How many agains makes the study wrong