According to this article a recent study has shown that math anxiety is a very real phenomenon with measurable physiological reactions.
However, the article goes on to prescribe the cure: "focus on the process of learning math instead of simply trying to get students to churn out the right answer... A classroom culture where students aren't afraid to fail and are encouraged to learn by talking through wrong answers is optimal."
As someone that's actually in the classroom, these sorts of answers just piss me off. It's not that there isn't a lot to be gained when students are willing to fiddle with an equation for a bit, or look over incorrect answers to learn to spot mistakes and learn from them. There absolutely is.
The problem comes in that students aren't willing to. All too often, students feel they've spent far too much time on a problem already, and having to look it over a second time is asking far too much. They're done. Time to move on. I've always offered students the option to take their homework, identify mistakes, correct them, and resubmit it for a good portion of the lost credit back. But even with this generous incentive, student's don't participate.
As far as process based learning, this too is a fantastic idea... but again overlooks the reality of the situation. Memorizing a set of steps to solve a quadratic equation by completing the square isn't nearly as good as teaching a student why completing the square works and having them work out the process behind the steps. But the reality is that this is a far deeper level of learning. Memorization of a few steps is the foundational level of Bloom's Taxonomy. Applying them is only the third. But a full understanding such that a student could fully understand the methodology is near the very top. And unsurprisingly, that deeper learning takes more explaining, more examples, and more time, which is often at a premium in the classroom.
And even if I had that time, students again, don't want it. It doesn't feel solid enough for them. As a recent example, I concentrated heavily on the concepts of solving quadratic equations in my algebra II class. Explaining, the zero product property with factoring, and why when you complete the square, you need to take half of the coefficient on the linear term and square it, going through the derivation of the quadratic formula. The steps were, of course, all there, but it got so lost with all the rest, the student failed the test with a 40%.
I let her have another week to study and again, concentrated, initially heavily on the theory behind the machinations, but again, she made it clear she was utterly confused. So on the last day, I had her write out a list of the steps for each method, use them as a guide on a few practice problems, and suddenly, she was immensely confident. On her retake she scored a 96%.
So what's the lesson here? Distilling math into rote memorization of simple processes lead to significantly decreased anxiety and subsequently, higher test scores. Of course, I fully realize that if tested on these methods in a year, this knowledge will be nearly completely gone. It's not an ideal situation, but given the alternative, frustration and failure, I'll take the battles I can win and hope that even if the process isn't fully remembered, it's easier to learn a second time around and crystallization comes then when the knowledge and experience base has been expanded.
It's not the best solution, but it's reality.