What is a Confidence Interval?

A confidence interval is the range around a study's result that shows how precise that result is. The headline number is only the best guess, and the interval shows the whole span of values the data support. A narrow range means the answer is pinned down; a wide one means it is not.

Researchers rarely measure everyone, so they study a sample and estimate from it. Every sample comes out a little different, so the estimate always carries uncertainty, and the interval is how that uncertainty gets reported. Say a trial reports recovery 2.1 days faster, with an interval of 0.4 to 3.8 days. The best guess is 2.1, but the data support anything inside that span. Width is the main takeaway: a narrow range means the effect was measured precisely, while a wide one means the study found something but cannot say how big it is. What sits inside the range matters just as much, because a range running from below zero to well above it means no effect at all is still consistent with the data. The 95% does not mean a 95% chance the truth sits in this particular range. It means that if the study were repeated many times, about 95% of the ranges produced would contain it.

Two forecasts for tomorrow's high temperature. One says 74 to 76 degrees, the other says 60 to 90, and both centre on 75. Only the first tells you whether to wear shorts or carry a jacket. The midpoint is the estimate, and the width is how much the forecaster actually knows.