Statistics on municipalities, or local government areas (LGA's), are frequently used as a basis for research studies. In reality, these categories do not define a socially homogeneous area, but merely arose through some historical accident or major population shift. In comparisons of crime rates between different municipalities, however, it is often assumed that problems occur uniformly throughout the area. This study, which questioned the appropriateness of this assumption, selected three local government areas for closer analysis regarding variations in juvenile offender rates. Areas chosen had a high degree of social mix, as judged by housing types and confirmed by recent census data. Using comments of local residents and census information, municipality A, with an average offending rate, was splite into five homogeneous areas; the very low rating municipality B was divided into seven areas; and the high rating municipality C was split into six areas. Analysis of the juvenile offender rates showed that municipalities B and C both had one geographically small area with a very high offending rate, apparently generated by residents of a public housing development. These findings indicate that an overall rate for a municipality may not constitute a realistic base for welfare and allied funding and that area specific data should be collected to target problems accurately. One figure and four references are provided.