Journalists and statistical thinking

Earlier this month, results of the baccalauréat came out. It is the French examination you have to pass to graduate from high school. B. Coulmont, a professor in sociology at Université Paris 8, made this plot representing the number of mention Très Bien (highest honors) according to the first name. He insisted that first name do not determine academic success, but instead are a good proxy to one’s social class. But soon all the french media were predicting one’s success according to one’s first name (“Tell me your first name, and I’ll tell you your grade”). This reminded me of the French presidential election last year. On the same day (13/03/2012), two different polls put each of the two major candidates ahead. But how could that be? Who was right? One had to be wrong! France (and its journalists) had just discovered confidence bounds.

H.G. Wells once said:

“Statistical thinking will one day be as necessary for efficient citizenship as the ability to read and write.”

I believe this is especially true for journalists, who spread information. And of course it also applies to the public that must be able to do critical thinking. Unfortunately, we are not there yet.

First names and mention Très Bien at Baccalauréat 2013 — by B. Coulmont

First names and mention Très Bien at Baccalauréat 2013 — by B. Coulmont