I can't argue against your local statistics if they work locally for you. But you must understand that that's your local sample. In my experience it doesn't matter at all, and the fact I know it's perfectly possible to prescind from a college degree entirely to be perfectly knowledgeable, competent and productive on a certain field (engineering at least, as per my sample), is enough of a proof (at least for that field), as this is provable by counter-example. I have college education, so I know from first hand experience that the quality of the material available on the internet is vastly superior to anything college can ever give you. In fact I realized after the fact that I wasted my time, I shouldn't have pursued a college degree on engineering, I should have learned from the internet directly.