First, I should mention something of my own background. My academic background is in psychometrics and also neurobiology, where I did my master's and doctoral work. Sternberg is mostly preaching to the choir with me, as I agree with many of his criticisms about the deficiencies of current and past I.Q. tests.
That having been said, I am mostly okay current psychometric, statistical, and mathematical theory and practice here, as long as one understands the strengths and limitations of the various approaches. I understand those very well, but most people don't and tend to get hung up on one or another aspect of it without having a systematic grasp of all the psychometric issues. If it weren't for that, books like Sternberg's wouldn't be necessary.
Sternberg's definition of successful intelligence is pretty common-sensical, although more difficult to quantify than the abilities typical I.Q. tests measure, but I'm okay with that.
However, the bottom line is that the real answers about intelligence are eventually going to come from the brain research areas, which was my main field. The neurobiology doesn't contradict the psychometric approach but does complement it and provides a more rigorous basis for the idea of intelligence and what it consists of. To give you just a brief example of the neurological picture, the human brain contains 60 trillion nerve cells organized into 14,000 major and minor brain centers and pathways, and each nerve cell is connected to between 3,000 and 100,000 other neurons, producing a neural network and web of almost unbelievable compexity. And in the past 50 years, neuroscientists have made considerable progress in understanding the neural basis of intelligence and of higher cognitive abilities, such as language processing and spatial ability, which have been found to be located in the temporal lobe in the case of language processing, and in the right hemisphere in the case of left-hemispheric dominant people (which is most of us).
But getting back to Sternberg's concerns, the most egregious and widespread problem with I.Q. testing, of course, is that people hung up on a single I.Q. high or low test score, which might not mean much in isolation, and the system doesn't help that situation since it attaches too much credence to them without understanding the other factors, qualifications, and exceptions to a single I.Q. score that must be taken into account.
Sternberg also spends a lot of time discussing examples of people (such as himself), who don't do very well on standard I.Q. tests and about the baleful effect such scores had on their lives. Appropos of that, I can give two much more glaring examples than Sternberg himself, notwithstanding his being a Yale professor, which I am perfectly willing to concede is pretty impressive.
In the late teens and 1920s an important Stanford psychologist, Lewis Terman, tested thousands of California schoolchildren to identify those with high I.Q.'s and then to follow them throughout their lives, to see of the early promise of their intelligence was fulfilled. Terman ended up with a group of 1300 children, who he followed from their early years until their deaths. I would suspect many if not all would be dead by now. Until they had passed away, their files remained sealed, and only Terman and his group knew their actual identities.
Anyway, many did have impressive careers as writers, scientists, lawyers, teachers, and other professionals. Despite most of them growing up during the depression in the last century, many more of them went to college and onto professional and graduate school than the overall population. That having been said, the test had two major faults or oversights in terms of the selection process: the test, which was the Stanford-Binet, an important and widely used I.Q. test, missed the two Nobel Laureates in physics, Luis Alvarez, and William Shockley. Shockley is familiar to many as the famous inventor of the transistor. Both were tested but fell below the minimum of 140 or a score of 135 for a sibling to be included. And none of the other 1300 children won a Nobel Prize. Hence, the test missed the only two Nobel Laureates in the entire group.
Also, James Watson, of Watson and Crick fame, only has an I.Q. of 115, if I remember right, and is the co-discoverer of DNA, for which they shared a Nobel Prize back in the 50's.
So obviously, I.Q isn't the whole story. I have many stories myself of people who had much lower test scores than I on any of the standardized tests, whether I.Q., the SAT, the GRE, or whatever, who did just fine in college and grad school and who often got higher grades than I, and who went on to become more successful in real life too. So as I said, Sternberg is sort of preaching to the choir in my case, and overall, I tend to agree with him that I.Q. should not be the overriding determinant in the selection and educational process that it often is, at least not without taking into consideration other factors such as special aptitudes and talents, creative abilities, grades, work and real world experience, self-discipline and willingness to work hard, and so on.