Food insecurity has remained pervasive for most Kenyan livelihoods despite the implementation of substantive interventions by the government and its development partners, since it gained independence in 1963. The inability to isolate distinct determinants of food insecurity for each livelihood group has led to interventions and solutions that have entrenched food insecurity rather than mitigate it. The key impediment to a livelihood-level analysis of food insecurity is the use of data and information collected at district-level administrative units, coupled with the absence of a robust analytical methodology. This study set out to identify determinants of food insecurity for three distinct livelihood groups in Kenya, namely the pastoral, agro-pastoral and marginal agricultural groups. The study also sought to empirically evaluate incremental impacts of identified determinants of food insecurity for each group. The outcomes were intended to inform the selection of particular indicators in order to target, monitor and identify important inter-relationships between variables for each livelihood group. Few studies have applied heterogeneous ordered logit regressions to livelihood-level data to evaluate food security determinants among livelihood groups and a comprehensive livelihood analysis of the determinants of food insecurity has not yet been undertaken in Kenya. Yet, Kenyan livelihoods are highly diverse, and livelihood characteristics transcend administrative boundaries. This study used a heterogeneous ordered logit to model determinants of food security in Kenya. The variables were: conflict, HIV/AIDS, rainfall, flooding, proximity to markets, migration patterns, food consumption sources, income contribution sources and own farm production. Results of significance tests and residual variability from the ordered logistic regression led to the identification of important determinants of food insecurity in each of the three livelihoods. The degree to which each of the variables was influential in accentuating food insecurity in each livelihood, was also evaluated. Determinants of food insecurity and their inter-relationships informed the selection of indicators for monitoring. Proximity to markets seemed to have a marked impact on food security in the pastoral, agropastoral and marginal agricultural livelihood groups. Conflict was influential in determining food insecurity, particularly for the pastoral and agro-pastoral livelihood groups. HIV/AIDS prevalence in the community was critical in determining food security status for the marginal agricultural and agro-pastoral livelihood groups. Rainfall was an important determinant of food insecurity in all the groups. Flooding had no significant impact on food insecurity. The results showed that an increase in the number of food sources improved food security in the pastoral, agro-pastoral and marginal agricultural livelihood groups. A diversity of income sources improved food security in the pastoral and agro-pastoral livelihood groups. The pastoral migration pattern seemed to have a substantial impact on food security especially in the pastoral and agro-pastoral livelihood groups. Own farm production was also influential in determining food security in both groups. The study outcome provides a basis for identification of important monitoring indicators including agro-climatic, trade and market processes, migration dynamics, income and food sources and the stability, settlement patterns, key livelihood and coping strategies in the three livelihood groups. The strong inter-relationships between variables suggest that multiple variables need to be monitored concurrently to address livelihood food insecurity in Kenya. The findings suggest that livelihood approaches are central to identifying determinants of food insecurity in Kenya. The outcomes of the study provide a basis for informing interventions intended to reverse food insecurity in Kenya for each distinct livelihood group. Further research could include an analysis of the impacts of seasonality, an in-depth analysis of the markets and their marked influence in affecting food security, and applications of similar methodologies to evaluate of the food insecurity of livelihood groups that were not covered in this study. Outcomes of this work are expected to provide a basis for formulating livelihood-specific interventions in Kenya. The results will provide a platform for further interrogation of important determinants of livelihood food insecurity by governments, researchers, and development partners. Aspects of the methodologies applied in this study can be replicated in adjacent countries with food security and livelihood characteristics similar to Kenya, such as Ethiopia and Somalia. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.