As an astrophysicist, I often find myself in awkward situations when I visit my hometown during Chinese New Year. Relatives love to bombard me with yes-or-no questions: "Do aliens exist?" "Have UFOs visited Earth?" "Is there an edge to the universe?"
I could take the easy way out and say "No, no, and no." But that would betray the spirit of science. Scientific questions rarely fit neatly into true-or-false boxes—especially in my specialty, astrostatistics. The honest answer is usually: "Probably not."
I know the word "statistics" might make you want to stop reading. Please stay. I promise this will be useful. Let me show you how statistical thinking applies to something very concrete: coronavirus testing.
I live in the United States. Suppose I test positive for COVID-19 here. You might think I'd feel reassured—after all, the US is a developed country with good healthcare. But statistically, I'd actually be more worried than if I tested positive in Malaysia. If I got a positive result in Malaysia, I'd stay calm and consider retesting. In the US? I'd be genuinely alarmed.
Why? To understand, you need to know a little about how probability actually works.
Context Changes Everything
At the time I'm writing this, over nine million Americans have tested positive. Unofficial estimates suggest one in five people in states like New Jersey may have already been infected. The virus is everywhere.
So if my test comes back positive, what's the probability I actually have COVID? Two factors matter. First: how accurate is the test? If it's 100% accurate, a positive result means I definitely have it. But no test is perfect. Even reliable tests produce some false positives.
The second factor is what statisticians call the "prior"—the probability I had the virus before I got tested, based on everything else I know about my situation.
Here's a simpler example. Suppose I test positive for a rare tropical disease that's only transmitted by a specific insect found in the Amazon rainforest. I've never been to South America. I've never been anywhere near that insect. Unless the test is literally perfect, my prior knowledge tells me this is almost certainly a false positive. I should get retested.
Now flip it around. If I test positive for COVID in the US during a massive outbreak, my prior is very different. The virus is rampant. People around me are getting sick. A positive result here probably means I have it—not because the test is more accurate, but because my circumstances make infection far more likely.
This is Bayesian reasoning: new evidence doesn't stand alone. It updates what you already knew.
The Real Goal of Science
Astrophysics—and science generally—works the same way. We're constantly updating our prior knowledge based on new observations. Sometimes the data matches our expectations. Sometimes it overturns them completely.
My parents have never quite understood what I do. They often ask: "What are you trying to prove with your research?" The truth is, I'm not trying to prove anything. I'm trying to find where our current understanding breaks down. Science advances by discovering inconsistencies, not by confirming what we already believe.
It's a bit like COVID testing. A negative result is reassuring but unexciting. A positive result is alarming—but also far more informative. In science, "positive results" that contradict our models are where the real discoveries happen.
Does this mean anything is possible? That scientists just make things up? Obviously not. We believe the sun rises in the east because everyone observes the same thing, day after day. We believe in black holes and dark matter because multiple independent lines of evidence all point the same direction. When the data is overwhelming, your prior becomes irrelevant. If you test positive for COVID ten times in a row, you have COVID—regardless of how unlikely you thought it was beforehand.
But we should be careful when something contradicts common sense. Carl Sagan put it well: "Extraordinary claims require extraordinary evidence." A single surprising observation isn't enough to overturn well-established knowledge. If you test positive once in a country with low infection rates, don't panic—but do get tested again.
Embracing Uncertainty
I've never believed science exists to explain everything with certainty. This is why I struggle with my relatives' yes-or-no questions. When we demand black-and-white answers, we treat science as a set of fixed rules. That misses the point entirely.
Science, like the humanities and philosophy, is a lens for understanding the world. At its core lies a single principle: we can never be completely certain about anything, and we must let new observations revise our beliefs. This applies far beyond the laboratory.
Voltaire said it best: "Uncertainty is an uncomfortable position. But certainty is an absurd one."