We provide a comprehensive analysis of special items and the characteristics of the firms that recognize them. Our analysis reveals that the temporal frequency, magnitude, and persistence of special items has increased significantly in the last 30 years, and that such increases are primarily driven by negative special items. More recently, however, our evidence is consistent with both a decline in frequency and magnitude of negative special items. On the other hand, we find that the frequency of reporting of positive special items, which remained relatively constant through 2002, has increased in more recent years. We also find strong evidence that subsequent special item reporting is an increasing function of the frequency of “prior” special item reporting. Using a random subsample of firms reporting special items, we document that 22 percent of the amounts reported in Compustat do not reconcile with the amounts reported on the firms' actual financial statements. Our comprehensive analysis should be of interest to regulators, academics, and managers interested in the implications of special items on firm-related consequences such as future earnings and firm value. Our examination can also serve as a catalyst for researchers interested in extending this important area of inquiry.