"Methodological Issues in the Measurement of Crime," in Hans-Joachim Schneider
(ed.) The Victim in International Perspective. Berlin and New York: de Gruyter, 1982
(English and German editions).
Attempts to assess the nature or frequency of crime using the survey method quickly
encounter several conceptual, methodological, and procedural barriers. The problems and
puzzles which make up these barriers have confounded efforts to generalize reliable and valid
measures of crime which are of utility to scholars or government officials. Here I review these
barriers; most of my attention will be devoted to the National Crime Survey currently being
conducted by the United States Bureau of the Census, if only because extensive
methodological investigations have laid bare most of the shortcomings of that effort. These
are five major difficulties facing those compiling "moral statistics" concerning criminal
“Innovations in the Analysis of Crime Surveys,” a paper presented at the
Conference on Measurement and Research Design in Criminal Justice, Griffith
University, Queensland, August 1990.
Written for a conference on crime surveys in Australia, this paper presents several excellent
recommendations for the conduct of victimization surveys. It does not simply focus on their
"analysis," for it would be mistake to hope that more sophisticated analytic methods can
somehow overcome neglecting to gather key data elements in the first place. The paper
discusses four ways in which the utility of large-scale household victimization surveys could
be enhanced: by bringing geography back into the picture, improving the treatment of
multiple victimization, more consciously constructing the surveys around a conceptual
framework which enhances their explanatory power and substantive appeal, and not doing
them at all. Alas, no one has ever taken any of these recommendation to heart.