gross error
/ grohss EHR-ur /
Most measurement errors are small, polite things. A gross error is the opposite: it is the moment you read the wrong line of a burette, knock over a flask, write down 5.2 instead of 2.5, or use the wrong reagent entirely. It is a mistake, a blunder — not a subtle limitation of the equipment, but something that simply went wrong, often in one obvious lurch.
A gross error is a large, one-off mistake caused by carelessness, malfunction, or misjudgement, rather than by the inherent random scatter or systematic bias of a method. Unlike random and systematic errors, which are properties of the measurement process, a gross error is an avoidable accident that usually affects just one result, throwing it wildly far from where its companions sit. The data point it produces is often called an outlier.
Gross error matters because a single blunder can quietly poison an average and ruin a conclusion if it slips through unnoticed. The correct response is not to keep it and average it away, but to recognise it for what it is: a result spoiled by an identifiable mistake should be discarded and the measurement repeated. The honest caveat is discipline — you may only reject a value if you have a genuine reason (a documented blunder) or a statistical outlier test justifies it, never simply because a number is inconvenient.
Five titration volumes read 24.32, 24.30, 24.31, 19.85, 24.29 mL. The 19.85 is wildly out of line — most likely the analyst misread the burette or used a half-empty pipette. That single gross error should be investigated and, if confirmed, dropped before averaging.
A blunder stands out — investigate before you delete.
A gross error often shows up as a statistical outlier, but the two ideas differ: 'outlier' describes how the number looks in the data, 'gross error' names the human or instrument blunder behind it.