I need to put this into different words?

Research on the neighborhood level shows consistently high correlations between socio-structural neighborhood characteristics and the number of juvenile residents committing crimes. It is often assumed that these relations are causal. In this paper this causality assumption is tested. By using multi-level techniques we try to separate contextual (neighborhood) effects from composition effects. The analyses are based on data concerning juveniles and neighborhoods in Rotterdam, The Netherlands. To ascertain that the empirical test is really critical, we use three independent measures of (the number of) juveniles committing crimes. Results show that correlations on the neighborhood level do not necessarily indicate causal (neighborhood) effects. Above all, differences between neighborhoods are caused by differences in composition. These results are interpreted in terms of the geographical distribution of social networks.|||No, if your audience is educated, it makes sense the way it is|||too long!