A study in the New England Journal of Medicine regarding differential publication of antidepressant trials - such that if you looked at only published results, you would think ninety-four of trials were positive for the drugs they studied.
However, the authors found many of the negative trials were not published. This leads to unrealistic estimates of drug effectiveness.
In fact, the FDA analyzed all the data (including trials that were not published) and found only fifty-one percent of all studies of antidepressants showed positive results.
When they looked more carefully at the effect sizes needed to show results, the positive outcomes came down to thirty-two percent overall.
This study indicates a need to look past headlines when evaluating both medical research and medical journalism.