representative sample
A representative sample is a small portion taken from a much larger amount of material that genuinely reflects the whole — so that whatever you find in the spoonful is also true of the pot. Think of tasting a big pot of soup: if you stir it well and take one spoon, that spoon tells you how the whole pot tastes. But if you scoop only off the oily top, or only from the salty bottom, your spoonful lies about the rest.
Formally, a representative sample has the same composition, in the properties you care about, as the bulk material it was drawn from, within the precision the analysis needs. You almost never test the entire lot — a truckload of grain, a lake, a shipment of pills — so the few grams you actually measure must stand in for all of it. If they don't, no instrument, however good, can fix the answer.
This is the quiet foundation of every chemical result: the analysis can only be as truthful as the sample is representative. A perfect measurement on an unrepresentative sample gives a precise, confident, wrong number. That is why thinking about how the sample was obtained often matters more than the cleverness of the instrument that measures it.
To check the lead level in a 20-tonne shipment of wheat, an inspector draws scoops from many bags across the whole load and combines them — not one handful from a single bag near the door.
Many small portions from across the lot stand in for the whole.
Representativeness is judged against the property being measured, not in the abstract. A sample can be representative for moisture yet useless for a trace contaminant that sits only in a few hot spots — so the question is always "representative for what?"