Jensen's critical thoughts on IQ-tests:

>If there's anything on which my judgment has changed significantly since 1969, it is the scientific value of typical IQ tests. Psychological tests are limited by the fact that they do not provide absolute scales, that is, those that have a true zero point and equal intervals throughout their range. As is well known in the physical sciences, the mathematical, and not just statistical, analysis of data is much greater with measurements based on absolute, or ratio-property scales. These are virtually absent in the psychological measurement. There's no doubt, however, that IQ tests and many other conventional psychological tests have real practical value. They are unquestionably valid predictors of certain kinds of performance in education and employment, and can be most useful in education selection, and in hiring and promotion decision.<

Intelligence, Race and Genetics - Conversations with Arthur R. Jensen, Frank Miele  (2002)