Measuring What Matters: Crime, Disorder, and Fear

Wesley G. Skogan

This paper considers two issues: measuring the possible effects of an innovative policing program, and doing so in a framework that could support the inference that the program caused variations that the measurements might reveal. Measurement involves the collection of data that represent -- sometimes only indirectly -- the problems that are targets of programs. These are the "outcome" measures, and it is vital that they represent as accurately as possible the scope of a program's intentions. The framework within which these data are collected is evaluation's research design, and it is crucial that the design account for as many alternative explanations for what is measured as is possible under the circumstances. Arguing that "the program made a difference" over the past month or year involves systematically discounting the potential influence of other factors that might account for changes in the measures, through the use of randomization, matched control groups or time series, and other design strategies. This essay focuses on measurement issues, but it bridges to design issues through some concrete examples of how measures have been used to make judgments about the impact of programs. It examines in sequence some of the experience of the evaluation community in taking the vital signs of a community via measures of crime, disorder, and fear.

PDF icon Download working paper PDF