XLSTAT: Grubbs test to detect outliers in Excel

 

Background

First, we state the null hypothesis that there are no outliers in the dataset. Here is how to calculate the Grubbs test statistic:

If this statistic is above the value of significance in the Grubbs table, we reject the null hypothesis and consider that the dataset contains outliers.

Dataset for Example

The data have been obtained from a normal distribution with mean 0 and variance 3. One outlier has been added. We wish to test if there is one outlier in the sample.

Setting Up

• Open XLSTAT.
• Click on the Testing for outliers menu and select the Grubbs test for outliers feature:

In the General tab:

• Select the data and the Grubbs test option (the Double Grubbs test can be used to detect two outliers).

In the Options tab:

• Select the two-sided alternative hypothesis, meaning testing if a relationship exists in both directions (positive and negative).
• Leave the default significance level at 5%, which corresponds to a 95% confidence threshold.

• Click on the OK button to launch the computations.
• The results are displayed on a new Excel sheet.

Interpreting the Results

The result is that the p-value for this test is smaller than 0.0001. That means that the null hypothesis should be rejected.

In the Z-scores table the outliers are displayed in bold.