Obtaining data from migrant workers through self adminis tered questionnaires will run into similar difficulties, i.e. how to locate the workers and how to get them to participate. However, when large datasets obtained with a similar questionnaire administered in multiple countries are available, making comparisons across countries and between professions still seems like an interesting option to us. The purpose of such analyses would not be to find ‘accident prone cultures’, but rather to explore the way workers with different national backgrounds approach safety and perceive the people who should support them with this, i.e. managers and safety professionals. Unfortunately, the aim of questionnaire designers is often to develop a questionnaire that is valid across professions or nations, i.e. a questionnaire that ‘works’ both in Denmark and Poland or Italy in the same way. Different scores are interpreted not in terms of the nationality of the respondent but rather in the way that safety is perceived at a particular location, usually ignoring any influences of the respondents’ national background. When more is known about the way in which differ- ent migrant workers approach safety and its direct supporters (i.e. supervisors, first-line managers), these very supporters can be trained in using other strategies with different migrant workers in getting the same results, as is already more common at educa- tional institutes enrolling multiple nationalities.