Science Isn't a Set of Facts — And Teaching It That Way Is Backfiring

For years, psychology textbooks featured a concept called ego depletion—the idea that willpower is a limited resource that gets used up over the course of a day. It was cited everywhere. Self-help books were built on it. Teachers referenced it. It felt like solid, established science.

Then researchers tried to replicate the original studies. Across dozens of attempts, many couldn't reproduce the effect reliably. The story got a lot more complicated, and the confident version that had spread through popular culture turned out to be, at best, an oversimplification.

Ego depletion isn't an outlier. It's part of a broader pattern that has come to be called the replication crisis—a reckoning with the fact that a significant number of published findings in psychology, medicine, nutrition, and other fields don't hold up when other researchers try to reproduce them. And while this crisis is often framed as a problem within science, it's also a problem with how we've been talking about science to the public.

What "Crisis" Actually Means Here

The word crisis makes it sound like science is broken. It isn't. What's actually happening is closer to a correction—a field catching up with its own standards, identifying where incentive structures, small sample sizes, and publication biases led to overconfident conclusions.

But here's the thing: the process of catching and correcting errors is exactly how science is supposed to work. The problem isn't that studies got revised. The problem is that nobody told the public that revision was normal.

For most Americans, the version of science they encountered in school went something like this: scientists do experiments, discover facts, and those facts accumulate into an ever-growing body of settled knowledge. Uncertainty was a gap waiting to be filled. Disagreement among experts was a sign that someone hadn't done the work yet.

That version of science is tidy, but it's not accurate—and it creates a fragile kind of trust. When people who were taught that science produces certainties discover that a widely-cited study didn't replicate, or that dietary guidelines shifted again, or that a drug approved by the FDA turned out to have serious side effects nobody caught in trials, the response isn't usually nuanced. It's often: "So they were lying? Why should I trust any of it?"

That's the backfire. And it's a direct consequence of how we set up the relationship between science and trust in the first place.

The Certainty Trap in Science Education

There are understandable reasons we've taught science this way. Uncertainty is hard to communicate without sounding like you don't know what you're doing. In a competitive funding environment, researchers have real incentives to present their findings with confidence. Teachers working through packed curricula don't always have time to explain why the textbook's version of a study is a simplified snapshot of an ongoing conversation.

But the cost of that simplification is steep. When people learn that science produces The Truth rather than our best current understanding based on available evidence, they develop a binary relationship with it. Either science is completely trustworthy or it isn't trustworthy at all. There's no framework for engaging with uncertainty as a normal feature of the process.

This makes people especially vulnerable to bad-faith actors who exploit scientific uncertainty for their own ends. Climate denial, anti-vaccine messaging, and supplement industry marketing all use the same basic move: find a study that didn't replicate, a researcher who changed their position, or a genuine area of scientific debate—and use it as evidence that nothing can be trusted. That argument only works on an audience that was never taught that uncertainty and revision are features, not bugs.

What Honest Science Communication Actually Looks Like

Rebuilding trust doesn't mean lowering confidence in science. It means building a more accurate model of what science is.

That starts with language. Phrases like "the science is settled" might be politically useful in certain debates, but they're epistemologically sloppy. Even in areas of strong scientific consensus—climate change, vaccine safety, evolution—the consensus is built on ongoing research, not a permanent verdict. Saying so doesn't undermine the consensus. It actually explains why the consensus is trustworthy: because it's been tested repeatedly and held up.

It also means being honest about the incentive structures that contributed to the replication crisis. Scientists are human. They work under pressure to publish, to secure funding, to achieve results that confirm their hypotheses. Pre-registration of studies, open data requirements, and replication initiatives are all responses to those pressures—and they're worth explaining to the public, not hiding from them.

When a study doesn't replicate, that's worth communicating too—not as a scandal, but as the system working. "We thought X, but when we looked harder, the evidence was weaker than we believed" is an honest sentence. It's also a more trustworthy one than pretending the original finding was never that confident.

Teaching Uncertainty Without Teaching Nihilism

There's a version of this conversation that goes badly: you explain that science is uncertain and revisable, and someone concludes that therefore all claims are equally valid, and therefore they should trust their gut over the CDC. That's not the lesson.

The lesson is that degrees of confidence matter, and that evidence-based reasoning—even when it produces provisional conclusions—is still vastly more reliable than anecdote, ideology, or motivated reasoning. The fact that some nutrition studies haven't replicated doesn't mean diet is unknowable. It means we should hold individual studies lightly and pay more attention to systematic reviews and meta-analyses.

This is actually a more sophisticated and more accurate picture of how knowledge works. And it's one that, once people genuinely understand it, tends to produce more durable trust—because it doesn't collapse the moment a headline announces that something scientists said was wrong.

At SkeptiCamp, we've seen this play out in workshops with adults who came in deeply skeptical of science after feeling burned by changing recommendations. What often shifts isn't the evidence—it's the frame. When they understand that "we updated our understanding" is a sign of integrity rather than incompetence, something loosens. That's not a small thing. That's exactly the kind of trust that's worth rebuilding.