The common presumption that orphans are less likely to attend school than non-orphans is re-examined using survey data from two regions in Tanzania. It is argued that orphans should not be compared simply with non-orphans since there are other vulnerable groups of children. Further, with particular reference to place of residence, it is argued that orphans should not be viewed as a homogeneous group. In Tanzania both orphans and a second, sizable, also potentially vulnerable group of children - children who have not lost a parent, but who live with only one or neither of their parents ('children from disjointed families') - are less likely than other children to attend school in urban and roadside settlements but hardly so in rural areas. This is explained by most orphans and children from disjointed families residing in female-headed households, and by female-headed households struggling more in urban/roadside than in rural areas, because of the different histories of women's economic activity and different livelihood structures between those areas.