How to formulate the theory of setting up an experiment?

Question

Test Driven Development

There was a need for an academic understanding of the theory of setting up an experiment, so that a correctly set experiment was not an experiment for the sake of an experiment.
In IT words: in the vast majority of cases, a test is written in order to pass it, and in order to achieve maximum coverage - mentioning code instances when executing the test. Naturally, this does not give a 100% result with 100% coverage and 100% successfully destroyed mutations.
I would like to categorically avoid this, but I could not find any practical approaches other than single-blind and double-blind.
I would be glad to discuss practical approaches to testing.