I remember reading about an experiment performed by behavioral economists where person A divides some cash and person B either accepts their division and gets their allocated share of the money or rejects it and neither party gets their allocated share. You could say the consequentialist solution is to always accept the division of money, which most folks don't do, so this could make a good trial exercise. On the other hand, if person A is someone person B is going to have repeated interactions with, one could argue that the social capital of training person A to divide things fairly might be worth forgoing cash... So maybe it wouldn't work in a scenario where the class meets again and again? (Unless things were anonymized somehow...)

There is also the Newcomb's Problem aspect to this, where having taught the class about consequentialism will make it appear as though you have made everyone who is Person B worse off.

Reading up on experiments behavioral economists have done in general seems like it could be a good source of ideas.