NetLogo comes with a library of sample models that demonstrate phenomena from physics, chemistry, biology, social science, earth science, computer science, etc. Here are some of my favorites:
This model shows how an interesting and complex fractal can emerge from the behavior of agents doing simple computations.
This is Wolfram's simple model from New Kind of Science (NKS). Using a 1D CA with only 256 possible update rules, he shows how four different levels of complexity can arise: fixed point, cyclical, chaotic, and universal.
When we think of a crowd we think of mobs, riots, stampedes. But crowds of agents doing simple things can solve complex problems.
Sometimes a small change can have a big effect. Avalanches, panics, and rebellions are examples.
How can a society be ethnocentric if its members are not?
There are four basic types of interactions between agents: cooperation, altruism, selfishness, vindictiveness.
The PD models reduce interactions between agents to a basic game.
Economies are emergent phenomena.
This is essentially SugarScape. It's a nice demo of Pareto Principle.
A general framework for competition and mating in an environment with a fixed amount of resources.
This model shows how we can think of a NetLogo model as a kind of virtual lab for experimenting with possible policies.