Scientists think if they design their experiments to be testable (TestDrivenDesign), they can then report the results and let others re-do the experiments and get the same results, or others can invent new experiments that leverage the prior not-yet-disproven results as foundations.
See CodeUnitTestFirst.