The paper gives a new method to identify significant effects in two-level factorial experiments,and compares the new method with the exiting methods using Monte Carlo simulation.The existing methods only perform well under the assumption of effect sparsity,but the new method performs well without effect sparsity.