I don't see how one person's blinded self-experimentation would yield more knowledge in this case, as the proposed experiment is not methodologically sound and as the data would not tell you anything more about the effect - as far as I can tell, anyway.