The aim of this paper is to apply exploratory techniques in order to uncover vulnerability dimensions that characterize clusters of households within the Zambian dataset. Specifically, our purpose was to discover latent factors and to build a typology of households according to the classification results. The relationship between poverty and AIDS impact is a strong one that is known already. What is less understood is the relative importance and dynamics of different morbidity profiles across wealth groups in relation to household food security. This paper identifies some possible relationships in this regard. The datasets we used belong to a quantitative baseline survey that was conducted in Northern Province Zambia, by FAO in February/March 2004 and funded by Development Cooperation Ireland (DCI). The complementary use of two techniques, multiple correspondence analysis and hybrid cluster analysis allowed us to derive a categorisation a posteriori which confirms and adds more information to the a priori taxonomies hypothesised in the research design of previous qualitative studies (FAO, 2004). Our results underline the presence of an apparent gender divide: female headed households seem to clearly cluster with vulnerability factors (chronic morbidity; the burden of keeping orphans; economic vulnerability factors; vulnerability to food insecurity) as opposed to male headed households, that in most of the cases are married and also prosperous in terms of economic status. Chronic morbidity is severely harming the capacity of certain types of households to work and cope with the already negative circumstances, making them progressively destitute and also socially excluded. We also discuss the implications for both applied research and practical development interventions of the use of such exploratory techniques to evaluate vulnerability taxonomies.