Relationships are derived that describe the effect of electrode polarization on the mean ionic transference number of a solid electrolyte and allow to re-affirm that the "modified emf method" is not at all qualified to provide the least piece of useful information about the transference number. The false impression that this "method" is supposed to distinguish itself by the capability to refine distorted input data and to correct them by the impact of electrode polarization comes from a fundamental mistake in the practiced data evaluation as a consequence of disregarding the complexity of the underlying phenomena. The approach is an unparalleled example of an artefact which to date has not yet been recognized by the scientific community. (C) 2011 Elsevier Ltd. All rights reserved.