A comparison of impulse approximation calculations for the (e,ep) reaction, based on the Dirac equation and the Schrodinger one is presented. Trivial (kinematics) differences are indicated, as well as how to remove them from the standard nonrelativistic formalism. Signatures of the relativistic approach are found where the enhancement of the lower components (spinor distortion or negative energy contributions) modifies TL observables with respect to the nonrelativistic predictions, what seems to be confirmed by the experiment. Finally, the relativistic approach is used to analyze several experiments for the reaction 16O(e,ep)15N taken at values of Q^2 from 0.2 to 0.8 (GeV/c)^2, not finding a significant Q^2 dependence of the scale factors over this range.