There is also something powerful in the concept of an ‘airport idea’. On the surface, it links with Dan’s plea for what students should do if they me… - David Franklin

In Dan’s case, the interest in policymaking is so central to the students’ presence at Harvard that it would be foolish to run a statistics course that did not acknowledge it. He goes as far as to build this into the purpose of the course: as we heard earlier, his purpose is “not just to maximise learning about statistics, but also to maximise learning of the skills that will be useful to have out there in the world.

With these issues in mind, Dan has a unique way of tying exams into the learning process. After attending a workshop run by the Nobel Prize winner Carl Wieman, he was converted to the idea of two-stage exams. The first stage of the exam is the same as any other. The second stage brings the students together in groups of four, and gives them all a subset of the questions they just answered in the first stage. The students are encouraged to discuss their answers in their groups, and they all submit a second version. Groups who manage to improve upon the students’ individual scores get a small boost in the final outcome.

Dan’s goal, taken from his mentor Richard Zeckhauser, is to get students to think probabilistically about the world. This means engaging with the challenge of understanding and accepting what 29% means in the context of a Trump victory, and fighting the brain’s natural inclination to see things in binary terms.