High spatial and temporal resolution magnetic tweezers experiments allow for the direct calibration of pulling forces applied to short biomolecules. In one class of experiments, a force is applied to a structured RNA or protein to induce an unfolding transition; when the force is maintained at particular values, the molecule can exhibit conformational switching between the folded and unfolded states or between intermediate states. Here, we analyze the degree to which common force calibration approaches, involving the fitting of model functions to the Allan variance or power spectral density of the bead trajectory, are biased by this conformational switching. We find significant effects in two limits: that of large molecular extension changes between the two states, in which alternative fitting functions must be used, and that of very fast switching kinetics, in which the force calibration cannot be recovered due to the slow diffusion time of the magnetic bead. We use simulations and high-resolution RNA hairpin data to show that most biophysical experiments do not occur in either of these limits.