http://www.theatlantic.com/business/archive/2012/03/do-we-really-need-more-scientists/254109/
But one thing missing from Lowe's arguments is the failure of math education in America. [We can reasonably argue the various ways of learning science and budgetary needs, but math education is strictly budget neutral.] You can toss a simple arithmatic problem at most Americans, and they'll become frozen with panic and fear if they're without calculators. That math-phobia extends to all math domains, and is remarkably broad-based across all classes. [Talking about the non-STEM population here, which is most Americans in all socio-economic classes.]
Why is that? The evisceration of the rote-learning paradigm in the 1970s is to blame. While it was a good corrective against rote memorization, most basic education require quite a bit of memorization to get started. Especially in math, proficiency with arithmatic can only come from repetitive drills over time. Even after New Math has faded, grade school teachers remain reluctant to train arithmatics in the age of calculators. The Kahn Academy and the Singaporean method both work by going back to the basic drills. Without that foundational understanding of numbers, kids cannot master higher-level material.
We don't need to make everyone into mathematicians. But they don't have to be math-phobes, either. In the rest of the world, people may not like math, nor want to be STEM. But at least they can handle numbers without calculators, they're comfortable with numbers. That's something I'd like to see here, too.