Multi-target tracking involves a data association step to update the predicted object trajectories with the appropriate measurements. The data association problem is combinatorial in nature and thus has high computational requirements. The symmetric measurement equation (SME) approach transforms the combinatorial data association task into a nonlinear filtering problem, introducing other difficulties, most importantly non-Gaussian densities. This paper investigates the performances of two common SME approaches, the sum-of-products and the sum-of-powers formalisms. The extended Kalman filter implementations are compared to a matched Kalman filter, which serves as a baseline estimator.