After scientists ask questions, they then answer them, but only tentatively. That tentative or temporary answer is called an hypothesis. Now comes the interesting part. In order to qualify as science, the process must continue, and the hypothesis must be tested. Those three steps, a question, a hypothesis, and the test of the hypothesis, are what constitute the core of science.
What are two non-scientific ways of dealing with the green cheese hypothesis? The first might be to vote on it, a popularity-based method. So, I could ask a large number of people whether or not they believed that the Moon was made of green cheese, and if more than 50% agreed, I would accept the hypothesis.
Another method might be to go to an authority, someone who knows. We have used this method all of our lives. For example, when we were younger we asked our parents for answers, they were the authorities. So, in this example, we might look for someone who knows cheese. We might ask the Wisconsin Cheese Board; they should know cheese. If they say the Moon is cheese it must be so.
Of course, we can now see how popularity and authority are not scientific. How would a scientist test the green cheese hypothesis? Obviously, the test would involve some type of visit to the Moon, and the collection of a lunar sample. If the lunar sample is green cheese, then the hypothesis is true. But what if it is not green cheese, is the hypothesis dropped? Not necessarily.
Confirmation of a hypothesis is powerful, but lack of confirmation is not. Back to the example. Suppose I went to the Moon and did not find green cheese. Someone back on Earth might ask whether I searched the entire Moon, including the side that never faces Earth, or 100 ft. under the surface, and so on. It is possible that the green cheese exists but I just did not find it.
Disconfirmation usually leads to changes in the hypothesis. For example, the hypothesis might change to: Green cheese is only on the side of the Moon that never faces Earth, or, Green cheese is 100 ft. below the surface of the Moon. Now, these new hypotheses will need to be tested as well. After many modifications of the hypothesis, and after many tests, and after none of those hypotheses and their tests have been confirmed, then that line of investigation may be dropped. (Loss of interest by funding agencies is also a large determinant for the cessation of research. A long history of research without confirming results will likely lead to loss of funds for that project.) That is why some hypotheses may linger on for years after they have been tested and seemingly disconfirmed.
The point is that one confirmation may be sufficient to prove a hypothesis, but no number of failures to confirm can disprove a hypothesis. The situation is profoundly asymmetric.