Grubbs
Outlier test based on the assumption that the data is normally distributed.
The steps to follow are:
- Define the average of the first 20 Datapoints; ( Average( Reeks));
- Define the standard deviation of the 20 Datapoints; ( StDev.s( Reeks));
- Define per Datapoint the absolute difference between the datapoint and the average;
- Define G per Datapoint. G is the absolute difference from point 3 divided by the standard deviation from point 2;
- The Datapoint is an outlier when G exceeds the critical value of 3,001.
- Choose the outlier with the highest G in case there are multiple outliers.
