We generally don't need to figure out "when to stop."
Because we don't stop. No, we don't run experiments to see if a pencil falls when you drop it, to use your example. I suggest this is not a good example because it's a 'test' the layman tests everyday. But, no matter. The point is that the science are doing now uses the falling pencil as a hypothesis.
Let's suppose gravity fails tomorrow (the layman would notice - again, not a great example for my point). We run an experiment which assumes that gravity exists, and is around 1g at the earth's surface. What happens?
We don't get the predicted results. Opps, something is wrong with one of our assumptions. Which one? Well, we will start out by assuming that the hypothesis under test is incorrect. However, meanwhile 1000's of other people are running tests with assumption of 1g gravity, and those tests will fail. Eventually, someone will get the bright idea to question the existance of 1g gravity, device a test, and find out, that indead, now we have -1g gravity, or whatever.
That's the neat thing about a world whose behavior is based on a few physical principles. Change the principles, and it has a cascade effect that you quickly notice. And, if it doesn't change anything you are looking at, then does it really matter that you haven't noticed it yet?
It's not like scientists sit in a room and have ponderous conversations to decide "the truth". Hypothesises slowly become accepted. "It's probably true plate tectonics exists" leads to experiments and observations. When these agree with the assumption, eventually all geologic science uses the theory of plate tectonics. There's no court of science, no judge. You think you have an idea about the nature of the universe? Get funding, and study/experiment. Guess correctly enough times, and you have a fruitful career in science. Fail? Well, there's plenty of community colleges needing Physics 101 instructors. The person who assumes plate tectonics is false does not get results. Reality sorts out the wheat from the chaff.
ETA: of course, the question is extremely important for individuals, funding reps, etc. "Do I build a hypotheses based the theory of tectonics, or do I try to prove/disprove tectonics itself" is a career make/break question (well, it was at one time). There's no easy answer, as figuring out the answer requires knowing the result you are trying to arrive at. So, we muddle through the best we can.
)
-9000 of "why can't we accept that goddittit?".
