The next tool for social science experimentation should allow for macro level, generalizable, scientific research. In the past devices such as rat mazes, Petri dishes and supercolliders have been developed when scientists needed new tools to do research. We believe that Virtual Worlds are the modern equivalent to supercolliders for social scientists, and feel they should be the next area to receive significant attention and funding. The advantages provided by virtual worlds research outweigh the costs. Virtual worlds allow for societal level research with no harm to humans, large numbers of experiments and participants, and make long term and panel studies possible. Virtual worlds do have some drawbacks, in that they are expensive and time consuming to build. These obstacles can be overcome, however, by adopting the models of revenue and maintenance practiced by the current game industry. The returns from virtual worlds being used as scientific tools could reach levels that would self fund future research for decades to come. However, at the beginning an investment of funding agencies seems to be necessary.
Keywords: Virtual Worlds, Macro Level Experiments, Research Infrastructure