This paper analyzes the impact of COVID-19 lockdown measures on children's school resilience. Using an individual fixed-effect linear probability model on Nigeria data, it exploits the quasi-randomness of these measures to estimate their effect on school attendance after the lockdown was lifted. The results show that COVID-19 lockdown measures reduced children's probability of attending school after the school system reopened. This negative impact increased with children's age, reaching a peak among those whose education was no longer compulsory. For schoolchildren in that age group, the negative effect of COVID-19 lockdown measures is likely to be permanent, which, if not reversed, will undermine the quality of the economy-wide future labor force. The paper also finds evidence that in the child marriage-prone North-West part of Nigeria that these measures increased gender inequality in education among children aged 12 to 18. This result suggests that COVID-19 lockdown measures may exacerbate harmful traditional practices such as child marriage.