Detection Method
Mathematical procedures identify a single outlier within a data set that assumes a normal distribution. Using the grubbs test allows a laboratory to determine if a specific fibre strength reading or a chemical concentration value is far enough from the mean to be discarded as a measurement error. This method calculates a G value by taking the absolute difference between the suspected outlier and the sample mean, then dividing it by the sample standard deviation.
Application Boundary
The calculation requires at least three independent measurements to function correctly but is most reliable when applied to larger batches of data. If the calculated G value exceeds the critical value for the chosen significance level, the point is classified as an outlier. Only one value can be tested at a time, so the grubbs test must be repeated sequentially if multiple suspect points exist.
Industrial Context
Production mills use this statistical tool to filter out anomalous results caused by temporary machine malfunctions or operator error during quality control checks. It prevents a single extreme reading from skewing the average moisture content or the reported yarn count of a large production lot.
Data Integrity
Indiscriminate use of the test to remove valid but undesirable data points compromises the accuracy of the quality report.
