An outlier is a point inconsistent with the bulk of the data, judged by its distance from the mean in standard deviations, by a test such as Grubbs, or by its residual from a fitted model. Before removing one, the analyst should find a reason: a measurement fault, an entry error, a unit from a different population. Deleting points just because they are inconvenient biases results, and some outliers are the most informative data in the set.
Wafer 9 is an outlier on every response. Check the log, it might have gone through right after a chamber door open.
Run the analysis with and without the point. If the conclusion flips, the outlier is the story.
Also heard as
- wild point
- anomaly
Related terms
Flyer
Single wild data point far from the rest, often a measurement or entry error rather than a real process change.
Residual
Difference between an observed value and the value a fitted model predicts for it.
Leverage
Measure of how far a data point's input values sit from the rest, which gives it outsized pull on a fitted model.
Fat-finger
Data entry error from a mistyped number, such as a slipped decimal or swapped digits, that shows up as a false signal.