Is Math Anxiety Caused by More than Lack of Practice and Skill?

You may have to register first, but you might want to link to the article Understanding Math Anxiety in a recent edition of Teacher Magazine. I would be interested in the perspectives of both math classroom teachers and those outside of the profession. Many teachers I've spoken to feel that anxiety decreases with increased student preparation but from my own personal observations, I suspect there's more to it than that. Do you believe that 'performance anxiety' on standardized math tests is real or a fabrication? If real, what are the implications for the classroom teacher? Simply make our classes more rigorous and our tests harder so that students will be accustomed to performing under more pressure? Or is something else needed...
One of the cognitive psychologists quoted in the article, Robert Siegler, is a professor of cognitive psychology at Carnegie Mellon University. Note below his current participation on the National Math Panel. Here is a quote:

Students feel more anxiety in math partly because they are dealing with so many concepts and procedures that are foreign to them, said Robert S. Siegler, a professor of cognitive psychology at Carnegie Mellon University, in Pittsburgh, who has examined children’s thinking abilities in math and science. Once students realize they do not grasp a math concept, the internal pressure grows.

“Math entails certain conceptual barriers that lead people to read the same passage over and over again and not understand it,” Mr. Siegler said. By contrast, in reading a history lesson, students are likely to recognize vocabulary, themes, and ideas, even if they do not understand all the implications of a particular passage.

“You don’t feel like you totally didn’t understand it, and you’re just floundering,” he said.

Mr. Siegler is one of 17 people serving on the National Mathematics Advisory Panel, a White House-commissioned group charged with identifying effective strategies for improving instruction in the subject. The panel includes a number of cognitive psychologists, along with education researchers, mathematicians, and others.