In today's (7/13/13) Seattle Times there was a short article about a 9 year old Tacoma boy who was beaten and starved by his mother and grandparents. I'll bet his MSP scores suffered as a result of this maltreatment, but those who look only at test data will never know the underlying cause. All they will see is a student who 'failed to meet standard'. They will look at the boy's teacher and question his or her methods; what could they have done differently to have reached this child? They will not realize that, perhaps, this boy would have scored at the top of his class if he had been fed on the days prior to the test and not thrown off of his 6 foot high porch by his grandfather. There aren't data points large enough to reflect the human factor that underpins each one of them; for better or worse.
Yes, test data provide information, but they do not provide all of the information. It's not even close.