A poorly designed experiment usually ends up with confounding variables. It is well recognized that confounding variables have a variety of unignorable impacts on interpretation of study results. Particularly, the impact on efficiency of testing primary endpoints is a big concern. Simulation studies are used to investigate the underlying contributing factors related to the impacts concerned such as Type I error rate and power. Sample size and correlations between confounding variables were identified as the two dominant contributing factors. In addition, whether the confounding variables are random or fixed effects also affects Type I error and power.