Pioneering work by Nicolas Ambraseys and many collaborators demonstrates both the tremendous value of macroseismic data and the perils of its uncritical assessment. In numerous publications he shows that neglect of original sources and/or failure to appreciate the context of historical accounts, as well as use of unreliable indicators such as landsliding to determine intensities, commonly leads to inflated magnitude estimates for historical earthquakes. The U.S. Geological Survey “Did You Feel It?” (DYFI) system, which now collects and systematically interprets thousands of first-hand reports from felt earthquakes, provides the opportunity to explore further the biases associated with traditional intensity distributions determined from written (media or archival) accounts. I briefly summarize and further develop the results of Hough (2013), who shows that traditional intensity distributions imply more dramatic damage patterns than are documented by more spatially rich DYFI data, even when intensities are assigned according to the conservative practices established by Ambraseys’ work. I further consider the separate intensity–attenuation relations that have been developed to characterize intensities for historical and modern earthquakes in California, using traditionally assigned intensities and DYFI intensities, respectively. The results support the conclusion that traditionally assigned intensity values tend to be inflated by a fundamental bias towards reporting of dramatic rather than representative effects. I introduce an empirical correction-factor approach to correct for these biases. This allows the growing wealth of well-calibrated DYFI data to be used as calibration events in the analysis of historical earthquakes.