Disguised Alpha Science
All articles

How science works

Why good measurements include uncertainty

A reported value tells you more when you know how precisely it was measured.

· 2 min read

First, take the reading again

On a kitchen scale, the reading for the same object might shift slightly between attempts. The spread of repeated readings gives you one way to judge how consistent the measurement is under those conditions.

A display can show the same number every time and still have limits. Its smallest increment, its calibration, and how you position the object on the scale can all affect what you can conclude from the result.

What the number beside the result means

Suppose a lab reports a length of 20.40 mm with an expanded uncertainty of 0.06 mm. The stated interval runs from 20.34 to 20.46 mm. The method and coverage information explain how the lab arrived at that interval and how to interpret it.

NIST distinguishes uncertainty evaluated statistically from repeated observations from uncertainty evaluated by other means. Calibration certificates and instrument specifications can supply information for the latter.

Use uncertainty when comparing results

Imagine two instruments display slightly different lengths for the same object. Before concluding that one instrument disagrees with the other, check their calibration and reported uncertainties. The difference might be small relative to what the instruments can reliably distinguish.

A careful report gives you enough context to make that judgment: what was measured, how it was measured, and how much uncertainty accompanies the result.

Sources

More to read