How Much Trust Can You Put in a Number?

In science, no number stands entirely alone. Every measurement you make—whether it’s the temperature of a liquid, the time of a fall, or the brightness of a star—carries with it a certain level of uncertainty. Learning to recognize and account for that uncertainty is at the heart of good experimental thinking.

But this doesn’t mean that you have to become a statistician. In fact, much of the time, it’s enough to ask a few simple but powerful questions:

If I repeat this measurement, how much do the values vary?
Is the change I’m observing real, or could it be noise?
Are there patterns in the error, or is it random?

Understanding why a number might change helps us see what’s going on beneath the surface. For instance, a thermometer might give slightly different readings each time you use it—but if it consistently reads 2 °C too high, that’s a systematic error. If it fluctuates up and down with no clear reason, that’s random error.

You need to understand:

Why taking multiple measurements is almost always better than one.
The difference between accuracy (how close your measurement is to the truth) and precision (how consistent your measurements are).
How simple tools—like averaging, error bars, and standard deviation—help us see the reliability of our data.

And of course, we’ll show you how to let ChatGPT help:

“I measured the fall time five times: 2.1 s, 2.3 s, 2.2 s, 2.2 s, 2.4 s. What’s the average, and how much do they vary?”

Understanding uncertainty doesn’t make your data weaker—it makes your conclusions stronger. It teaches you not only how to measure, but also how to think.