It is well known that sample introduction via electrokinetic mode leads to a bias in conventional CE, which is proportional to the difference of electrophoretic mobilities between species. In electrophoretic separation chips using the gated injection mode, flow distribution at the crossjunction, which is linked to the electric field strength distribution during the loading step, induces an additional contribution to species discrimination. This second-order bias has a similar effect on quantitation like usual electrokinetic bias: the higher the analyte's apparent mobility, the larger the amount injected into the separation channel. The present paper assesses by numerical simulations the influence of several parameters, namely the injected amount, the electric field distribution, and the analyte-apparent Peclet number on this second-order bias.