I am researching the case of student exchanges between European countries, within the European Union ERASMUS programme. The network that I am studying is a closed non-binary network, in which the nodes represent the countries participating in the programme and the ties indicate the existence and size of student exchanges between any two nodes.
I am trying to run a few QAP regressions (eg. using cultural proximity, geographical proximity, etc), however, I have encountered a theoretical problem: not all permutations of the shuffling of rows and columns for QAP are logical for the network that I want to explain. In particular, within the
ERAMUS programme students from EU member states can choose to go to EU member states, to candidate member states and to non-EU members of the European Economic Area (EEA). However, students from non-member EEA countries and applicant states, can only go to higher education institutions in an EU member country. Students from different home countries thus have different choice sets, which constrains the direction of their migratory patterns.
Does anyone know if there is a way to incorporate these constraints in QAP regressions using UCINET?