A Psychometrics Primer
It's actually not too surprising that the results are accurate. Though a lot of people attribute this kind of stuff to psychobabble and crackpot theories, the science behind it is often pretty good.
The way personality exams are generally created (the scientifically worthwhile ones anyway) is by first norming the test against a population. I.e. administer all the test questions (or color cubes in this case) to as wide of a population as you can get your hands on and then record the results. Then ask all these people a bunch of qualitative questions - "What do you want to be when you grow up?", "What makes you happy?", "Do you like to eat bagels?" or whatever.
If you gather enough data, and properly and stringently analyze it, patterns will start to emerge. You'll find that, for hypothetical example, people who choose yellow first tend to like bagels, people who pick red last want adventurous careers, and so on.
There's a misconception that psychologists dream up some theory about personality based on factor X (in this case, color cube preference) and then try to substantiate it. But in actual fact, studies are FIRST done that reveal consistent patterns of behavior and preference. The psychologist THEN induces a theory that might account for these observed patterns.
Of course, there's plenty of flaky pop psychologists who do it in reverse order, write a book about it and become millionaires. But real psychology-type people generally don't approve of that sort of thing.
The ColorGenics test is really just telling you that, given your color preferences, you probably have ideas similar to people with similar color preferences. The scientifically interesting thing about ColorGenics is, why on earth is color preference so strongly correlated with particular patterns of higher order reasoning?
OK, enough psychology for now. You can all wake up now and resume your daily lives.