8.1-8.3

8.1

What Is Intelligence? Intelligence is notoriously difficult to define, but this has not kept people from trying. Part of the difficulty is that intelligence can reasonably be described at three levels of analysis: as one thing, as a few things, or as many things (Carroll, 1993). Intelligence as a Single Trait Some researchers view intelligence as a single trait that influences all aspects of cognitive functioning. Supporting this idea is the fact that performance on all intellectual tasks is positively correlated: children who do well on one task usually do well on others (Geary, 2005). These positive correlations occur even amongst dissimilar intellectual tasks—for example, remembering lists of numbers, knowing word meanings, and folding pieces of paper to reproduce printed designs. Such omnipresent positive correlations have led to the hypothesis that each of us possesses a certain amount of g, or general intelligence, and that g influences our ability to think and learn on all intellectual tasks (Carroll, 2005; Spearman, 1927). Numerous findings attest to the usefulness of viewing intelligence as a single trait. Measures of g, such as overall scores on intelligence tests, correlate positively with school grades and achievement test performance (Gottfredson, 2011). At the level of cognitive and brain mechanisms, g correlates with information-processing speed (Coyle et al., 2011; Deary, 2012), speed of neural transmission (Vernon et al., 2000), and brain volume (Pietschnig et al., 2015). Measures of g also correlate strongly with people’s general information about the world (Lubinski & Humphreys, 1997). Such evidence supports the view of intelligence as a single trait that involves the ability to think and learn. Intelligence as a Few Basic Abilities There are also good arguments for viewing intelligence as more than a single general trait. The simplest such view holds that there are two types of intelligence: fluid intelligence and crystallized intelligence (Geary, 2005): Fluid intelligence involves the ability to think on the spot—for example, by drawing inferences and understanding relations between concepts that have not been encountered previously. It is usually measured by reasoning on unfamiliar tasks, such as answering riddles, choosing the number that would come next in a sequence, being presented a matrix with eight entries filled in and choosing the best item to complete the matrix, or recognizing the figure that is a rotated version of another figure. It is closely related to adaptation to novel tasks, speed of information processing, working-memory functioning, and ability to control attention (Geary, 2005). Crystallized intelligence is factual knowledge, such as knowledge of word meanings, state capitals, answers to arithmetic problems, and identifying a country from a shaded figure on a map. It reflects long-term memory for prior experiences and is closely related to verbal ability. The distinction between fluid and crystallized intelligence is supported by the fact that tests of each type of intelligence correlate more highly with tests of the same type than they do with tests of the other type (Horn & McArdle, 2007). Thus, a child’s performance relative to peers on one test of fluid intelligence tends to more closely resemble the child’s performance on other tests of fluid intelligence than it does the child’s relative performance on tests of crystallized intelligence. Similarly, children’s relative performance on different tests of crystallized intelligence tend to be more similar than their performance on one test of crystallized and one test of fluid intelligence. The two types of intelligence also have different developmental courses. Crystallized intelligence increases steadily from early in life to old age, whereas fluid intelligence peaks around age 20 and slowly declines thereafter (Salthouse, 2009). The brain areas most active in the two types of intelligence differ too: the prefrontal cortex usually is highly active on measures of fluid intelligence, whereas areas within the temporal cortex, such as the hippocampus, are most active on measures of crystallized intelligence (Goriounova & Mansvelder, 2019). A somewhat more differentiated view of intelligence proposes that human intellect is composed of seven primary mental abilities: word fluency, verbal meaning, reasoning, spatial visualization, numbering, rote memory, and perceptual speed (Thurstone, 1938). The key evidence for the usefulness of dividing intelligence into these seven abilities is similar to that for the distinction between fluid and crystallized intelligence. Scores on various tests of a single ability tend to correlate more strongly with one another than do scores on tests of different abilities. For example, although both spatial visualization and perceptual speed are measures of fluid intelligence, children tend to perform more similarly on two tests of spatial visualization than they do on a test of spatial visualization and a test of perceptual speed. The trade-off between these two views of intelligence is between the simplicity of the crystallized/fluid distinction and the greater precision of the seven primary mental abilities. Intelligence as Numerous Cognitive Processes A third view envisions intelligence as comprising numerous, distinct processes. Information-processing analyses of how people solve intelligence test items and how they perform everyday intellectual tasks such as reading, writing, and arithmetic reveal that a great many processes are involved. These processes include remembering, perceiving, attending, comprehending, encoding, associating, recognizing, recalling, generalizing, planning, reasoning, forming concepts, solving problems, generating and applying strategies, and so on. Viewing intelligence as “many processes” allows yet more precise specification of the mechanisms involved in intelligent behaviour than do approaches that view it as “a single trait” or as “several abilities.” A Proposed Resolution How can these competing perspectives on intelligence be reconciled? After studying intelligence for more than half a century, John B. Carroll (1993, 2005) proposed a grand synthesis: the three-stratum theory of intelligence (Figure 8.1). At the top of the hierarchy is g; in the middle are several moderately general abilities (which include fluid and crystallized intelligence and other competencies similar to Thurstone’s seven primary mental abilities); at the bottom are many specific processes. g influences all moderately general abilities, and both g and the moderately general abilities influence the specific processes. For instance, knowing someone’s g allows for fairly reliable prediction of the person’s general memory skills; knowing both g and general memory skills allows quite reliable prediction of the person’s memory span; and knowing all three allows very accurate prediction of the person’s memory span for a particular type of material, such as words, letters, or numbers. Thus, for the question “Is intelligence a single trait, a few abilities, or many processes?” the correct answer seems to be “All of the above.”

8.2

Measuring Intelligence Although intelligence is usually viewed as an invisible capacity to think and learn, any measure of it must be based on observable behaviour. Thus, when we say that a person is intelligent, we mean that the person acts in intelligent ways. One of Binet’s profound insights was that the best way to measure intelligence is by observing people’s actions on tasks that require varied types of intelligence: problem solving, memory, language comprehension, spatial reasoning, and so on. Modern intelligence tests continue to sample these and other aspects of intelligence. Intelligence testing is highly controversial. Critics such as Ceci (1996) and Sternberg (2008) argue that measuring a quality as complex and multifaceted as intelligence requires assessing a much broader range of abilities than are assessed by current intelligence tests; that current intelligence tests are culturally biased; and that reducing a person’s intelligence to a number (the IQ score) is simplistic and ethically questionable. In contrast, advocates (e.g., Gottfredson, 1997; Horn & McArdle, 2007) argue that intelligence tests are better than any alternative method for predicting important outcomes such as school grades, achievement test scores, and occupational success; that they are valuable for making decisions such as which children should be given special education; and that alternative methods for making educational decisions, such as evaluations by teachers or psychologists or essays, are subject to worse bias. Knowing the facts about intelligence tests and understanding the issues surrounding their use is crucial to generating informed opinions about these issues. The Contents of Intelligence Tests Intelligence is manifested in different abilities at different ages. For example, language ability is not a part of intelligence at 4 months of age because infants this young neither produce nor understand words, but it is obviously a vital part of intelligence at 4 years of age. The items on tests developed to measure intelligence at different ages reflect these changing contributions. For instance, on the Stanford–Binet intelligence test (a descendant of the original Binet–Simon test), 2-year-olds are asked to identify the objects depicted in line drawings (a test of object recognition), to find an object that they earlier saw someone hide (a test of learning and memory), and to place each of three objects in a hole of the proper shape (a test of perceptual skill and motor coordination). The version of the Stanford–Binet presented to 10-year-olds asks them to define words (a test of verbal ability), to explain why certain social institutions exist (a test of general information and verbal reasoning), and to count the blocks in a picture in which the existence of some blocks must be inferred (a test of problem solving and spatial reasoning). Intelligence tests have had their greatest success and widest application with children who are at least 5 or 6 years old. The exact abilities examined, and the items used to examine them, vary somewhat from test to test, but there is considerable similarity amongst the leading tests. The most widely used intelligence test for children 6 years and older is the Wechsler Intelligence Scale for Children (WISC). In Canada, the WISC-V-Canadian is used (Cormier, Kennedy, & Aquilina, 2016). This version of the test relies on the same test items as the WISC-IV but is normed on a Canadian sample of children. Thus, it captures the abilities of the current population of children in Canada rather than relying on comparisons to samples of American children. The conception of intelligence underlying the WISC-V is consistent with Carroll’s three-stratum framework, proposing that intelligence includes general ability (g), several moderately general abilities, and a large number of specific processes. The test yields not only an overall score but also separate scores on five moderately general abilities—verbal comprehension, visual-spatial processing, working memory, fluid reasoning, and processing speed. The WISC-V measures these abilities because they reflect skills that are important within information-processing theories, correlate positively with other aspects of intelligence, and are related to important outcomes, notably school grades and later occupational success. Figure 8.2 illustrates several types of items that appear on the WISC-V (the actual items are protected by copyrights and thus cannot be reprinted). FIGURE 8.2 Five abilities tested by WISC-V This figure shows examples of the types of items used on the WISC-V to measure several aspects of children’s intelligence. On most subtests, the measure of performance is simply whether answers are correct, but on some subtests, such as “perceptual speed,” the measure of performance is the number of correct answers that are generated in a limited time. These are not actual items from the test but rather are of the same type.

8.3

IQ Scores as Predictors of Important Outcomes IQ scores are strong predictors of academic, economic, and occupational success for particular populations (Brown, Wai, & Chabris, 2021; Nisbett et al., 2012). They correlate quite strongly with school grades and achievement test scores, both at the time of the test and years later (Ritchie, 2016). Substantial relations between IQ score and performance in intellectually demanding occupations are present not only when the person is hired but for at least 10 years after (Sackett, Borneman, & Connelly, 2008). These positive relations between childhood IQ and subsequent life outcomes continue through at least age 72 (Herd, Carr, & Roan, 2014) and are seen in diverse aspects of life, including occupational and educational success, health, subjective well-being, and quality of friendships (Brown et al., 2021). In part, the positive relation between IQ score and occupational and economic success stems from the fact that standardized test scores serve as gatekeepers, determining which students gain access to the training and credentials required for entry into lucrative professions. Even amongst people who initially have the same job, however, those with higher IQ scores tend to perform better, earn more money, and receive better promotions (Schmidt & Hunter, 2004). A child’s IQ score is more closely related to the child’s later occupational success than is the socioeconomic status (SES) of the child’s family, the school the child attends, or any other variable that has been studied (Ganzach et al., 2013). These relations hold even at the top of the test score distribution. Although popular books such as Outliers (Gladwell, 2008) claim that people with fairly high test scores achieve grades and occupational success equivalent to those of people with very high scores, empirical research indicates that even at the top of the distribution, the higher the test score, the higher that subsequent achievement is likely to be (Lubinski, Benbow, & Kell, 2014). Exceptionally early readers, such as this 4-year-old, often continue to be excellent readers throughout life. Consider a long-term study of 320 children who, by age 13, took the SAT (a standardized test for university admission in the United States) as part of a national talent search and who scored in the top 0.01% (1 in 10,000) in verbal or math ability. Amongst their accomplishments by age 23 were adapting Pink Floyd’s The Wall into a multimedia rock opera, developing one of the most popular video games in the United States, and inventing a navigation system that was used to land a rocket on Mars (Lubinski et al., 2001). As a group, they had published 11 articles in scientific and medical journals and won numerous major awards in areas ranging from physics to creative writing. By age 38, more than half of the original sample had received a PhD, MD, or JD (Kell, Lubinski, & Benbow, 2013). Their rate of PhDs was more than 50 times higher than that for the general population, and their rate of patents was 11 times that in the general population. Even within this elite sample, higher initial SAT mathematics scores predicted higher achievement. For example, the higher the score on the SAT math test at age 13, the greater the number of patents and publications in scholarly journals—especially those in science, engineering, and mathematics—at age 38. You probably have heard of some of the exceptional achievers who met the 1 in 10,000 threshold: Mark Zuckerberg, who founded Facebook; Sergei Brin, who cofounded Google; and award-winning singer-songwriter Lady Gaga (Clynes, 2016). Other Predictors of Success Although IQ scores are strong predictors of academic, economic, and occupational success, they are far from the only influence. Characteristics such as motivation to succeed, conscientiousness, intellectual curiosity, persistence in the face of obstacles, creativity, physical and mental health, and social skills are also very important (Makel et al., 2016; Ritchie, 2016). For instance, self-discipline—the ability to inhibit actions, follow rules, and avoid impulsive reactions—is more predictive of changes in report card grades between grades 5 and 9 than is IQ score, though IQ score is more predictive of changes in achievement test scores over the same period (Duckworth, Quinn, & Tsukayama, 2012). Similarly, “practical intelligence”—skills useful in everyday life but not measured by traditional intelligence tests, such as accurately reading other people’s intentions and motivating others to work effectively as a team—predict occupational success beyond the influence of IQ score (Cianciolo et al., 2006; Sternberg, 2003). As these studies demonstrate, the same dataset can provide evidence for the impact of IQ and of other factors on educational, economic, and occupational success. Consider the data in Figure 8.4. Consistent with the importance of IQ, the figure shows that amongst people with the same level of education, those with higher IQ scores earned higher incomes. Consistent with the importance of other factors, the figure shows that amongst people with comparable IQ scores, those who completed more years of education had higher incomes. Thus, while IQ is a key contributor to educational, occupational, and economic success, motivational and environmental factors are also crucial. Video: Interview with Ellen Winner