I would say that equal distribution may be valid on a global scale, but in a microcosm therein there can be clustering. For instance, my home town exists to support the employees of the neighboring nuclear research facility. It's a small town, about 4200 people, and the vast majority of those people are employees of AECL, the children of employees or retired employees. If you believe that certain types are naturally predisposed to certain areas of study, then you have to agree that the logical conclusion is that an isolated region built to foster one specialized and technical facility would have an uneven distribution of types. That doesn't mean that it exludes certain types entirely, but one would expect a huge influx of logicals (which seems to hold true).