No, experiments are really only science in the usual sense if you have an expectation of how they will turn out and you run the experiment to test your hypothesis. And properly constructing the hypothesis, to include past knowledge and some causal mechanism is vitally important, the causal part being often neglected even in "real" science.
There was a lot of this during Covid (which I know is a bad example because of it's political baggage, but anyway). Various "x cures / helps with covid" trends (sometimes including studies) but not really based on a anything. It's not really scientific to just imagine something could work, and then try it. Contrast it with vaccines, where, regardless of your view on the covid one, there's an actual plausible causal pathway from being exposed to it to getting immunity, so it's possible and a legit thing to do to run an experiment to see if it works.
"so it's possible and a legit thing to do to run an experiment to see if it works. "
I think it is possible and legit scientific to run experiments on allmost anything, no matter how weird.
What matters is how you do the experiment and what is the result - does it work?
And if it does, is it reproducible?
Thats the key difference between a scientific and a esoteric experiment.
But you totally can investigate whether divine blessed guano shit cures cancer in a scientific way. Most researchers just won't waste their time with it. Experiments are expensive. This is why most experiments are indeed only conducted, if there is a solid theory about the process and not a shot in the blue.
(And another actually sad part today is, that many new if not most scientific papers out there are actually not reproducible anymore (at least not with the data provided), but no one cares or does something about it, except complaining, because there is nothing to win, from redoing an experiment of someone else. Reputation and grants demand new research.)
I wish there was an app where I could design experiments, and people could opt-in to participating in them. I'd of course participate in other people's experiments . The app could handle the statistical reasoning. Every now and then, there would be a strong enough signal in the data to attract "real" scientists to problems that would otherwise go unexplored.
People are gonna do weird stuff anyhow, we might as well see if some good can come of it.
There was a lot of this during Covid (which I know is a bad example because of it's political baggage, but anyway). Various "x cures / helps with covid" trends (sometimes including studies) but not really based on a anything. It's not really scientific to just imagine something could work, and then try it. Contrast it with vaccines, where, regardless of your view on the covid one, there's an actual plausible causal pathway from being exposed to it to getting immunity, so it's possible and a legit thing to do to run an experiment to see if it works.