"The Validity of Official Crime Statistics: An Empirical Investigation," Social Science
Quarterly, 55 (June, 1974), 25-38.
Studies of crime, deviance, and social disorganization, and the impact of social conditions and
police decisions upon their scope, have been hampered by widespread suspicion of the
validity of official measures of crime. Crime statistics, the most notable and accessibly of which
are those contained in the FBI's yearly Uniform Crime Report, are reputedly invalid indicators
even of the limited actions which they purport to measure directly. Their validity is threatened
in the sense that the reported figures are not one-to-one reflections of events; the number of
burglaries "known to the police" in official parlance do not equal the number of burglaries
which have taken place. This report suggests that the data may still be useful. While error in
official estimates of the incidence of crime may present serious problems for those interested
in the development of social indicators, it may not be devastating for other enterprises. Official
crime statistics may be quite useful as such, if we make modest demands of the data. This
report compares official city crime statistics in two categories (robbery and auto theft) with
survey-generated measures of the incidence of victimization in the same communities. The
comparison suggests that official figures may be useful indicators of the relative distribution of
crime across cities.
“Comparing Measures of Crime: Police Statistics and Survey Estimates of Citizen
Victimization in American Cities. Proceedings of the American Statistical Association,
1974 Social Statistics Section.
This paper examines in detail the processes by which crimes are reported in survey interviews
and to the police. It compares the error structures of the resulting data.
"Crime and Crime Rates," in Wesley G. Skogan (ed.) Sample Surveys of the Victims of
Crime. Cambridge, MA: Ballinger Publishing Co., 1976, Chapter 6, 105-120.
Victimization surveys were conceived to provide new and more accurate measures of the
incidence of crime. Direct interviews with victims enabled us to bypass the fallible
data-gathering activities of local police departments and gather wide-ranging and detailed
information on the experiences of citizens both with crime and with the criminal justice system.
One of the first uses of the victimization survey data collected for the federal government by
the Bureau of the Census was to compare them with the official crime statistics published
yearly by the FBI. This was an obvious first step, for the discrepancy between crime estimates
made through surveys and crime reports filed with the FBI was quite large. The Census
Bureau's National Crime Survey uncovered about three times as many crimes as had been
recorded by the police. The media focused on the differences between official and survey
crime figures for specific cities. The New York Times, for example, printed the two side by side
for eight large cities. In a front page article, David Burnham argued that the publication of such
statistics would increase pressure on local criminal justice agencies to improve their
performance and end FBI and local police department domination of crime statistics. This
chapter examines the relationship 1) between official and survey measures in crime and 2)
between each type of measure and the "true" crime rate. The chapter also examines the
consequences of these observations for criminal justice planning, management and evaluation.
Measurement Abstracts