SJBIdaho points us to this article and asks:
"Frennzy, Do you think it is possible to skew the results of this "juke box test" to  have a reverse effect?"

Well, if I understand your question correctly, I would say yes.  Results of a program like this would be very similar to any statistical sampling.  Basically, a percentage of conformity with a set of pre-determined conditions, in a specific set of pre-determined criteria.  In this case, the program would be fed the signatures of current hit music, and would then test new music to determine how well that new music matches those signatures.  That said, all statistical results can be skewed to support just about any argument (unless you live in Florida during an election year). 

Now if you were asking more along the lines of "could we manually alter the input criteria, and would it create a new type of 'hit' song?", again, I would probably say yes...if enough of the public were monitoring the results that come out of the program.  Support for this argument comes from Fox Television, in the form of 'American Idol'.  The presumption there being that anybody good enough to win the contest will automatically be a smash success.  There is a sort of Pygmalion effect here.  Perception becomes reality.  Success in the music industry today has much less to do with the substance of the music than it does with image, promotion, and marketing.  While I don't begrudge anyone the right to use those tools to further their musical career, I firmly believe that they should have some talent prior to using them. 

Finally, if you were asking, "could we do some subtle manipulation of the data, in order to prove that people like Britney Spears actually have no talent, and there is no reason that she should be a star?", then the answer can only be:  I sure hope so.

That's the FAQ, Jack.  Keep 'em coming.