For our logistic regression analysis we looked at the determinants of average or worse health. The variables were considered in 3 domains: demographic, socioeconomic and psychosocial. This categorisation is somewhat artificial and there are lots of different ways of categorising the variables but we wanted to emphasise the differences between demographic variables which we considered as givens and socioeconomic factors which we restricted to variables with a direct impact on material status. The latter therefore fit most closely with neomaterialist explanations for health inequalities. Education was considered as a demographic factor as it is not as clearly linked with status in the FSU as it is in the west.