We calculate corner frequencies and stress drops for 201 earthquakes in four earthquake sequences that are potentially induced by wastewater injection in Oklahoma. Specifically, we determine stress drops for 35 events in the 6 November 2011 5.6 Prague sequence, 40 events in the 13 February 2016 5.1 Fairview sequence, 73 events in the 3 September 2016 5.8 Pawnee sequence, and 53 events in the 7 November 2016 5.0 Cushing sequence. Although the stress‐drop estimates show large scatter for individual sequences, we find high stress drops for three of the four 5+ mainshocks (17–34 MPa, Brune stress drop) and lower stress drops for most of the foreshocks/aftershocks in each individual sequence. The exception is the 2011 Prague sequence, which has stress drops ranging between 0.03 and 76 MPa, and the mainshock has a low stress drop of 3.25 MPa. Compared with the other three sequences, the 2016 Fairview sequence exhibits more constant stress‐drop estimates, with a two to three times higher median stress drop. We find significant scatter in the stress‐drop estimates of small earthquakes () and note that the earthquakes greater than 4 have systematically larger stress drops than the average stress drop observed for earthquakes below 4. We observe no clear evidence for depth dependence or temporal patterns of stress‐drop estimates. The large spatial variability of stress drops reflects strong fault heterogeneity in this area, which is likely influenced by the injection of fluids into the subsurface. Given the large variations of stress drop, our results do not support the suggestion of using low stress drops in ground‐motion prediction models for seismic hazard assessment of induced earthquakes in the central and eastern United States (CEUS).