precision
/ prih-SIZH-un /
Back at the dartboard: if you throw five darts and they all land in a tight little cluster — even if that cluster is up in the corner, nowhere near the bullseye — that is precision. Precision asks not 'did you hit the target?' but 'do you keep landing in the same spot?' It is about agreement among your own repeated tries, not about being right.
Precision is the closeness of repeated measurements to one another. You quantify it by measuring the same thing several times and looking at how much the results spread, usually with the standard deviation or relative standard deviation: a small spread means high precision. Precision is degraded by random error — the small, unpredictable fluctuations that nudge each reading a little this way or that — and it has nothing to do with whether the average is correct.
Precision matters because it tells you how reproducible your measurement is, and a result you cannot reproduce is hard to trust. The essential caveat is the one chemists repeat constantly: precise is not the same as accurate. A miscalibrated balance can give beautifully consistent (precise) readings that are all wrong (inaccurate). Good precision is necessary for a dependable method, but it is never, by itself, proof that you are measuring the truth.
Six replicate titrations give 24.31, 24.29, 24.33, 24.30, 24.32, 24.31 mL. The values barely differ, so the precision is excellent — whether they are accurate depends entirely on whether 24.31 mL is the true volume.
Tightly clustered numbers show precision, not correctness.
Repeatability (same operator, same day, same instrument) and reproducibility (different labs or conditions) are two different flavours of precision — repeatability is the optimistic best case.