By that logic, you could just not do the study at all and claim it as evidence of absence. Statistically, it's the same problem. You need data (and lots of it,

By that logic, you could just not do the study at all and claim it as evidence of absence. Statistically, it's the same problem. You need data (and lots of it, relative to the noise) to provide evidence of anything.