I've actually tried that peanut butter bagel experiment myself and didn't notice any significant difference. Of course experiments with a population of 1 and no control group are meaningless.