Science is a process, and it's really simple:
- Form a hypothesis (make a statement of what you want to test, ensure it is testable; if it's not testable it's not amenable to science)
- Come up with how to test the hypothesis
- Test the hypothesis
- Publish the results and details
There are technical bits that make that easier, but the big parts are keeping records of what you've done and what the results were
Some hypotheses are expensive to test, best leave those to the well funded pros
That's a) impossible; and b) if limited enough to be possible would take too long to be practical