It is because of this test, there is a lot of problems in modern science, and often some of the facts are accepted without a clear confirmation of objectivity. But these situations are relatively rare and often associated with the economic side of confirmation. But science is not only a collection of facts. For even more details, read what Michael Chabon says on the issue. Facts are useless in principle, as they say, what happened, but not what happens. If Maria hit Pete yesterday, it does not mean that it will hit it today. Therefore, the facts in science is not the goal but a means. A goal is to determine the cause-effect relationships. Knowing only that Mary hit Pete, it can be assumed when it hit it next time.
Assuming that the cause is hurt, we form the hypothesis that needs to be tested. Masha asked why she hit Pete and heard "offended", our hypothesis passes into the category of theories. And if after a long time, the theory will give correct results, then it goes into the category of law. Scientists learn about our world, forming hypotheses, theories and laws (laws usually have strict scope). Hypothesis of kind are the approximation of (joint) of the facts. A way to "connect" the facts there is a different set. Therefore, testing of hypotheses in science are too strict. As time shows, even this severity can not to create a theory that explains everything: many theories give accurate predictions in the same area, but in a different area, they give the opposite result.