The Critical Differences between Replicates technique is used if there are too few replicates in a sample to use an outlier method. For instance, if only 2 replicates are tested it is not possible to statistically identify either of them as an outlier.
However, if a lab has an established history of conducting a specific analysis always having 2 replicates per sample then that history can be statistically analyzed to establish a mean and standard deviation of the differences between the replicates. Once established, the differences between replicates of new samples can be compared to this historical difference.
The Critical Differences template identifies any sample whose replicate difference is greater than the statistically established “critical” difference.
Example:
Precision Scientific Software 