This suggests to me that there might be a follow-on problem with fabricated results. Given the "publish or perish" thing in academia, and devaluation of negati

This suggests to me that there might be a follow-on problem with fabricated results. Given the "publish or perish" thing in academia, and devaluation of negative results, it seems there's a strong incentive to obtain positive results. And this is likely to increase the incidence of researchers being pressured into tweaking data to obtain affirmative results.