@damacri First, true experiments are the best available evidence. If not, analyses that control for or stratify by age, given what we know about how daycare has different effects at different ages. As I said above, lumping infants in with toddlers and 3 year olds data will of course produce less...