The case of the missing victims: Gunshot woundings in the National Crime Survey

Abstract

National Crime Survey (NCS) data yield an estimate that 171,000 Americans were nonfatally shot in criminal assaults, robberies, and rapes for the period 1973–1979. Comparing this estimate with the number of firearms homicides during this period suggests either that the death rate in gunshot cases is very high (over 1/3) or that the NCS estimate is low. Based on police-generated data appropriate to estimating the true death rate from gunshot wounds, it appears that the NCS estimate is low by a factor of about 3.0 compared with the number of criminal gunshot woundings known to the police. It is common knowledge that survey-based estimates of assault rates tend to be relatively unreliable, a fact that has been attributed to problems with respondents being willing and able to recall threats, fist fights, and other minor assaults. The current result indicates that the estimation problem is not limited to minor assaults. There is reason to think that the underestimate of gunshot woundings is the result of problems with the NCS sample as well as problems with respondent cooperation.