What do you do when you encounter a study with a validity/bias issue?