Question: how have you tested and proved your theories?

Melanie Stefan answered on 15 Mar 2011:

Generally speaking, it’s impossible to prove a theory. We can only prove a theory wrong (by finding a counter-example) or gather support that will favour a theory over another (until a third, and even better, theory comes up)

In my research, the process is something like this: I have a question (say about how a specific protein in the brain is regulated) and I come up with a theory about it. I usually go through the following questions:

1. Does the theory make any sense at all? (i.e. if I build a computer model based upon it and run a simulation, does it actually behave the way I think it will)? If it doesn’t, my theory is obviously wrong, so I have to come up with something new.
2. Is the theory compatible with everything we know about the protein? I.e. do my simulation results match known measurements and experiments? If they don’t, then my theory is wrong, or at least, incomplete. Note that this is an ongoing process, because as soon as someone gathers new data about my protein, I will have to check whether it fits my theory.
3. Can my theory explain what happens under non-ideal circumstances, for instance if the protein is mutated? If it can’t, it’s not a very good theory, so I have to abandon (or change) it
4. Are there other theories about how the protein works? If yes, can they also explain all the known data? If they can’t, then my theory is obviously better. If both theories are equally good at explaining the data, I have no way of deciding which is better (for the moment).
5. Can my theory make testable predictions, i.e. can I suggest an experiment that has never been done and correctly predict the outcome? If yes, that makes a strong case in favour of my theory. Although it still does not prove it!