I always wait until a particular study has been reproduced at least once before I take it seriously.