(FORTUNE Magazine) – Familiar question: What kind of intelligence gets you ahead in the world of business? In Practical Intelligence: Working Smarter in Business and the Professions (Harper & Row, $17.95), Roger Peters offers some answers worth pondering, although I would recommend against buying them outright. Peters, a psychologist at Fort Lewis College in Durango, Colorado, takes his lead from the new and flourishing school of ''cognitive psychology.'' The school emphasizes the idea that intelligence comes in multiple forms: It can be interpersonal (understanding other people), intrapersonal (understanding yourself), verbal, logical-mathematical, physical (possessing such skills as the ability to work with your hands), and the list does not end even there. Cognitive psychologists have performed a sizable number of experiments on these different kinds of intelligence. The value of the experiments is limited, however, by the fact that so many of them are done on college students -- an unrepresentative group, compressed within a relatively narrow range of intelligence. Each of the chapters in Peters's book is devoted to one of the ''multiple intelligences.'' (The phrase is associated with Howard Gardner, a Harvard- based cognitive psychologist and MacArthur Fellow to whom Peters acknowledges a special debt.) The author's approach is didactic: He wants to help his readers understand, as they focus on the different dimensions of practical intelligence (PI), how they can improve their own thinking in each area. A major component of interpersonal intelligence, for example, is an ability to detect the motives and feelings of others. People obviously vary in the keenness of their interpersonal insights, and Peters argues persuasively that many of the skills involved are learnable; specifically, you can learn to pick up on the ''postures, expressions, and tones of voice'' that send signals through what is commonly called ''body language.'' Peters suggests that you keep testing yourself on such matters. At parties and business meetings, for example, make explicit predictions about people you don't know -- predictions about where they will sit, who will talk to whom, who will turn out to be the boss -- and make a point of checking the answers. I give Peters high marks for setting forth his case plainly and with little of the superfluous jargon that beclouds much academic writing on psychology. Unfortunately, he is an unreliable guide with respect to one major matter. He insistently distinguishes between PI, which is more or less synonymous with ''street smarts,'' and ''school smarts,'' which are measured by IQ scores. IQ testing is the heavy in the tale he tells. Peters asserts that PI is more multifaceted than what the IQ measures, that it is more teachable, and that it is more relevant to the demands for intelligence you meet in the business world and practical life in general. Like most cognitive psychologists, Peters severely underestimates the practical significance of IQ measurements. I suspect that one reason for this persistent error is the sampling problem noted above: When the people you test are drawn from a relatively narrow range on the IQ scale, IQ will inevitably seem not to have much explanatory power. But in the world at large, the practical effects of IQ variations are gigantic. No fact about tested intelligence has been as widely confirmed as the ability of the tests to predict outcomes not only for school grades but also for such intensely ''practical'' outcomes as incomes. Peters seems oblivious, for example, to some enormously important recent work on IQ as a guide to job skills. The work has been directed by John E. Hunter of Michigan State University and Frank L. Schmidt, formerly of the U.S. Office of Personnel Management and now at the University of Iowa. Hunter and Schmidt have made what looks like a genuine breakthrough in the understanding of human capital, and it is worth pausing here to zero in on their findings. They have taken advantage of new statistical procedures (called ''meta- analysis'') that make it possible to combine the results of thousands of small-scale studies. The studies in question measure worker productivity in industry, government, and the military while also capturing the individual traits of the workers involved -- including such traits as personality, motivation, educational background, and ''cognitive ability'' (taken from standard IQ tests and their equivalents). The studies indicate that, after correcting for sampling fluctuations and the unreliability of measurement, intelligence as measured by IQ is the single most valid predictor of worker productivity available to industrial psychologists. That remarkable statement holds, whether the productivity is gauged by objective indexes of worker output or by the judgments of supervisors. To be sure, IQ's predictive power is not the same for all jobs. Its validity is greatest for the more complex jobs. It does better, that is, at assessing the likely productivity of a worker who must set up an assembly-line machine than of one who merely moves the machine from one point to another. But even for the simplest jobs, IQ test scores are highly significant predictors of job performance -- a much better predictor than, say, tests measuring educational level or individual motivation. And IQ is even better at predicting performance in on-the-job training programs. These data have some large economic implications, and Hunter and his colleagues have made some efforts to gauge them. Question: If IQ is as good as it seems to be at measuring worker productivity, what is the cost (that is, in lost productivity) of not using IQ-like standard tests on typical populations of job applicants? The Hunter-Schmidt calculations leave little doubt that the cost is huge. They estimated that in 1981, for example, the federal government could have saved something like $13 billion by ignoring all other criteria and simply selecting applicants with the highest test rankings from the available labor pools. For the economy as a whole, they estimated that in 1980 hiring the highest-ranking applicants would have been worth between $80 billion and $160 billion. Books on practical intelligence, including Peters's, can be extremely useful in helping employers to see how they might raise the intelligence of workers, not to mention themselves. The focus of such books tends to be on improving the performance of given individuals, by increasing their self- awareness, self-discipline, and motivation. But one large fact, which happens to have intensely practical applications for employers, tends to get left out in such books. The fact is that the discipline of psychology is much better at selecting the best available workers than it is at improving the performance of workers in place. Better selection is what the much maligned IQ test makes possible. And better selection still looks like the largest unexploited means for enhancing productivity.