Science is the best way to understand the natural world. However, the black box of statistical science, is not the most rational approach to science. How can you reason inside a black box when you cannot see anything in the darkness? Margin of error does not allow for any facts. Facts are sharp points and not fuzzy dice. This is not fully rational science. It is more like oracle science; fortune teller, and not exclusively based on reason. It is somewhere between religion and reason. It still has spirits like Lady Luck and Murphy law.
Rational science has a much higher standard. For example, Einstein's theory of relativity are expressed with math logic; equations. If anyone was to run an experience and one data point did not touch these math curves, the entire theory would be re-examined.
With statistical science you draw your curve, after the fact and curve/ theory is the best fit to the data even it does not touch any of the data points. Rational theory has a higher much higher standard. I am not sure if statistical theory can define reality. It can correlate but that is not the same as explain with reason. Statistical should be called Science lite.
Another problem with science lite, is since you start in a black box, how do you know you are running the correct experiments for what is in the black box. If you run the wrong experiments your data will never touch the correct line to create a self fulfilling prophesy of randomness. A rational theory knows how things work in advance; open the box and see. Therefore it can run the correct experiments so all the data touches. This is what is needed for applied science.
The Golden Age of science was a more rational time. Most of the advances, now are in Applied Science and Technology; tools, but little in pure science, due to too much science lite. Applied science has to be more rational, since irrational is not reliable or stable in reality.