This research explored associations between refugee housing characteristics and desired refugee outcomes – namely health & well-being and self-reliance – and examined plausible mediating mechanisms that may explain the associations between these variables. It is one of the first studies to directly examine the relationship between refugee housing and refugees’ ability to be self-reliant, in alignment with Housing First models. A sample of 82 participants was randomly selected in three zones of Bidibidi Refugee Settlement in Uganda, where residents are largely responsible for constructing and maintaining their own earthen housing. These housing structures and layouts were described, and a matrix of bi-variate correlations was examined. A series of significantly correlated variables was tested for mediation based on the hypotheses, and this mediation model was then tested for moderation by demographic variables (zone, sex, marital status, hut tenure, and age). Results identified several statistically significant housing variables relevant to self-reliance and health & well-being. Presence of pests within households was associated with resource access, and sleep quality served as a significant mediator in this model. No demographic variables served as significant moderators for the pests-sleep quality-resource access mediation model. Ultimately, these findings help identify specific future points of intervention for settlement housing, thereby informing projects, policies, and the design of refugee settlements going forward.