This paper shows that the Demographic and Health Survey Wealth Index need to be used
with caution in the analysis of household poverty in Kenya. Thus, we depict a case of varied
results when the wealth index is directly used with regional comparison, and discover that in
both the binary and ordered logistic models, the years of education of household head, their
marital status, the size of a given household and the region of residence (province) strongly
determine household welfare status. We also observe that these characteristics are even more
important in explaining household probability to poorest, and thus, lay emphasis on results
obtained while controlling for household region of residence (province) to those that
distinguish between rural and urban households.