But look what happened after 2000, when pre-registration was brought in. Suddenly there were only a couple of positive results, with the rest of the studies reporting null effects, clustering around zero. The success rate for trials before registration was required was 57 per cent; afterwards, it plummeted to 8 per cent.
