Confidence Intervals: An Estimate’s Wiggle Room

It sometimes seems as if a new headline promises to change the way we think about our health every few weeks:

A New Medication Cuts the Risk of Heart Disease

Coffee Might Help You Live Longer

A Popular Diet Reduces the Risk of Diabetes

Some findings eventually become accepted medical advice, while others fade away, and every so often, two studies seem to reach completely different conclusions.

So, how do scientists know when they’re confident enough to make a claim? Scientists rarely claim to be certain. Instead, they ask a different question: How confident should we be that this result reflects what’s really happening?

That distinction lies at the heart of modern science.

Every study is based on limited information. Researchers can’t observe every patient, test every treatment, or measure every outcome. Instead, they collect the best evidence they can, analyze it carefully, and estimate what is most likely true. They also measure how uncertain those estimates are.

When the Same Number Doesn’t Mean the Same Thing

Imagine two research teams test different medications for high blood pressure. Both studies report the same result: Patients’ blood pressure dropped by an average of 10 points. If that’s all you know, the studies seem equally convincing. But the researchers know there’s another question to ask: How precise is that estimate?

The first study reports the average reduction is likely between 9 and 11 points, while the second reports it could be anywhere between 2 and 18 points. Suddenly, those identical headlines tell divergent stories.

The first study gives scientists a much clearer picture of the medication’s effect. The second suggests the treatment might work, but the estimate is much less certain. Researchers might need to study more people or collect additional evidence before drawing strong conclusions.

That difference matters because science isn’t just about finding an answer. It’s about understanding how much confidence we should place in that answer.

One of the tools statisticians use to express that confidence is called a confidence interval. Rather than presenting a single number as if it were exact, a confidence interval describes a range of values that are consistent with the evidence collected.

In other words, instead of saying, “The answer is exactly 10,” scientists are saying, “Based on what we’ve observed, this is the range where the true answer is most likely to be.”

Why Confidence Matters

Most people read headlines, not scientific studies. Those headlines usually summarize a single finding; they rarely tell how much uncertainty lies behind that finding. That’s one reason scientific studies can seem confusing—or even contradictory.

Imagine one study reports a treatment reduces pain by an average of 15%, while a second study finds an average reduction of 12%. At first glance, it might look as though the researchers disagree. However, if both studies have relatively wide confidence intervals, the results could be more similar than they appear. The apparent disagreement could simply reflect the natural uncertainty that comes from studying different groups of people.

As more studies are conducted, those estimates often become more precise. Scientists begin to see not only whether a treatment works, but also how well it works and how confident they can be in that conclusion.

In other words, science doesn’t usually move forward because of one dramatic discovery. It moves forward as evidence accumulates and uncertainty gradually shrinks.

As the National Academies of Sciences, Engineering, and Medicine notes in Chapter 7 of its report Reproducibility and Replicability in Science, reporting uncertainty isn’t a limitation of science—it’s “a central tenet of the scientific process.” By showing not only what they found but also how confident they are in those findings, researchers give others the information they need to evaluate, replicate, and build on the work.

Confidence Isn’t Certainty

If you’ve ever heard a scientist say, “The evidence suggests …” or “The results are consistent with …,” you may have wondered, “Why won’t they just say yes or no?” The answer is because science rewards accuracy over certainty.

Researchers know every study has limits. Every experiment measures only part of a much larger picture. New evidence may strengthen today’s conclusions, refine them, or occasionally overturn them. That doesn’t mean scientists lack confidence in their work. It means they’re careful about matching their confidence to the strength of the evidence.

Rather than claiming to have all the answers, scientists describe what the evidence supports today while remaining open to what tomorrow’s evidence might reveal. It’s a mindset built on curiosity, transparency, and continuous learning.

The Takeaway

The next time you read about a breakthrough treatment, a new nutrition study, or the latest poll, look beyond the headline. Ask not only about what the researchers found, but also about how confident they are in the result.

Science rarely offers absolute certainty. Instead, it offers the best answers we have today, along with an honest assessment of how much confidence we should place in them.