I think there's one more step that goes a bit earlier in the process. Before an observation is made, there needs to be a choice on what experiments are done. Typically, those measurements will be done with one of two goals. The first is to measure a free parameter in the theory that isn't well-constrained by existing experiments. (e.g. "According to Maxwell's equations, the speed of light should be constant. Let's measure that constant value." [0]) The second is to find an unintuitive prediction from current models, then measure what happens in those circumstances. (e.g. "According to special relativity, passage of time should vary based on velocity. That's weird. Let's measure and see if it is the case." [1])
The purpose of these is to narrow down which experiments are being performed, trying to choose experiments that will try their hardest to break an existing theory.
[1] https://en.wikipedia.org/wiki/Hafele%E2%80%93Keating_experim... (This article currently has an incorrect edit made yesterday. It seems to be based on a mistaken assumption that the rotating earth is a stationary reference frame. The opening paragraph should read "were consistent" rather than "were not consistent".)
No, science starts with an observation which is not explained by the prevailing theory. Then you form a hypothesis, etc.
At least in principle. The reality of science is a lot messier.