calibration
/ kal-ih-BRAY-shun /
Before you trust a new kitchen scale, you might place a known 100-gram weight on it and check that it reads 100. If it reads 103, now you know to subtract 3. Calibration is exactly this: checking an instrument against something whose value you already trust, and adjusting or correcting it so its readings can be believed.
More precisely, calibration is the procedure of comparing the readings of a measuring instrument against a reference standard of known value, and establishing the relationship between what the instrument shows and the true quantity. The result is either an adjustment of the instrument or a correction curve that translates raw readings into accurate values, traceable back through a chain of ever-more-precise standards.
Why it matters: an uncalibrated instrument can be perfectly precise — giving the same answer every time — and yet consistently wrong. Calibration is what links a lab's numbers to agreed standards so that results mean the same thing everywhere. A caveat: calibration drifts. Instruments age, conditions change, and a calibration done last year may no longer hold, so it must be repeated and dated.
To measure an unknown solution on a spectrometer, an analyst first runs several standards of known concentration and plots absorbance against concentration. That calibration line then converts the unknown's absorbance into its concentration — without it, the instrument's raw reading would just be a meaningless number.
Calibration turns an instrument's raw signal into a trustworthy, comparable value.
Calibration and adjustment are not the same step. Calibration only compares the instrument to a standard and finds the discrepancy; adjustment then changes the instrument to remove it. You can calibrate without adjusting — simply recording the correction to apply later — and a calibration says nothing useful unless the reference standard itself is sound.