Estimating the impact of COVID-19 on the multiple dimensions of child well-being requires quasi-random variation in exposure to it, which is unlikely to occur during a pandemic. Recent developments in econometrics have highlighted the relevance of subjective evaluations of treatment effects in the absence of randomization. This paper delivers new evidence, based on primary data collected in Luxembourg in Spring 2021 about their subjective appraisal of the effects of COVID-19 on multiple dimensions of children’s well-being. Effects are recovered through specific survey questions, asking children to compare actual outcomes with counterfactual ones, that they believe would have occurred in the absence of COVID-19. Children report negative effects of COVID-19 on subjective health and on schooling outcomes, as well as disruptions on the time used to interact with the family. The paper explores the sources of heterogeneity behind these results.