When testing hypotheses, it is important to carefully scrutinize the data for the possibility of influential cases, especially in analyses with relatively low numbers of cases, as in most cross-national multilevel studies.