Can Math Save Journalism?

From Mathematician to Pulitzer Winner: Julia Angwin's Case for Journalism as Proof

Julia Angwin is the founder of Proof News and the newly launched Independent Media and Audience Project at Harvard. Her latest book “On Courage” is out now. Full conversation on Breaking Math on YouTube, Apple Podcasts, and everywhere else streaming.

---

There’s a particular kind of person who majors in math and then spends the rest of their life explaining why that wasn’t a mistake. Julia Angwin is one of them. So am I. So, it turns out, is nearly everyone worth talking to on this show (occasionally other science fields included).

Julia told us that announcing her math major at University of Chicago parties functioned like a small, precise repellent. I believe her completely; I have done a similar experiment in rooms full of writers and watched the exact same reaction — a polite nod, a step backward, and a hunt for the cheese table. Noah pointed out this isn’t unique to journalism circles. It’s purely the reaction math gets everywhere it wasn’t expected to show up. I swear that we are all much more livelier than the title holds.

Julia expected it to show up everywhere. She grew up in Palo Alto with a mathematician father who spent two decades as the lead mathematician translating two-dimensional maps into digital space at one of the original digital mapping companies — so if your GPS has ever routed you into a lake, that’s arguably his professional lineage, though he’ll tell you it’s the routing software’s fault, not the underlying data. Her mother was a nuclear chemist working on pipe corrosion, the unglamorous math-adjacent work that keeps radiation where it belongs. Julia has said she grew up believing there were exactly two life choices: hardware or software. Math wasn’t a subject in her house. It was the only known language.

What struck me most, listening back to this conversation, wasn’t the math pedigree. It was what she did with it once journalism got its hands on her.

The quote-hunting problem

Julia has a phrase for a huge amount of daily journalism: quote hunting. You go out into the field, you bag a few quotes, bring them home, and call that a story. She’s not wrong that this works fine for low-stakes reporting — nobody needs a rigorous sampling methodology to cover a Safeway opening. But for anything that matters — systemic injustice, algorithmic manipulation, and the actual mechanics of power — three anecdotes aren’t evidence. They’re a tip, not proof.

This is the sentence that reorganized the whole conversation for me: journalism’s job is to establish that something happened, and establishing that something happened is, functionally, an exercise in proof. Now, Julia has spent her career treating it that way, first at the Wall Street Journal and ProPublica, then at The Markup, and now at Proof News — each outlet a step further from anecdote and closer to something you could defend in front of a skeptical mathematician.

The Amazon story

Now you may be wondering what this proof looks like… The clearest demonstration of what this looks like in practice is the investigation Julia’s team ran at The Markup into Amazon’s search rankings. They’d already done the legwork on Google, measuring how far down the page independent, non-Google results landed. (More than half the time: off the first screen on your phone.) Amazon posed a harder problem, because its results aren’t a single scrollable line, they’re a grid, and there’s no obvious place your eye lands first.

So they talked to statisticians. They landed on a decision-tree approach: pick one slot — the first result — and test which attributes predicted landing there. Reviews? Fulfillment status? Whether the product was Amazon’s own house brand? The finding was stark enough to survive translation into plain English: being an Amazon-owned product was eight times more predictive of the top slot than any other factor they tested.

Readers understood it. More importantly, the House Antitrust Committee understood it — and used the finding to confront Amazon’s own congressional testimony, ultimately referring the company to the Department of Justice over a possible perjury. That’s the kind of accountability Julia is chasing: not “some people feel like Amazon favors itself,” but a number, arrived at transparently, that a committee could put in front of an executive under oath.

Ingredients labels for proof

At Proof News, Julia built something close to a nutrition label for journalism: hypothesis, sample size, techniques, findings, limitations, printed on every story. It’s an honest answer to a dishonest problem. A huge amount of the collapse in trust toward journalism, she argues, comes down to corporate ownership misaligned with the actual goals of reporting — but some of it is also a habit of over-declaring certainty that was never really there. Proof News is her attempt to borrow the discipline of the scientific method — hypothesis, test, transparency about limitations — without pretending journalism can ever reach the closed, self-contained certainty of a mathematical proof. In math, you build the whole cathedral yourself and then prove things inside it. In the real world, you’re stuck measuring a cathedral someone else built, with tools that are always slightly wrong.

From algorithms to authoritarianism

The second half of our conversation went somewhere different than our usual episode, but it is still quite timely: Julia’s new book “On Courage,” written with Ami Fields-Meyer on resisting authoritarianism, built from interviews with dissidents around the world. The connective tissue, she argues, is that the same handful of companies that decide what shows up first in your Amazon search also decide what news lands in your feed — and control over that layer of intermediation is, increasingly, the actual game of authoritarian power. Not gulags first. Media capture first. The police state is the 10% that shows up after the narrative’s already won.

Her prescription for ordinary people is deliberately unglamorous: a “Swiss cheese” approach to safety, where no single defense is complete but enough imperfect layers stacked together close most of the gaps — a password on your encrypted chat app, photos moved out of your camera roll, a neighbor who notices when someone’s watching your house. And a hard warning against perfectionism: the man in front of the tank in Tiananmen Square didn’t stop anything. A kid in front of tanks in Nepal did. Nobody can tell you in advance which piece of sand jams the gears, so you throw as much sand as you can.

Noah gave her a phrase mid-conversation that made Julia audibly delighted: mathematical dissident. Someone who spent a career hiding behind the pretense of journalistic neutrality and finally decided to just state her axioms — pro-democracy, anti-authoritarian — and build everything else as a corollary. It’s a very mathematician thing to do: pick your starting assumptions honestly, out loud, and then let the logic go where it goes.

There isn’t, by Julia’s own admission, much actual math in the new book. But there’s a mathematician’s temperament running underneath all of it — the same one that made her walk out of college parties alone. Agree on the axioms first. Then see what you can build.