Yes, so lets revise that. Science must have physical data, data verified by the senses. And those data can be gathered from observation out in the field or in the laboratory.
Hypotheses can be formed and evidence gathered from observation to support them, but the observations have to have priority over theory.
Yes, we agree on this. But I would like to add that there is theoretical science as well. For example in physics, theories like superstring theory and M-theory are purely theoretical. The theory is quite elegant, but it proposes vibrating loops or stings so small, that there are no instruments sensitive enough to observe them. When do we get the data to check the theory? Nobody knows, if ever.
This has been the problem with astronomy since Hubble. An idea (redshift as exclusively the product of recessional velocity) gained ascendency and now shoehorns observations (such as quasars as ejecta from certain types of galaxies) to fit the theory or ignores them entirely.
That is the sort of thing that most people do not realize about science. Sometimes scientists are not very scientific when they ignore data that contradicts their favorite theory. Here is another example. Again in physics, relativity and quantum mechanics contradict each other. They cannot both be true. So for about 50 years, physicists worked in one field or the other, and simply ignored people working in the other field. But that has changed over the past 15 years or so. Physicists are indeed talking to each other to resolve that contraction.