This postulate is ascribed to Einstein. Many distinguished scholars quote it these days in order to fight the present testing craze.
However, this is a misquote. Einstein actually said that “Not everything which is important can be measured.”
Anyway, we should beware of the pendulum swinging from extreme, thoughtless testimania back to extreme, thoughtless testhostility. I agree that current tests and testing practice are bad, because they hardly, if at all, test what they pretent to test, but measure mainly reading speed, test anxiety, and test wiseness. This is, as Peter Sacks had rightly argued 2000 in his book “Standardized minds”, the true reason for the strong correlation between test scores and economic status! If we do not say this, we unintenionally reinforce the false belief that these tests would be valid. These tests are merely statistical constructs with a grain of this and that ability, but they are not psychometric instruments: There is no psycho in their metric.
We do have excellent, valid tests of important human traits which many do not know. They are psychologically made tests, not statistically fabricatred ones. But their construction costs time and money and they do not provide the test industry with profits as bad tests do. In fact, they are off limits for high-stakes testing of people. They are only made for evaluating teaching methods, lesson plans, and educational policy-making. If we teach all kids well, we do not need to select them and can focus instead on self-selection and cooperation (Deming). Admission to scarce positions would then be made on the basis of a true lottery, not on the basis of a pseudo-scientific lottery.
Addendum: I am preparing a paper on this topic. Where, will be announced on my web-site.