systematic error

Imagine a bathroom scale that always reads two kilograms high, or a clock that runs steadily slow. However many times you step on the scale or check the clock, the answer is wrong in the same direction by the same amount. Systematic error is a bias that repeats identically from measurement to measurement, so no amount of averaging can remove it.

Formally, systematic error is the component of uncertainty that shifts every measurement in a consistent, reproducible way: from a miscalibrated instrument, a neglected physical effect, a biased procedure, or a flawed model. Because it does not fluctuate randomly, it does not average down with more data; taking N measurements leaves the bias exactly where it was. You reduce it not by collecting more but by understanding it: better calibration, control experiments, cross-checks against an independent method, and modelling the effect so you can subtract it and assign a residual systematic uncertainty for whatever you could not fully pin down.

In precision physics the systematic budget usually dominates the final error bar and is the hardest, most judgement-laden part of the whole analysis. The famous cautionary tale is the 2011 OPERA claim of faster-than-light neutrinos, eventually traced to a loose fibre-optic connector and a clock miscalibration: pure systematics masquerading as a revolutionary discovery. Honest practice therefore always quotes the statistical and systematic uncertainties separately, because they behave so differently.