By Tim Harford.
"Being open about uncertainty is not just a case of reporting some kind of statistically-derived “margin of error”. There are many ways for a conclusion to look statistically robust but be wrong. What is needed is to be clear about the underpinning assumptions and open-minded about what would happen if the assumptions were mistaken.
...Politicians are creatures of certitude: they join a tribe of like-minded people, convinced that the tribe on the other side is wicked and stupid. The media love certitude, too. Newspaper editors hate headlines with “may” or “might” in them."