I'm not linking it, but Time's cover article is about education in the 21st century and it's need to adapt. Lots of interesting points, and some things that I, and many educators already knew. I've got a personal feeling to shed on the language issue, but will do that in another post. My big point here is simple: no child left behind (NCLB) is crap.
I'm pretty sure I've mentioned something like this before, but here we go again...
Testing is the centerpiece of NCLB and it is causing more problems than it is fixing. Testing sounds like a good idea to people who are not educators, because they tend to think, "Gosh, when I was in school we took tests to 'prove' what/that we had learned, so testing kids will 'prove' that they are learning and that schools are succeeding." Feel free to think this way all you want, no matter how stupid it is, and it is stupid. Let's take a look at a simple benchmark of learning and determine how to test it: children should know how to multiply and divide. Well, we've all taken math tests before, most of us had to memorize multiplication tables up through 12. If I ask you what is 12 x 11? Does your answer represent understanding? a real ability to multiply? memorization? Okay, so let's make it 324 x 17... This is more challenging, so it will better gauge a student's abilities. But there are different approaches to doing this problem. One is to write the numbers on top of each other and multiply row by column then sum. This is how most people learned. Another approach may be to make use of the distributive property (324x17 = 10x324 + 7x300 + 7x25 -7). The latter demonstrates a significantly better understanding of mathematics and how multiplication relates to addition and subtraction, but it hasn't been tested. The test may have another question related to the distributive property and another dealing with the order of operations, but someone who just uses them automatically has a better understanding of how they all work. It isn't tested, though. You can have children show their work, but then you have to make a qualitative decision as to the better answer even if two kids get the answer correct. Moreover, anyone who does the work in their head may not show any, and how is that to be marked? When these kids get out into the real world would it be better if they know how to multiply 324x17 or would it be best if they can figure it as being ~20x325 = 20x300 + 5x4x25 = 6500. What if they have $70 and need to pick up 17 (items) that cost 3.24 (and they live in OR, where there is no sales tax)? Of course, real world problems suck and there really is no legit comparison. The real answer is that the student who can apply different aspects of their learning in one area will be better able to do so in others and will be better able to extrapolate what they have learned to real problems and, therefore, be better able to function in a dynamic society.
Testing does not (and will never) demonstrate this. The fundamental flaw of (standardized) testing is that it is goal oriented, with the wrong goal. This is not to say that all testing (or even all standardized testing) is bad. Nonstandardized testing, as practiced in many classes, is very valuable if not necessary to gauge student progress, and can, in fact, examine their understanding at a higher level. So long as the test is not a benchmark for passing and failure for a student or school, standardized test results can be useful to model simple aspects of education, and disparity within it. The SECOND they determine a student's and/or a school's fate they become meaningless measures of mediocrity...methinks. Our education system already does a poor job of nurturing creativity and intellectual curiosity, NCLB testing makes it worse.