Despite some popular The General assertions, a single factor for intelligence, called g, can be measured with IQ Intelligence tests and does predict Factor success in life by Linda S. Gottfredson No subject in psychology has pro- downplayed or ignored. This misrepresen- mental tests are often designed to mea- voked more intense public controversy tation reflects a clash between a deeply sure specific domains of cognition—ver- than the study of human intelligence. felt ideal and a stubborn reality. The ideal, bal fluency, say, or mathematical skill, From its beginning, research on how implicit in many popular critiques of spatial visualization or memory—people and why people differ in overall mental intelligence research, is that all people are who do well on one kind of test tend to ability has fallen prey to political and born equally able and that social inequali- do well on the others, and people who social agendas that obscure or distort ty results only from the exercise of unjust do poorly generally do so across the even the most well-established scientific privilege. The reality is that Mother board. This overlap, or intercorrelation, findings. Journalists, too, often present a Nature is no egalitarian. People are in fact suggests that all such tests measure view of intelligence research that is unequal in intellectual potential—and some global element of intellectual abil- exactly the opposite of what most intel- they are born that way, just as they are ity as well as specific cognitive skills. In ligence experts believe. For these and born with different potentials for height, recent decades, psychologists have other reasons, public understanding of physical attractiveness, artistic flair, ath- devoted much effort to isolating that intelligence falls far short of public con- letic prowess and other traits. Although general factor, which is abbreviated g, cern about it. The IQ experts discussing subsequent experience shapes this poten- from the other aspects of cognitive abili- their work in the public arena can feel tial, no amount of social engineering can ty gauged in mental tests. as though they have fallen down the make individuals with widely divergent The statistical extraction of g is per- rabbit hole into Alice’s Wonderland. mental aptitudes into intellectual equals. formed by a technique called factor The debate over intelligence and Of course, there are many kinds of analysis. Introduced at the turn of the intelligence testing focuses on the ques- talent, many kinds of mental ability and century by British psychologist Charles tion of whether it is useful or meaning- many other aspects of personality and Spearman, factor analysis determines the ful to evaluate people according to a character that influence a person’s minimum number of underlying dimen- single major dimension of cognitive chances of happiness and success. The sions necessary to explain a pattern of competence. Is there indeed a general functional importance of general mental correlations among measurements. A mental ability we commonly call “intel- ability in everyday life, however, means general factor suffusing all tests is not, ligence,” and is it important in the prac- that without onerous restrictions on as is sometimes argued, a necessary out- tical affairs of life? The answer, based on individual liberty, differences in mental come of factor analysis. No general factor decades of intelligence research, is an competence are likely to result in social has been found in the analysis of per- unequivocal yes. No matter their form inequality. This gulf between equal sonality tests, for example; instead the or content, tests of mental skills invari- opportunity and equal outcomes is per- method usually yields at least five dimen- ably point to the existence of a global haps what pains Americans most about sions (neuroticism, extraversion, consci- factor that permeates all aspects of cog- the subject of intelligence. The public entiousness, agreeableness and openness nition. And this factor seems to have intuitively knows what is at stake: when to ideas), each relating to different sub- considerable influence on a person’s asked to rank personal qualities in order sets of tests. But, as Spearman observed, practical quality of life. Intelligence as of desirability, people put intelligence a general factor does emerge from analy- measured by IQ tests is the single most second only to good health. But with a sis of mental ability tests, and leading effective predictor known of individual more realistic approach to the intellectual psychologists, such as Arthur R. Jensen of performance at school and on the job. It differences between people, society could the University of California at Berkeley also predicts many other aspects of well- better accommodate these differences and John B. Carroll of the University of being, including a person’s chances of and minimize the inequalities they create. North Carolina at Chapel Hill, have con- divorcing, dropping out of high school, firmed his findings in the decades since. being unemployed or having illegitimate Extracting g Partly because of this research, most intel- children. ligence experts now use g as the working By now the vast majority of intelli- Early in the century-old study of definition of intelligence. gence researchers take these findings for intelligence, researchers discovered that The general factor explains most granted. Yet in the press and in public all tests of mental ability ranked individ- differences among individuals in perfor- debate, the facts are typically dismissed, uals in about the same way. Although mance on diverse mental tests. This is 24 Scientific American Presents Human Intelligence Copyright 1998 Scientific American, Inc. Y AR T LIBR BRIDGEMAN AR Ad Parnassum, by Paul Klee HIERARCHICAL MODEL of intelligence is akin to a pyramid, with g at the apex; true regardless of what specific ability a intelligence researchers can statistically other aptitudes are arrayed at successively test is meant to assess, regardless of the separate the g component of IQ. The abil- lower levels according to their specificity. test’s manifest content (whether words, ity to isolate g has revolutionized research numbers or figures) and regardless of the on general intelligence, because it has way the test is administered (in written allowed investigators to show that the or oral form, to an individual or to a predictive value of mental tests derives stitutes intelligence in action. Indeed, group). Tests of specific mental abilities almost entirely from this global factor intelligence can best be described as the do measure those abilities, but they all rather than from the more specific apti- ability to deal with cognitive complexity. reflect g to varying degrees as well. Hence, tudes measured by intelligence tests. This description coincides well with the g factor can be extracted from scores In addition to quantifying individual lay perceptions of intelligence. The g fac- on any diverse battery of tests. differences, tests of mental abilities have tor is especially important in just the Conversely, because every mental also offered insight into the meaning of kind of behaviors that people usually test is “contaminated” by the effects of intelligence in everyday life. Some tests associate with “smarts”: reasoning, prob- specific mental skills, no single test mea- and test items are known to correlate bet- lem solving, abstract thinking, quick sures only g. Even the scores from IQ ter with g than others do. In these items learning. And whereas g itself describes tests—which usually combine about a the “active ingredient” that demands the mental aptitude rather than accumulated dozen subtests of specific cognitive exercise of g seems to be complexity. knowledge, a person’s store of knowledge skills—contain some “impurities” that More complex tasks require more mental tends to correspond with his or her g reflect those narrower skills. For most manipulation, and this manipulation of level, probably because that accumulation purposes, these impurities make no prac- information—discerning similarities and represents a previous adeptness in learn- tical difference, and g and IQ can be used inconsistencies, drawing inferences, ing and in understanding new informa- interchangeably. But if they need to, grasping new concepts and so on—con- tion. The g factor is also the one attribute The General Intelligence Factor Exploring Intelligence 25 Copyright 1998 Scientific American, Inc. Matrix Reasoning 1. 2. A B C D E A B C D E Number Series Analogies 3. 2, 4, 6, 8, _, _ 7. brother: sister father: 4. 3,6,3,6, _,_ A. child B. mother C. cousin D. friend 5. 1,5,4,2,6,5, _, _ 8. joke: humor law: TFREDSON A. lawyer B. mercy C. courts D. justice T 6. 2,4,3,9,4,16, _,_ GO LINDA S. Answers: 1. A; 2. D; 3. 10, 12; 4. 3, 6; 5. 3, 7; 6. 5, 25; 7. B; 8. D 8. B; 7. 25; 5, 6. 7; 3, 5. 6; 3, 4. 12; 10, 3. D; 2. A; 1. Answers: SAMPLE IQ ITEMS resembling those on current tests require in the images, numbers or words. Because they can vary in the test taker to fill in the empty spaces based on the pattern complexity, such tasks are useful in assessing g level. that best distinguishes among persons pendent of g (or each other). Further- by-product of one’s opportunities to considered gifted, average or retarded. more, it is not clear to what extent learn skills and information valued in a Several decades of factor-analytic Gardner’s intelligences tap personality particular cultural context. True, the research on mental tests have confirmed a traits or motor skills rather than mental concept of intelligence and the way in hierarchical model of mental abilities. aptitudes. which individuals are ranked according The evidence, summarized most effec- Other forms of intelligence have to this criterion could be social artifacts.