Experimental studies have shown that the steric effect in chemical reactions can decrease (e.g., for Ba+N2O-->BaO*+N-2) Or increase [e.g., for Ca(D-1(2))+CH3F-->CaF*+CH3] with increasing translational energy. Decreasing (negative) energy dependences have successfully been modeled with the angle dependent line of centers model. We present a classical model in which a positive energy dependence of the steric effect is explained by an isotropic, attractive long range potential. In this "trapping" model we assume the reaction-apart from a cone of nonreaction at one side of the molecule-to be barrierless. This model shows that a positive energy dependence of the steric effect is not indicative of reorientation of the molecule, as has been suggested in the literature. Rather, the positive or negative energy dependence of the steric effect is shown to correlate with the absence or presence of a barrier to reaction and an attractive or repulsive long range potential. For the reorientation effects which occur in the case of anisotropic potentials, we consider the application of the standard quasiclassical trajectory (QCT) method and we introduce a modified QCT method. We argue that the latter is more suitable for the computation of the orientation dependent reactive cross section.