We need evidence because human judgment alone is unreliable. Our brains are wired with shortcuts and biases that lead us to wrong conclusions, and evidence is the tool that catches those errors before they cause harm. Whether in medicine, science, public policy, or everyday decisions, evidence provides a way to test what we believe against what is actually true.
Human Intuition Gets It Wrong
Four decades of research into cognitive biases has shown that human reasoning is prone to systematic irrational patterns, especially when the subject matter is complex, novel, or ideologically charged. We engage in “motivated reasoning,” selectively searching for arguments that support what we already believe while ignoring or undervaluing evidence that contradicts us. This tendency is so deeply embedded that we rarely notice it happening.
Confirmation bias is the most familiar example: we naturally seek out information that supports our existing opinions and dismiss information that doesn’t. But it’s far from the only one. Status quo bias makes us prefer the current state of affairs simply because it’s familiar. Loss aversion makes us weigh potential losses more heavily than equivalent gains. Omission bias makes us judge harmful actions as worse than equally harmful inaction. These aren’t occasional glitches. They are the default operating mode of the human mind, and they affect everyone from voters to doctors to scientists themselves.
Evidence acts as a counterweight. By requiring that claims survive testing and scrutiny, it forces us to confront the possibility that our instincts are wrong. Science, as the philosopher Karl Popper argued, progresses by making precise claims, subjecting them to rigorous tests, and discarding the ones that fail. The power of science isn’t its tight logical structure (the process can be quite messy), but that it aims to test claims and seriously consider contradicting evidence.
Medicine Without Evidence Was Dangerous
The strongest case for why we need evidence comes from looking at what happens without it. In medicine, the number of practices later shown to be ineffective or actively harmful is staggering. Empirical reviews of high-impact medical journals have identified over 140 reversed medical practices, treatments that doctors used confidently for years before evidence proved them wrong.
One example: for decades, patients with coronary artery disease and stable chest pain routinely received stents to open blocked arteries. It seemed logical. The artery is blocked, so open it up. But a major 2007 trial found that for these patients, stenting was no better than medication alone. Thousands of people had undergone an invasive procedure with real risks for no measurable benefit. In another case, the FDA granted approval for an intracranial artery stent in 2005 based on a single, uncontrolled study that showed the device could widen narrowed arteries. It sounded promising, but that study couldn’t tell anyone whether patients actually lived longer or felt better. When proper trials were finally done, the device didn’t hold up.
These aren’t ancient mistakes from centuries ago. They happened in modern hospitals, performed by well-trained physicians who genuinely believed they were helping. Without rigorous evidence, good intentions and logical-sounding reasoning led to real harm.
How Evidence Strength Is Measured
Not all evidence is equal. Researchers rank different types of evidence by how likely they are to contain bias, the systematic errors that distort results.
At the top of this hierarchy sit systematic reviews of randomized controlled trials. These pool data from multiple studies where participants were randomly assigned to receive either a treatment or a comparison, which helps cancel out the confounding factors that can skew results. Below that come individual randomized trials, then well-designed observational studies that follow groups of people over time, and then case-control studies that look backward from outcomes to possible causes.
At the bottom sits expert opinion. This isn’t because experts are unintelligent. It’s because an individual expert’s conclusions are filtered through their personal experience, shaped by the particular patients they’ve seen or the specific outcomes they remember most vividly. There’s no control for confounding factors, and no mechanism to catch the biases that affect all human reasoning. A surgeon who has performed a procedure hundreds of times may genuinely believe it works, because they remember the successes more clearly than the failures, or because they never followed up with patients who didn’t improve.
Evidence Protects You From Unsafe Products
The drug approval process is built entirely on the principle that good intentions and theoretical promise aren’t enough. Before any medication reaches your pharmacy, it goes through a structured sequence of testing. First, laboratory and animal studies determine whether the drug works as expected and appears safe enough to try in people. Then a series of human trials begins, each designed to answer specific questions about safety, dosing, and real health benefits.
The U.S. Food and Drug Administration generally expects results from two well-designed clinical trials before approving a drug, specifically to make sure the findings from the first trial aren’t the result of chance or bias. In rare diseases where multiple trials aren’t feasible, convincing evidence from one trial may suffice, but the bar remains high. Approval means that the drug’s benefits outweigh its known and potential risks for the people who will take it.
This system exists because drugs that seem effective in small, uncontrolled studies regularly fail in larger, more rigorous ones. A promising lab result is just that: a promise. Evidence is the process of checking whether the promise holds up in the messy reality of human bodies and real-world conditions.
Evidence Keeps Science Honest
The standard tool for evaluating evidence in research is the p-value, a statistical measure of how likely a result would be if there were actually no real effect. The conventional threshold is set at less than 5%, meaning there’s less than a 1-in-20 chance the finding is a fluke. For fields where false positives carry serious consequences, the bar is set far higher. In genetics research, for instance, the threshold can be as strict as 1 in 100 million.
Even with these safeguards, the system isn’t perfect. A 2016 survey published in Nature found that more than 70% of researchers had tried and failed to reproduce other scientists’ experiments, and more than half couldn’t reproduce their own. This “reproducibility crisis” isn’t an argument against evidence. It’s an argument for more and better evidence, for replication, transparency, and higher standards. When a finding can’t be reproduced, the system flags it. Without the expectation of evidence, there would be no way to catch these failures at all.
Better Decisions in Policy and Practice
Evidence doesn’t just matter in laboratories. In healthcare systems, clinical decision support tools now integrate real-time patient data with evidence-based guidelines to help doctors manage the sheer complexity of modern medicine. These systems reduce cognitive overload, catch potential errors, and provide recommendations tailored to individual patients. They work because they’re built on accumulated evidence from thousands of studies and millions of patients, not on any single person’s memory or instinct.
In public policy, evidence-based approaches have produced measurable results. Evaluation of Medicare’s innovation models and shared savings programs, for example, found that evidence-guided redesigns generated real cost savings while improving quality of care. The alternative, making policy based on ideology, anecdote, or political convenience, has a long track record of wasting resources and failing the people it was meant to help.
The core principle is the same whether you’re a scientist testing a hypothesis, a doctor choosing a treatment, or a government designing a program. What you think is true might not be. What worked once might not work again. What seems obvious might be completely wrong. Evidence is the discipline of checking, and checking again, before acting on assumptions that could affect people’s lives.