Laboratory experiments and humans
All laboratory conditions are flawed when studying human beings, and games are no better than other labs. At Terra Nova, the hot new group blog discussing economic and legal issues of games, Dan Hunter asked I wonder whether it will be possible to get past the "hey, it's just a game" factor - in order to test out legal systems later to be implemented in the real world.
Personally I don't see why it is necessary to get past this, as gamers would be sure to test all realistic and fantastic loopholes in a system, as that is what gamers do. But if it was supposed to be used as a model of society, I fear that simulations will fall in the same trap as laboratories: human beings are aware of the context they are in, aware that they are part of a test, an experiment or a simulation, and they act accordingly. The information will always have to be interpreted and analysed with an eye to errors due to this self-awareness of human subjects.
And while you're looking in at Terra Nova, read Greg Lastowka's post on Games, intellektual property and ownership.