In scientific terms, a statement is a hypothesis until it's proven beyond a reasonable doubt...
Further, when hypothesis testing, by convention we do not test the research hypothesis, rather the null. And if the results are significant, we reject the null, with such results suggesting support for the research hypothesis. If the results are insignificant, then we accept the null, and find that the research hypothesis was not supported; i.e., we would not use the terms "disprove" or disproven when referring to the lack of support.
Furthermore, as pertains to the results of the scientific method, there is always some measure of "reasonable doubt." For example, some reasonable doubt can be attributed to error in measurement, systematic error, etc., consequently we attempt to measure the probability of this error, as well as to estimate the confidence levels (and confidence intervals in certain types of research), essentially to give us more confidence in what the results suggest.
There are other sources of "reasonable doubt" when applying the scientific method. It is difficult to test theory directly, but we do operationalise theoretical concepts into measurable variables, stating the assumed relationships between variables in research hypotheses, before testing them in the null form. In addition to the problems associated with error mentioned earlier, we always struggle with attempting to identify all the necessary conditions that may be sufficient to suggesting a cause-and-effect relationship. Once again, we never prove or disprove, rather such tests may support or not support the hypothesis, suggesting some result.