The World is a Laboratory

Much of what we know about the world comes from experiments. But few aspects of human society can be contained in a laboratory. History offers a powerful alternative tool: natural experiments.

Thank you for reading! Some of this essay is only available to paid subscribers, so please consider upgrading to unlock all of it—as well as full access to 160+ other essays. You might also consider giving the gift of a paid subscription to an intellectually curious soul this holiday season, or giving them a copy of my new book, FLUKE. Also, a ridiculously lovely, surprising thing happened: I was recently named one of the world’s 25 “Top Thinkers” by Prospect Magazine. If you’d like to vote for me, you can do so here.

---

I: Britain’s Great Inadvertent Experiment of 1953

Why was a British baby born after September 1953 significantly more likely to die from disease later in life than a British baby born before September 1953?

Pause for a moment before reading further. Try to consider why that might be the case; play social scientist and develop your best hypothesis.

---

In many ways, September 1953 was a relatively unremarkable month: several planes crashed, JFK got married, and Nikita Khrushchev became the head of the Soviet Central Committee. But one largely forgotten event did take place on September 26, 1953—and it would shape the fortunes of countless unborn children for decades to come.

Benevolent writer that I am, I offer one clue, before the grand Poirot-style reveal:

On September 26, 1953, the British government ended sugar rationing. Shortly after the onset of World War II, from January 1940 most British adults were allowed a maximum of 40 grams of sugar per day, a limit far below what most people would have preferred. This enforced rationing persisted well after the war ended, with restrictions ending more than eight years after the end of hostilities. Then, suddenly, British consumers could buy and consume sugar in great heaps, an abrupt shift on a single day. As the BBC reported of the occasion:

Children all over Britain have been emptying out their piggy-banks and heading straight for the nearest sweet-shop as the first unrationed sweets went on sale today. Toffee apples were the biggest sellers, with sticks of nougat and liquorice strips also disappearing fast…Adults joined in the sugar frenzy, with men in the City queuing up in their lunch breaks to buy boiled sweets and to enjoy the luxury of being able to buy 2lb boxes of chocolates to take home for the weekend.

This graph showcases how the end of sugar rationing created an abrupt spike in sugar intake for an entire population, a huge surge in the span of just weeks.

Little did the British government recognize that they had just laid the groundwork for researchers in 2024 to reveal insights from a profoundly interesting natural experiment. Unlike most experiments, natural experiments aren’t guided by scientists in lab coats swirling beakers and pipetting test tubes. Rather, researchers exploit naturally occurring phenomena from history to answer a research question. In the 1953 case, there’s a clear-cut—though unintentional—experiment, with the entire British population as guinea pigs:

What happens when a population abruptly goes from low to high sugar intake levels?

The answer: babies conceived and born in the rationed, low-sugar period are significantly less likely to develop type 2 diabetes, hypertension, and obesity later in life than those born in the free-for-all sugar intake period. Specifically, the researchers found that “early-life exposure to sugar rationing led to a reduction in T2DM and hypertension risk by about 35% and 20% and delayed the onset of these diseases by about 4 and 2 years, respectively.”

The positive health effects of being conceived, born, and raised during the sugar rationing were therefore substantial—equivalent, as Lauren Leffer points out, to quitting smoking or being a lifelong vegetarian.

What’s particularly powerful about this study is that it has automatic controls built-in to account for other factors. The “before” and “after” populations were basically identical in every way except for their access to sugar in their diet. What mattered for later health, it turned out, was the level of sugar in an infant’s diet during the first 1,000 days after conception.

Moreover, this “natural experiment” made research possible that would have been impossible (or deeply unethical) in a lab. Imagine taking thousands of babies and randomly assigning them to high or low sugar diets for the first 1,000 days—just to see how many of them got diabetes fifty years later. Not exactly the kind of study that would pass the ethics review board!

II: Our World is a Giant Laboratory

As

Adam Mastroianni

put it, “the ten-second history of science goes like this: for about 1200 years, people scribbled in the margins of Aristotle. Then one day Francis Bacon said “hey guys let’s do science” and people were like “sounds good.”

Bacon’s inductive reasoning was a crucial precursor to the birth of modern science, which now uses carefully controlled experiments as its gold standard. These experiments rely on a rigorous logic of deliberately isolating and manipulating key variables of interest—while holding other aspects constant—to identify how one thing changes another.

Within human populations, randomized controlled trials are often favored for testing medicines, for example, because by randomly sorting people into two groups, researchers are making reasonable approximations for “holding all else constant.” If you take 2,000 people and randomly allocate them to two groups, there will be differences between the two sets of 1,000 people, but those differences won’t be systematically biased—you’ll tend to have a similar distribution of height, weight, class, race, health risk factors, and so forth.

If you then give one group medicine and the other gets a placebo, that approach allows pretty solid inference that any variation in the experiment’s results is due to the medicine rather than to some other confounding factor from sorting bias. In this realm, natural experiments are viewed as merely the researcher’s second best option, the bargain replacement when the gold standard isn’t possible. But is that really true?

There are two major problems with a straight-up experimental approach for studying some of the biggest questions we care about: those involving large scale change within complex human societies.

First, there are those pesky problem of ethics and logistics. For my doctoral research, for example, I studied whether rigged elections increase the subsequent risk of political violence, particularly in the form of coups d’état and civil wars. (Spoiler alert: the answer is “yep, by a lot!”)¹

Now, imagine if I were to try to test that same hypothesis using an experiment. I would quite literally need to randomly assign a bunch of countries to two groups, rig elections in precisely half of them, and then see how many erupted into terrible violence. You can probably see why I didn’t take that approach.²

Second, there’s an often overlooked problem even with carefully designed experiments: the simplicity of the laboratory can create a profound mismatch with the complexity of the real world—not to mention that people being studied behave differently.

Our lives unfold embedded within social bonds, skewed cultures, our physical and natural environments, a web of infinite confounding. And the real story of change in our societies and our lives is often written in the margins, not always from big, obvious variables. This never-ending complexity means that some problems that we care about deeply cannot be so simply isolated in a laboratory—and that even if we could do it, the results from the proverbial test tube might actually differ from what would happen outside the laboratory, in our endlessly baffling world.

Thankfully, history offers a remedy, because history can be thought of as one giant human experiment. Even without any directed attempts by boffins in lab coats to manipulate our world for research, with clever examination of the past, we can still infer incredible truths about our species—about who we are, and about how key factors create chaos or order, destitution or prosperity, a thriving society or a collapsing one.

So, buckle up, because the world of natural experiments is a beguiling one.³ And, as we’ll see, it can provide us with crucial insights into a mind-boggling diversity of human change—from how Russian holidays affect the behavior of digital trolls hoping to influence American elections; to how geology and geography pushed one Polynesian society toward war and cannibalism and another toward prosperity under centralized God-kings; to lessons for humanity from spiders and snakes in Guam; and how ethnic hatred and terrifying violence can sometimes be unleashed—or avoided—by seemingly minor, arbitrary tweaks, such as some well-timed rainfall before a tragic genocide.