outlier
/ OWT-ly-er /
Picture a class where everyone scores between 70 and 80 on a test, except one student who scores 12. That lonely score is an outlier: a value that sits far away from the rest of the pack. In a set of measurements, an outlier is the reading that just does not seem to belong with the others.
An outlier is an observation that lies an unusually long way from the bulk of the data. It may arise from a genuine mistake — a misread burette, a spilled sample, a transcription slip — or it may be a real, surprising signal that the rest of the experiment missed. Statistical rules such as the Q-test help judge, by a fixed criterion, whether a value is far enough out to treat as suspect.
It matters because a single outlier can badly distort the mean and the standard deviation, so spotting it protects your conclusions. The honest caveat cuts both ways: deleting outliers just because they are inconvenient is data-tampering, yet keeping a value caused by an obvious blunder corrupts the result. The disciplined path is to investigate the cause, apply an agreed test, and document whatever you decide.
Six replicate absorbances read 0.452, 0.448, 0.451, 0.449, 0.450 and 0.612. The lone 0.612, traced to a fingerprint on the cuvette, is a clear outlier that should not be averaged in with the rest.
One stray reading, traced to a cause, stands apart from the rest.
An outlier is not the same as a gross error, though one often causes the other: a gross error is the blunder, the outlier is the odd-looking number it produces. The median is far less disturbed by outliers than the mean.