In reality you need to consider probabilities. What if there are thousands of times more kids in the second category than the first?

This is a pattern of thinking that I’ve seen a lot on the pro-lockdown side: reasoning from the axiom that a human being dying has infinite negative utility. While normally a good policy heuristic in practice when all else is equal, it of course can’t be strictly true. If it were, then it would be worth it for everyone in the world to experience unbounded misery if that extended the life of one person by one day.

Clearly that is absurd. So the moral calculus must be more nuanced than “if it saves lives, it must be worth it”.

One option harms every child and potentially saves some parents' lives.

The other harms practically no children and will likely lead to some more parents dying, although the data is unclear (as death rates in states with much more open schools is not consistently higher).

Harm all kids to save a few parents on the margin...potentially? Do the cost-benefit.