Summary Objectives - To undertake a case-control analysis of the health, nutrition and caring practices of orphans enrolled in primary schools in Ethiopia. Methods - Pupils of both sexes aged 7-17 who were randomly selected from Grades 3 and 4 of primary school during a national survey of schoolchildren in Ethiopia and who were classified as an orphan were matched by age, sex and school with non-orphans. Logistic regression was used to compare children in terms of indicators of anthropometric and nutritional status, chronic infections, personal hygiene, diet, caring practices and self-reported sensory disability. Results - Of the 7752 children in the national survey, 1283 (16.9%) had lost either both parents or one parent. Of these orphans, 1048 were uniquely pair matched for the case-control analysis. About 60% of orphans had lost their father, and about 20% each had lost their mother or both parents. Orphans had better anthropometric measurements and indices than non-orphans, although the differences were small, and they were less likely to have a goitre (OR = 0.68, P = 0.011). There were no differences in the odds of infections. Orphans were less likely than non-orphans to have eaten breakfast or fruit and vegetables on the previous day and were more likely to report having trouble seeing and hearing. Conclusion - Orphans were slight better nourished than non-orphans, but this could have been a result of asystematic bias in underestimating the age of orphans. The indicators suggested that orphans were less well cared for than non-orphans, but the differences were small.