Decision Making
Wisdom of crowds



October 28, 2023

Sometimes teaching turns out to be very different from the usual lecture on a Wednesday. I was asked (well I signed up) to teach in the Kinderuni (children’s university) at my home institution, the University of Bern. This was part of program where children between the ages of 8-12 are invited to participate in a lecture (Friday) and then in a seminar (Saturday). In what follows I will mainly talk about the lecture.

Figure 1: Kinderuni


I talked about decision making and experiments as a methodological approach. My aim was to have many examples and run experiments with them in class. 75 children showed up on Friday for the lecture. Having a mobile makes running in-class experiments really easy - but I could not rely on that - hence back to paper and pencil. We started off with an anchoring task, the classic: multiply the following numbers 1*2*3*4*5*6*7*8*9 for one group and the reverse for the second seemed a little to hard for 8-years old so I used addition instead 1+2+3+4+5+6+7+8+9. This worked reasonably well (a brave helper typed in the numbers, while I worked through another example with them) and we could demonstrate quite a substantial anchoring effect (400 v 1400) which was explained back to me, by one of the kids in great detail - pretty cool.

Wisdom of the crowds

As mentioned above, while a lot of typing was going on I let them estimate three quantities: I had 1) a big glass vase with ping pong balls. 2) a picture of many Legos and 3) I let them estimate my weight.

Figure 2: Estimations

There were 44 balls, 88 Legos and my own weight (on the morning of the lecture) 76 kg. The average of the whole group was off by 1 ball (pretty good!), 4 Legos and 5 kg (I just look too slim I guess) - we talked a bit about individual choices and the averages - again tricky for the little ones, pretty ok for the older.


We moved on to decision making without risk (food choice) and with risk. Here I tried to show them base rate neglect with several picks of a certain gummybear color (green or red) from different distributions.

Red and Green We started out with an easy task - do you prefer A or B if you want to draw a red gummybear?

Easy call for red After some further easy tasks - something more tricky - after bringing these in order, it became clear that the two distributions were the same - so same probabilities for A and B.

Figure 3: Equal probabiliites

Finally the real tricky one:

Figure 4: Tricky

Just going with the number of red ones (if you want a red) is misleading … we do a bit of maths and decide that A is advantageous. Clearly the 8-year olds had troubles with this - but the older ones totally got my point and explained that A simply has the larger probability of a red gummybear - hence this one is preferable - again - pretty cool.

Figure 5: Risk

Game theory

Finally - and definitly most chaotic of all - we played two rounds of game theory. First a dictator game, imagining one of the other kids as their partner. Then I paired them up and let them play an Ultimatum game in pairs with (Sugus)[] sweets.

Figure 6: Dictator game

Figure 7: Ultimatum game

Both rounds were pretty wild but we ended up in a nice discussion on fairness - a fair offer (3:3 when 6 Sugus were distributed) was clearly identified and other cases mentioned. Some interactions ended in interesting distributions like 6:0 cause one of the two kids did not like Sugus - so well coordinated :).

Take Aways

  • I was way off how long things take - favorite example: I told them to form pairs of two for the ultimatum game … it took us about 10 minutes to achive this, planned about 2 :)
  • responses in paper and pencil are tricky to live analyse - I normally run FormR as a frontend and a R script in the background that analyse/visualize data automatically - easy, once setup. Here we did two things - entering values from the response papers into Google Sheets and sticking dots onto a flipchart to generate a distribution graph - both works but both takes a lot of time (even with 3 people in the room)
  • Kids (at least in this class) are super responsive and engaged - I am used to getting not a lot of feedback/questions from my students - with the kids 7-10 hands are up for every questions - great, but again a time issue if you want to address the questions properly


Here is a pdf export of the Keynote slides I did.