A spreadsheet full of measurements can contain one number that doesn’t look like the rest, but no statistics software is required to check it properly. This guide walks through running a Grubbs test entirely inside Excel: setting up the data correctly, calculating the G statistic and critical value with built-in functions, and reading the comparison so you finish with a documented, defensible decision about that suspicious value.
A Grubbs test on Excel checks whether the single most extreme value in a numeric dataset is a statistically significant outlier. You calculate the sample mean and sample standard deviation with AVERAGE and STDEV.S, find the largest absolute distance between the extreme observation and the mean, and divide that distance by the standard deviation to get the G statistic. Separately, you calculate a critical value from the sample size and a chosen significance level using T.INV.2T. If the G statistic is greater than the critical value, the test flags the extreme value as an outlier at that significance level; if not, the value stays in the dataset.
Grubbs test on Excel workflow from data to outlier decision
The Grubbs test, also called the Grubbs’ outlier test or the extreme studentized deviate test, is a statistical hypothesis test built for deciding whether the single most extreme value in a univariate, approximately normal dataset is inconsistent with the rest of the sample. For broader context on this type of analysis, see the Grubbs outlier test Excel methods overview.
The test compares two numbers you build in Excel: the Grubbs statistic (G) and a critical value (Gcritical) that depends on your sample size and significance level. When G exceeds Gcritical, you reject the null hypothesis that no outliers are present at the selected significance level.
Prepare Excel data before running the Grubbs test
Before any formula goes into a cell, the data itself needs to satisfy a few conditions. The Grubbs test assumes the underlying variable is approximately normally distributed and is designed for one extreme value at a time.
Calculate the Grubbs statistic with Excel formulas
Four numbers feed the Grubbs statistic: sample size, mean, standard deviation, and the largest absolute deviation. Excel functions such as COUNT, AVERAGE, STDEV.S, ABS, MAX, MIN, and T.INV.2T can build the calculation without add-ins.
Interpret Grubbs test results and verify the outlier
A significant Grubbs result is a statement about probability under the test assumptions, not automatic proof that a value is wrong. Review the flagged observation before removing it.
Troubleshoot incorrect Grubbs test calculations in Excel
Most incorrect results trace back to mismatched ranges, wrong standard deviation functions, incorrect alpha values, or invalid assumptions about the dataset.
Choose Grubbs test versus other Excel outlier methods
The Grubbs test is useful for one suspected extreme value in approximately normal data. Other approaches such as z-score screening and IQR methods may be better for broader outlier detection tasks.