Measuring crime is one of criminology’s most central tasks. Self-report studies constitute one means of doing so. By asking people, primarily youths, about their experience of involvement in criminal acts, self-report studies are intended to provide knowledge on the extent and structure of crime and on crime trends over time, while also providing opportunities to study the causes of crime. The dissertation’s objective is to examine the use of self-reported crime as a research method. It problematises the use of self-report studies as an instrument for measuring both the extent and structure of youth crime, and also trends in youth crime over time. Problematising the method in this way both illustrates the significance of methodology and measurement instruments for the production/construction of criminological data and makes possible a more nuanced and aware approach to the use of such data. The dissertation comprises a discussion of the basic assumptions of the self-report method viewed from the perspective of theory of science, a review of Swedish self-report studies and a detailed study of the Swedish School Survey on Crime, Sweden’s nationally representative self-report study of youth in year nine (aged 15). The dissertation’s most important conclusions are that researchers, when designing a self-report study, should consider both how crime is viewed and what the study is intended to examine. It is also important, when using self-report data, to consider what the data represent. This is of significance in relation to both how questions might best be formulated and to what extent the mechanisms that affect the results will involve problems for the quality of the measures obtained.