method validation
/ METH-od val-ih-DAY-shun /
Before a restaurant lets a new dish onto the menu, the chef cooks it many times to be sure it always comes out right, tastes the same each evening, and works even when the regular oven is swapped for another. Method validation is the lab's version of that trial run: you prove, with evidence, that an analytical method really does what you claim before you start using it on real samples.
More precisely, method validation is the documented process of checking that a method performs well enough for its intended purpose. You measure a set of named characteristics — among them accuracy, precision, the limit of detection, the limit of quantitation, linearity, range, specificity and robustness — and show each one meets a pre-agreed acceptance criterion. The result is a validation report that says, in effect, this method is fit for this job.
It matters because an unvalidated method gives numbers nobody should trust: you would not know whether a low reading means a clean sample or a method that simply cannot see the analyte. The honest caveat is that validation proves fitness only for the conditions you tested — the same matrix, concentration range and instrument. Move the method to a very different sample or a new lab and at least part of the validation must be repeated.
A pharmaceutical lab validates a new HPLC assay for a drug tablet by showing accuracy of 98–102% recovery, a relative standard deviation under 2%, a straight calibration line from 50 to 150% of the target dose, and no interference from the tablet's fillers — then files the results as the validation report.
Proving a new drug-assay method is fit for purpose before routine use.
Do not confuse validation with verification: validation is the full proof done once for a new or modified method, while verification is the lighter check that a method already validated elsewhere still works in your hands.