This paper examines the adverse effect of natural resources scarcity on children's schooling and the possible gender bias of resource collection work against girls' schooling. It uses cross-sectional data on 316 children aged 7–18 years collected from 120 rural households in Tigray, northern Ethiopia. The two-stage conditional maximum likelihood estimation technique is employed to take care of endogeneity between schooling and collection intensity decisions. The results revealed that a 50 per cent increase in collection intensity reduces the likelihood of child schooling by approximately 11 per cent. However, we find no evidence of gender bias against girls' schooling.