Most Science Papers Don’t Replicate Here is Why That Should Scare You

On Saturday, 24 May, 2025, J.D. Vance (VP) remarked with sharp clarity that “there is an extraordinary reproducibility crisis in the sciences, particularly in biology, where most published papers fail to replicate.” He went further to argue that universities are plagued by bureaucracy that blocks innovation, dominated by politically homogenous faculty, and even guilty of discriminatory practices that may violate civil rights. His point wasn’t just criticism for the sake of it it was a call to reflect on how institutions respond to failure: will they reform themselves or resist accountability under the guise of victimhood?

This observation sets the stage for a deeper exploration of why trust in both science and the humanities is faltering and what that really means for how we understand and build knowledge today.

Is Science Failing to Repeat Itself?

Let’s begin with something you probably didn’t expect to hear today: a lot of published science doesn’t hold up. And I know what you might be thinking “Really? I thought science was supposed to be solid. Facts. Evidence. Truth.” That’s what we’re all told. But here’s the uncomfortable truth: much of what we accept as scientific fact hasn’t actually been tested again. And when it has, it often fails to pass the test. In 2016, a Nature survey of over 1,500 scientists showed that more than 70% couldn’t reproduce another researcher’s experiment. Even more surprisingly, over 50% failed to replicate their own studies. That’s not just a glitch. That’s a crack in the foundation. And it raises a big question: what is going wrong in science and why aren’t we fixing it?

What Do We Mean by Reproducibility, and Why Should You Care?

Source- Nature

Reproducibility simply means that if I follow your method and use your data, I should get similar results. It’s the bare minimum standard of “truth” in the sciences. Without this, we’re not really building knowledge. We’re stacking guesses on top of guesses. Now, in cancer biology where the stakes couldn’t be higher Amgen scientists tried to reproduce 53 landmark studies. They succeeded with only six. That’s a success rate of around 11%. In psychology, one of the most active fields in replication reform, only 36 out of 100 major studies could be replicated. In ecology, a field that studies living systems, only 0.023% of studies even identify as replications. That’s not a typo it’s two out of every 10,000.

Source — Nature

So yes, this is real. This is serious. And while most people don’t know the numbers, they often feel the symptoms: conflicting health advice, changing environmental models, or new studies that seem to “reverse” the old ones. It creates confusion And confusion undermines trust.

Why Is This Happening Across So Many Fields?

Let’s break this down, not with blame, but with clarity.

First, we have publication bias. Journals prefer to publish new, surprising, positive findings. Nobody wants to read (or fund) a study that says “we didn’t find anything.” So researchers often tweak their data or cherry-pick results to make them look exciting. You can almost hear the inner voice: “If I don’t find something big, I won’t get published, I won’t get promoted, I won’t get my next grant.” Second, statistical power is often low. Many studies use small sample sizes or flawed analysis. They find “statistically significant” effects that are actually just noise. But the paper gets published anyway, and then someone else tries to replicate it and… nothing. Just silence where the result should be. Third, incomplete reporting is a huge issue. Studies don’t include enough detail for others to replicate the work. Methods are vague. Data is hidden. It’s like giving someone a recipe that says “cook until it looks right.” What temperature? For how long? What’s the sauce made of? No one knows.

And lastly, there’s the incentive problem. Academia rewards novelty, not reliability. Publishing something new makes you look brilliant. Confirming an old study? Not impressive. So even when scientists want to do careful, solid replication work they’re discouraged from doing so.

Is This Crisis the Same in the Humanities?

Now you might ask, “What about fields like history, philosophy, or literature? Do they have the same problem?” Not exactly. The humanities don’t usually test hypotheses in the lab. But they do rely on interpretation, and that brings its own set of issues. In the humanities, the challenge isn’t “reproducibility” in the strict sense. It’s credibility. It’s about whether arguments are grounded, whether sources are used responsibly, and whether different views are being considered. Here too, we face citation bias, theoretical echo chambers, and sometimes, a lack of clear method.

Some essays quote only authors who agree with the thesis. Some schools of thought dominate departments so strongly that dissenting views get ignored. Some interpretations are so convoluted, they serve more to impress than to clarify. That’s not healthy scholarship. That’s intellectual performance. So while the reproducibility crisis is more urgent in the sciences, the humanities aren’t immune to similar credibility issues. And the result in both cases? Erosion of trust, and the growing sense among students, policymakers, and the public that academic knowledge is shaky at best.

What Do the Philosophers Say? Are We Asking the Right Questions?

Karl Popper had a clear rule: if your theory can’t be tested…and possibly proven wrong it’s not science. Falsifiability was his gold standard. But today, a lot of studies are designed in ways that make them practically impossible to disprove. That’s a problem. Then came Thomas Kuhn, who said that science moves through paradigm shifts. It’s not linear. It’s messy. Normal science fails, and then a revolution changes the rules. Maybe this crisis isn’t the end of science. Maybe it’s a Kuhnian shake-up.

But then there’s Paul Feyerabend. He said, basically: science has never followed strict rules, and that’s okay. Progress happens through chaos, rebellion, even contradiction. That sounds liberating …right! but also dangerous if misused. Do we really want to say “anything goes” when it comes to evidence? These thinkers push us to ask: are we expecting too much from science or not enough?

What Happens If We Don’t Fix This?

This isn’t just a debate for academics. If we can’t trust a cancer drug trial, people die. If climate predictions can’t be tested or explained, public action stalls. If university research is seen as biased or opaque, society loses its ability to think critically. Worse, it gives space to conspiracy theories, fake news, and anti-intellectualism. People start to think, “Well, if experts don’t agree, I’ll just believe whoever I like.” That’s already happening. And if trust breaks down further, we’re in trouble.

Everyone Talks About Solutions So Why Aren’t We Fixing It?

Now here’s the twist. We actually know what many of the solutions are. We just don’t act on them. Why? We know we need to fund replication studies. We know journals should publish null results. We know we should reward transparency, not just novelty. Yet very little changes. Why is that? Maybe it’s because acting on these solutions means changing the way academic prestige works. Maybe it means slowing down. Publishing less. Being humble. And institutions like people don’t change easily when it means losing status or power.

This is the part that frustrates me the most. And maybe it frustrates you too. Because it’s not just a technical issue it’s a cultural one. A leadership one. If more scientists, educators, and research bodies said: “We will prioritize truth over trend,” maybe we’d see change. But here’s the question that haunts me: Why aren’t we asking these questions more loudly? Why are professional associations and academic conferences so often just about presenting polished slides, rather than opening real conversations about failure, reform, and renewal?

Why aren’t we as scholars, students, or leaders demanding something better?

What Would a Better Academic World Look Like?

Let’s imagine it. What if replication was respected, even rewarded? What if humanities departments welcomed conflicting interpretations without punishment? What if journals published fewer, better studies? What if conferences were places for real dialogue, not performance? That wouldn’t make everything perfect. But it would make knowledge stronger. It would make learning more honest. And it might rebuild the trust we so desperately need in this moment.

It’s Not About Blame — It’s About Will

Let me be clear: I’m not blaming individual researchers. Most are doing their best in a broken system. This is about structures, cultures, and incentives that make good people play a flawed game. But we do have agency. And courage. And maybe, a little wisdom. If we don’t use it, we’ll keep building knowledge on foundations we can’t verify. That’s not just bad science it’s risky for all of us….

So the next time you hear about a “groundbreaking study,” ask…Has anyone tried to repeat it? Can we see the data? Is it building on solid ground?Because asking that question might be the most scientific and the most human thing we can do.