Robustness

An inference procedure is robust if it still produces roughly accurate results when the conditions for its use are not met. Procedures can have different degrees of robustness against different influences. Robustness can be influenced by a number of factors, e.g. the number of observations in the samples used for the procedures.