch hands?
A. Yes. Across Europe, North America and Australia, left-handed pupils were routinely required to write with the right hand from the late 19th century into the mid-20th, sometimes enforced physically. I couldn't find one authoritative "end date" for the practice in the US — but the prevalence curve tells the story: about 3–4% left-handed among people born before 1920 versus 11–12% among those born after 1950.
The next time someone tells you a scary statistic
If you take one habit away from this, make it this one: ask who was counted, and when.
Not "is the number big?" Not even "was it peer-reviewed?" Just — who ended up in the sample, and could the sample itself have produced this answer?
That single question is what separates a nine-year gap from 0.02 years. Same world, same data, different sampling frame.
My left-handed friend, for the record, is fine. He's also now the person at dinner who says "wait, was that a death cohort or a birth cohort?" — which I consider a public service.
Which one were you before you read this?
- I'd heard the 9-year thing and half-believed it 🤔
- Never heard of it until now 😮
- I'm left-handed and I've been hearing this my entire life 🖐️
Drop a 1, 2, or 3 in the comments — I'm genuinely curious how far this one traveled.
Comments
Post a Comment