three methods to compute the factors main effect on model responses are compared. The first one is the improved Fourier amplitude sensitivity test [Tarantola, S., Gatelli, D. and Mara, T.A., 2006, Random balance designs for the estimation of first-order global sensitivity indices. Reliability Engineering and System Safety, 91(6), 717–727.]. The second one is the extension of the previous technique to the method of Sobol. At last, an original approach that combines a sampling-based method (Monte Carlo) with iterated one-dimensional fittings is also investigated. In our works, we show that the three methods are able to estimate the factors main effect. On the one hand, the improved FAST and the proposed sampling-based strategy are the less expensive methods as they only require one single sample set of simulation runs. On the other hand, when factors are correlated or an uncertainty analysis is also investigated, the former is the most suited. Besides, it is also demonstrated that the estimates of the main effects with the proposed sampling-based strategy is more robust and is computationally efficient.