This is an oldie, but not dreamt up by me. I know it as the Barnum Effect, named after the circus man who had “something for everyone”, but Wiki has it as the Forer Effect.

For this to work and be student-led you need confederates, a couple of students who “run” this activity. You need a generalisable, vague description of traits and behaviours which seems to be highly individual. Then you print this as many times as you have students in the class, perhaps with a fake web address at the bottom, and put each copy in an envelope, each envelope having the name of someone in the class.

The stooges tell the class that they have accessed a computer programme which gives accurate horoscope type predictions of personality based on birthdates-based star signs. After all, astrologers believe that their subject is a true science. They take a list of names plus birthdates/signs, and disappear for 10 minutes supposedly to print off the readings but actually to add the birth date/sign on each envelope. When they return, each person is asked to read their own personality prediction silently and secretly – in case they find something they don’t want others to know. They replace their prediction in the envelope, and then one by one give an assessment of accuracy of that prediction – could be on a 1-5 scale, or a Likert scale, or similar.

You could have a discussion on how to present the data descriptively, or any other discussion, but eventually the stooges have to reveal what really occurred. Whenever this has been done participants have given their predictions a high score of accuracy, and are amazed when they read identical predictions for others. There is the ethical issue of deception here, but this is a class exercise and with full debriefing it has not produced any discomfort in the past.

Derren Brown did this for his TV show some years back, but hopefully they won’t have seen or will have forgotten it.