Heki (2011) and Heki and Enomoto (2013) claimed that anomalous, yet similar, increases of ionospheric total electron content (TEC) started ~40 min prior to the 2011 Tohoku‐Oki, as well as before other Mw > 8 earthquakes. The authors concluded that the reported TEC anomalies were likely related to the pending earthquakes, suggesting also that TEC monitoring may be useful for future earthquake prediction. Here we carefully examine the findings of Heki (2011) and Heki and Enomoto (2013) by performing new analyses of the same TEC data. Our interpretation is that the 40 min onset of the ionospheric precursors is an artifact induced by the definition of the reference line adopted in analyzing TEC variations. We also discuss this repeatability in the tectonic and geodynamic context of the earthquakes. By performing a Superimposed Epoch Analysis of TEC data, we show that, however, the TEC increase reported by Heki (2011) was not particularly anomalous. We conclude that the TEC precursors reported by Heki (2011) and Heki and Enomoto (2013) are not useful for developing short‐term earthquake prediction capabilities. Key Points We examine TEC precursors starting about 40 min prior to strong EQs We report our TEC analysis attempting to reproduce previous findings We show that the TEC precursors are not associated with seismic activity