If using the Grubbs’ method, start at Alpha = 0.01.Ĥ. It is recommended to start at Q = 1% if using the ROUT method. Placing the slider further to the right is more liberal and will remove more outliers.ĭeciding on where to place the slider is user dependent. Simply, placing the slider further to the left will apply a more aggressive approach and will remove less outliers. This is achieved by dragging the slider left or right. The next step when removing outliers in GraphPad is to determine how aggressive you want to be in the process.

How aggressive should you be in outlier removal? It is advised to avoid using this method altogether. It does so by simply repeating the Grubbs’ method using the cleaned data to detect a further outlier and repeats the process until no further outliers are present.

The Iterative Grubbs’ method is based on the prior namesake, but can detect more than one outlier. It is recommended to only use the Grubbs method when you are certain your data contains either one or no outliers. Alpha is similar to the Q value in the ROUT method. The Grubbs’, also known as the Extreme Studentized Deviate, method can only identify a single outlier.įor the Grubbs’ method you specify an alpha, rather than a Q, value. It is recommended to use the ROUT method when wanting to detect multiple outliers. Think of this value as being the maximum percentage of outliers that are false. To learn more about the ROUT method, have a read of the published manuscript.įor this method, Q is specified. This method was developed by GraphPad and is based on the False Discovery Rate (FDR). The ROUT method can be used to find any number of outliers. So, ensure to test your data for normality in GraphPad before proceeding. Note that these methods assume your data has been sampled from a Gaussian distribution. There are 3 methods available in GraphPad to identify and remove outliers: ROUT, Grubbs’ and Iterative Grubbs’.

What is graphpad prism used for