A theoretical analysis of the stimulated gradient-echo method for achieving T2*-preparation in magnetic resonance imaging is performed. The results show that T2*-weighting without the "shading effect" due to magnetic field variation can only be achieved if the dephasing requirement of the stimulated echo is satisfied. It is also shown that the effective echo-time for T2*-weighting depends on the polarities of the dephasing and the rephasing gradient pulses that generate the stimulated echo. The results from the theoretical analysis are experimentally validated.