Knowledge on both the prevalence of and the risk factors for juvenile delinquency rely almost exclusively on data generated by self-report research. However, several errors can occur during the process of data gathering, all of which may have an impact on the validity and reliability of the results. This has become more apparent as self-report data on (juvenile) delinquency are increasingly being used to assess the prevalence of delinquency in society. In this article we try to determine how different administration modes may result in different results with regard to the prevalence of juvenile delinquency being reported as well as the predictors of delinquent behaviour. In order to decide whether these differences can be seen as mode effects, the possible influence of selection effects – or differences in the sample population – must be controlled for. For this aim, the technique of case-control matching is used to match the samples of a mail survey and a school survey, both administered on a representative sample of Flemish youth attending secondary education. Using the technique of case-control matching, we aim to control for differences in non-response error, in order to disentangle non-response effects from effects related to administration mode.