Given that the research questions and hypotheses which require linked macro and micro data are ones that postulate variation between the macro units either in the distributions of some of the macro variables or in the patterns of association between micro variables, it is frequently sensible to ask whether or not there is variation of this kind, before proceeding to link micro and macro data. This section discusses how this can be done. The main aim of these analyses is to identify the micro level variation that may be explicable with macro level factors. However, you should not conclude that if there is little between macro unit variation, there is no need to proceed to linking the macro data. It is possible in multivariate analyses that different processes cancel each other out, but the processes may be visible with careful modelling of the linked macro and micro data.