P value less than 0.05 is considered as the gold standard to determine the reliability of experimental data. This standard supports most published scientific conclusions, and it is difficult to publish papers that violate this standard, and it is also difficult to get funding from academic institutions. However, even Fisher understands that the concept of statistical significance and the P value supporting it have considerable limitations.

P value research

The value of p is often misinterpreted, and the statistical meaning is not equal to the actual meaning. In addition, in order to make the data more beautiful, many researchers adjust the P value up or down intentionally or unintentionally. "You can prove anything with statistics," said Sander Greenland, emeritus professor, statistician and epidemiologist at UCLA.

He is one of the scientists who call for statistical reform. Only relying on statistical research often leads to inaccurate scientific conclusions. This standard can judge whether the real thing is fake or not, and it can also judge whether the fake thing is true. After Fischer retired and moved to Australia, he was asked if he had any regrets in his long career. He clearly replied: "We shouldn't have proposed 0.05 in the first place."