Should You Follow the Data? Don’t Believe the Hype

“Follow the data,” that now-ubiquitous aphorism, might seem like a novelty of our digital age. But it’s a throwback, really, to a long-standing philosophical doctrine called empiricism. In the 17^th century, empiricists like Francis Bacon and John Locke urged their countrymen, too, to follow the data or, in other words, the facts given to them by way of observation or sensory experience.

Today, this advice may seem obvious, trite, or banal. Back then, it was revolutionary. In those days, authority figures worked overtime to convince the masses that their curiosity was a mere distraction. All that needed to be known was already known, they claimed, and there it was for everyone to see in holy books, ancient texts, and the official pronouncements of elites. Scout’s honor.

It was amid this dehumanizing, authoritarian climate that the empiricists swaggered onto the stage, snatched the microphone, and boldly warned a la the rap group Public Enemy: “Can’t Truss It.” In other words, don’t you dare take their word for it. New knowledge is both desirable and attainable if you rely on the evidence of your own senses. Anyone willing to put their theories and prejudices aside, the empiricists promised, could derive from this data certain, justified knowledge.

What an explosion this optimistic message sparked. Soon, people started thinking for themselves. They started observing and experimenting and performing what we now call modern science. They began attempting, more generally, to change their world for the better.

There was just one problem: empiricism was, in many cases, wrong. Staggeringly wrong.

First, even if we could somehow clear our minds of all theories, the data we then observed would be meaningless. Theories—our explanations of how the world works—are the way we make sense of data. So then, the idea of following the data is a canard—the very same data when viewed in light of different theories can result in starkly different conclusions. A smoldering building for example might seem like an accident to a firefighter, arson to a police detective.

Secondly, data, like all sources of knowledge, is frequently wrong and always incomplete. The data we receive from our eyes, though magnificently rich, fails to represent an astonishing amount of what’s actually out there in reality—bacteria, microwave radiation, distant stars, for example. Even the “big” data that social media companies amass, housed in gargantuan data centers distributed across the globe, represent just a tiny sliver of the infinitely complex hopes, dreams, preferences, and personalities of their users.

More to the point, none of this data can be assumed true. As the physicist David Deutsch points out in The Beginning of Infinity, “amending the data, or rejecting some as erroneous, is a frequent concomitant of scientific discovery.” Indeed, heliocentrism (the idea that the Earth revolves in orbit around the Sun) was discovered as Galileo put it, in “violent opposition” to sense data—namely, the feeling of the Earth’s stillness beneath our very feet.

So if following the data is not only unwise but literally impossible, how were empiricism-inspired scientists actually using data to gain new knowledge? The answer: by using it to criticize theories that had already been guessed.

The philosopher Karl Popper once summed up this more sound conception of how science proceeds with three words: problems—theories—criticism. First, rather than starting with data, scientists begin by identifying problems, that is to say weaknesses or insufficiencies in their existing theories or expectations. Then, they conjecture new theories that they hope might solve those problems. Finally, they criticize those theories as severely as they possibly can. It’s during this last step where observation and data play their most important role. Rather than being the source of our theories, data serves as a potential destroyer of them. Data is a bazooka we fire at our theories to see which ones can withstand the blast.

Should one theory survive while its rivals do not, we’re right to consider it our best explanation. But, and this is crucial, we don’t anoint it as “true” or even “probably true,” nor do we claim it as “justified” in any sense. In stark contrast to the certainty professed by empiricists, Popperians view even their best theories as conjectural, problematic, and provisional—expected to be overthrown in the future by even better theories.

Fantasies Essential Reads

How Our Sexual Fantasies Change Throughout Our Lives

Why Animated Porn Is So Popular

We owe empiricism and its “follow the data” mindset an enormous debt of gratitude for propelling science forward and for initiating an unprecedented growth in knowledge. But we mustn’t forget: it’s a fabrication, a fairy tale, a fiction. Granted, insofar as it reminds us to rebel against authority with respect to knowledge—or as Public Enemy put it, “Fight the power”—perhaps it remains a useful fiction. But insofar as it misleads us into believing that conclusive knowledge can somehow be mechanically derived from data, Public Enemy once again said it best: “Don’t believe the hype.”