Law enforcement agencies are adopting a variety of new surveillance technologies at a fast pace. These technologies could have substantial benefits in terms of public safety, but, for many of them, their ability to reduce crime is unknown. Although a small experimental literature addresses some of these technologies, many of the implementations have been too small to provide an accurate measurement of their potential. In this paper, I explore the advantages and make general suggestions about the use of quasi-experimental methodologies in estimating the public safety benefits of police technology. I also consider the specific case of license plate readers and provide some examples of difference-in-differences approaches that could be used to study their efficacy.
