Objectives: Compare the spatial offending patterns of arrested offenders to that of non-arrested offenders, in order to assess selection bias in arrest data.Method: Data used for this study are from the Dutch DNA database for criminal cases. DNA allows reliable linkage of serial crimes committed by the same offender, whether or not the offender has ever been arrested. Spatial offending patterns of arrested and non-arrested offenders are measured by calculating the mean intercrime distance (MICD) of the offense locations. Survival analysis is performed to study whether the MICD has an influence on the duration until arrest.Results: No large differences are found between the MICD of arrested offenders and the MICD of non-arrested offenders. The MICD does not affect the duration until arrest.Conclusions: Because no differences in the MICDs are found between the arrested and non-arrested offenders, arrest data are probably less selective than has been suspected in the past, and results based on these data are unlikely to be strongly biased.