pretty sure there's more than 23 test subjects for monkeys.

If anyone would be cherypicking a dataset, it'd be picking the 15 survivors of the group of 23 given the holistic survival rate is closer to 2%