The experiment doesn't have to be on people, it can be on budgets and resources. Tracking whether or not or how many people get help vs punished might be a plac

The experiment doesn't have to be on people, it can be on budgets and resources. Tracking whether or not or how many people get help vs punished might be a place to start