#622 1662 · John Graunt · Statistics / public health
A haberdasher read 70 years of plague-tracking records and invented demography by asking a different question of the same data
the problem
Decades of detailed death records existed, but only ever got read for one narrow purpose: knowing when to flee a plague outbreak
background
London's parish clerks had recorded weekly burial and christening counts since the late 1500s, standardized from 1603 onward, for one specific operational reason: tracking plague severity so city authorities and residents knew when outbreaks were worsening and it was time to shut theaters, restrict gatherings, or flee the city. For seven decades, the Bills of Mortality were read purely through that lens — a rising weekly death count meant more plague, nothing else.
John Graunt, a haberdasher and cloth merchant with no scientific training or institutional standing, had no stake in the plague-monitoring purpose the Bills were built for. He instead began treating the same seven decades of raw parish records as a dataset that might answer questions nobody collecting it had ever intended to ask — how mortality varied between the sexes, between city and country, by cause, and by age.
what everyone would do
For seven decades, everyone with access to the Bills of Mortality — parish clerks, city authorities, residents deciding whether to flee — read them for exactly the purpose they were collected: tracking weekly plague deaths to judge how bad an outbreak was getting. No one trained in medicine or natural philosophy treated the records as raw material for any question beyond that operational signal.
what they saw
Graunt, a tradesman with no institutional stake in the plague-monitoring purpose the records were built for, saw that seventy years of weekly numbers held far more information than the single question anyone had ever asked of them — patterns by sex, by cause, by age could all be extracted from data nobody had looked at that way, even though the raw records never directly recorded the thing he actually wanted to know.
the move
Working from records that didn't even list age at death, Graunt used ratio-based estimation techniques to reconstruct the first-ever life table, showing the probability of surviving to any given age for a population that had never been directly measured that way. He published his findings in 'Natural and Political Observations Made upon the Bills of Mortality' in 1662, presenting the Royal Society with 50 copies in person.
why it works
Seventy years of parish burial and christening records already existed as a byproduct of plague-monitoring, collected and funded for another purpose entirely, so Graunt could mine them at essentially zero additional cost beyond his own analysis. Because the records didn't list age at death directly, he used ratio-based estimation across the causes and volumes of death that were recorded to reconstruct that missing dimension indirectly, producing the first life table — a real, numeric answer to how long a person of a given age could expect to live, something no one had ever been able to state with actual data before. Because the underlying dataset was already running for another reason, the discovery cost Graunt only his reading and reasoning time, a return on effort no newly designed survey from scratch could match.
the payoff
Graunt's book effectively founded demography and vital statistics as fields of study, going through five editions, and the Royal Society elected him a Fellow on the strength of the work alone — a notable break from convention, since Fellows were normally drawn from gentlemen and scholars rather than tradesmen; King Charles II is reported to have told the Society to admit more tradesmen like him if they could find them.
where it breaks
This only works when a rich, long-running operational dataset already exists for some other purpose and someone has the analytical skill to extract signal the routine users never needed — a dataset too sparse, too short-lived, or too narrowly recorded to estimate around offers nothing to mine. It also depends on someone actually having the outsider's freedom to ask a different question: an insider embedded in the original purpose has every incentive to keep reading the data through its intended lens and little reason to wonder what else it might reveal, which is exactly why it took a cloth merchant with no stake in plague monitoring, not a physician or city official, to see it. And results extracted this way still need independent validation — Graunt's estimation techniques could be, and were, second-guessed, and the method only became broadly trusted once decades of subsequent demographic and actuarial work confirmed and extended it.
what came after
The life table method Graunt pioneered from repurposed plague-tracking records became, and remains, the foundational tool of actuarial science and public health statistics, and his core move — mining an existing operational dataset for a question its collectors never intended it to answer — is still the way demography and epidemiology routinely begin new inquiries into old records.
references
- [1]John GrauntWikipedia, 2025en.wikipedia.org
- [2]John Graunt at 400: Fighting disease with numbersSignificance (Wiley, Royal Statistical Society), 2020rss.onlinelibrary.wiley.com
- [3]History of Statistics 1: The Bills of Mortality, and the Beginning of StatisticsVermont Mathematics Initiative (American Statistical Association), 2010higherlogicdownload.s3.amazonaws.com