How do you choose the level of significance when doing a hypothesis?
Default level is
We can use any level of significance, it all depends on our tolerance for type 1 error. For example, in research papers published in physics journals, the significance level is usually extremely low, of the order of 1 in a million (5 sigma level). But for practical purposes, for example, for diagnosing a disease, a test significant at level of
But if there is no preference, the default level is usually taken as