Mediation is said to occur when a causal effect of some variables X on an outcome Y is explained by some intervening variables
M. Simple mediation model involves a series of regression equations. The Ordinary Least Squares (OLS) is the most popular technique to estimate the parameters of the model. However, this technique is easily affected by an outlying
observation. In order to rectify this problem, we may turn to robust methods which are not sensitive to any deviations from some ideal assumptions. In this paper, we compare the OLS and MM parameter estimation methods on simple mediation analysis. We do screening steps from the data to make sure that the data clean enough. Then we contaminate the clean data with different outlier scenarios and then examine their impact on the mediation estimates. The results from the
numerical examples indicate that the performance of the MM-estimator is more efficient than the OLS estimator in x, m and y-direction. A numerical example is created using simulated data set with the Proc Robustreg of SAS version 9.13.