Miscounting suicides.

A review of research on the classification and counting of deaths reveals little hard evidence supporting claims that suicides are seriously underreported in the United States. The literature contains hints that much of what underreporting does occur is counterbalanced by "overcounting" (i.e., erroneous certifications of false suicides). An analysis of detailed cause-of-death mortality data for the United States indicates that the maximum likely undercount possible, using generous assumptions concerning misclassification frequency and leaving aside the issue of overcounts, was about 26% for 1980. The maximum likely net undercount, taking compensating overcounts into account, is estimated to be under 10%.