Here's how I would do that: Imagine 10000 people. 0.01 of them, or 100, have a headache. 0.04 of those people, or 4, have fever. But we also know that 0.02 of the original 10000, or 200, have a fever. So 4 of the 200 people with a fever also have a headache. Given that person has a fever, there is a 4/200= .02 probability they also have a headache.