Many people have very strong beliefs about intelligence testing. All too often those beliefs are negative and unrepresentative of intelligence research. For intelligence researchers, it is a bemusing, irritating, and depressing state of affairs.
Steven Pinker, being interviewed at the International Society for Intelligence Research conference in Montreal in 2017, when asked why public understanding of intelligence was so bad, said he could not understand it, but someone should have a look at how the psychology of intelligence was being taught.
Without a further word everyone pointed at Russell Warne, who the day before had presented a study on the intelligence content of all the most popular US psychology textbooks. He had found that the topic accounted for 3.3% of textbook space, and what was in those few pages was not always a fair representation of research findings. In that vein, Russell has now looked at this sorry state of affairs in more detail, and sought to correct the many misconceptions about intelligence.
In the Know. Debunking 35 Myths about Human Intelligence. Russel T. Warne
What are the myths, and does he debunk them? Is he the avenging angel who will protect the righteous from the calumny of false witnesses?
I am listing out the 35 Chapter headings, so that you have a detailed look at what the book covers.
Section 1 The nature of intelligence
1 Intelligence Is Whatever Collection of Tasks a Psychologist Puts on a Test
2 Intelligence Is Too Complex to Summarize with One Number
3 IQ Does Not Correspond to Brain Anatomy or Functioning
4 Intelligence Is a Western Concept that Does Not Apply to Non-Western Cultures
5 There Are Multiple Intelligences in the Human Mind
6 Practical Intelligence Is a Real Ability, Separate from General Intelligence
Section 2 Measuring intelligence
7 Measuring Intelligence Is Difficult
8 The Content of Intelligence Tests Is Trivial and Cannot Measure Intelligence
9 Intelligence Tests Are Imperfect and Cannot Be Used or Trusted
10 Intelligence Tests Are Biased against Diverse Populations
Section 3 Influences on intelligence
11 IQ Only Reflects a Person’s Socioeconomic Status
12 High Heritability for Intelligence Means that Raising IQ Is Impossible
13 Genes Are Not Important for Determining Intelligence
14 Environmentally Driven Changes in IQ mean that Intelligence Is Malleable
15 Social Interventions Can Drastically Raise IQ
16 Brain-Training Programs Can Raise IQ
17 Improvability of IQ Means Intelligence Can Be Equalized
Section 4 Intelligence and education
18 Every Child Is Gifted
19 Effective Schools Can Make Every Child Academically Proficient
20 Non-cognitive Variables Have Powerful Effects on Academic Achievement
21 Admissions Tests Are a Barrier to College for Underrepresented Students
Section 5 Life consequences of intelligence
22 IQ Scores Only Measure How Good Someone Is at Taking Tests
23 Intelligence Is Not Important in the Workplace
24 Intelligence Tests Are Designed to Create or Perpetuate a False Meritocracy
25 Very High Intelligence Is Not More Beneficial than Moderately High Intelligence
26 Emotional Intelligence Is a Real Ability that Is Helpful in Life
Section 6 Demographic group differences
27 Males and Females Have the Same Distribution of IQ Scores
28 Racial/Ethnic Group IQ Differences Are Completely Environmental in Origin
29 Unique Influences Operate on One Group’s Intelligence Test Scores
30 Stereotype Threat Explains Score Gaps among Demographic Groups
Section 7 Societal and ethical issues
31 Controversial or Unpopular Ideas Should Be Held to a Higher Standard of Evidence
32 Past Controversies Taint Modern Research on Intelligence
33 Intelligence Research Leads to Negative Social Policies
34 Intelligence Research Undermines the Fight against Inequality
35 Everyone Is About as Smart as I Am
In the spirit of correcting misapprehensions quickly, here are some snap answers to the first 6 questions:
1 In fact, when the same people are given very different intelligence tests, including tests constructed in the belief that there is no general factor, the general factors extracted from the disparate tests correlate at above the .9 level.
2 Mental tasks correlate with each other, and it is easy to extract a general factor (and also some group factors) so it is not unwarranted to summarize people’s general level of ability with one number.
3 Brain size is weakly related .2 to .4 with intelligence, frontal lobes probably in the higher part of that range. Brighter people have more neurons in their brains, and those neurons are more densely packed together and, perhaps counter-intuitively, have fewer connections branching off each neurone. So, intelligence does have a relation to brain function, but research is at an early stage.
4 If intelligence really varies in character between different cultures, then it should be very difficult to extract the “Western” general factor, yet in 31 countries, and using a wide variety of tests, 94 of the 97 (96.9%) samples produced g either immediately or after a second factor analysis. Moreover, the g factor is about as strong in the non-Western samples as it is in typical Western samples. Most countries find “Western” intelligence tests very useful, once they have been translated and some language and specific knowledge items altered or removed. To cap it all, dogs, rats, mice, donkey and primates show g factors. It looks like an evolutionary adaptation.
5 Everyone seems to want multiple intelligences, particularly educationalists. However, even when researchers attempt to measure these multiple intelligences, the result is a series of correlated variables that produce a general factor, which is exactly what should never occur, according to the theory. Moreover, the proposer of the theory did not think it necessary to make it testable.
6 If practical intelligence could be measured, American Football teams would find it extremely useful. Instead, they use the Wonderlic intelligence test, because it correlates with some of the more complicated playing abilities. The proposer of the theory does not specify what results will prove that practical intelligence differs from general intelligence.
I hope I have given you enough brief answers to the first items to make you want to read the book. Some sections are longer and deeper, serving to explain how intelligence tests are constructed, tested for representativeness and fairness, and evaluated against real-life attainments.
Every score achieved on a test, including academic examinations, medical results, brain scans and astronomical surveys of the universe, contains some measurement error, so the underlying true score is never in perfect focus. However, error terms can be reduced by refining methods, increasing sample sizes and examining individual items carefully. In terms of intelligence tests, the practical implication is that an intelligence test result is correct to within 3 to 4 IQ points either way. Since the usual range is from 70 to 130, the error is roughly 5% to 6% of the range.
Academic attainment tests can certainly pick out “college ready” students at 85-89% correctness levels. Those who demand perfect standards of tests should have a look at the lower reliabilities of other selection methods, the ones traditionally used when testing is prohibited. Stopping testing always means testing by other, less reliable, methods.
It is worth giving a longer summary of what Warne says about the most contentious issues, like test bias (chapter 10), one of the most common misconceptions in psychology text-books. Warne describes all the procedures through which tests must pass before they are fit for use. Warne says that this is a complicated procedure, and gives further references. He does not mention Jensen’s rule of thumb: if a test is biased against a group, that group will do better in real life than the test predicts. Incidentally, it is good to have the Ellis Island procedures spelt out, and to find that the major causes of rejection were not low ability, which between 1892 and 1931 was diagnosed in only 0.02% of immigrants.
In Chapter 28 Warne says that “Racial IQ differences are completely environmental in origin” is a myth. Warne makes a good case for dismissing this “official” mainstream view, and to some extent achieves this in one table which I reproduce below. To begin with it makes for a difficult read, but persist, and it rewards.
Heritability within groups and heritability between groups are algebraically related. This means that one can work out, in the case of the 15 IQ point difference between African Americans and European Americans (d = 1.00) how big the environmental difference would have to be to create a 15-point difference in intelligence.
You can find that out by looking at the first column on the far left, ranging from .00 to .90, these being estimates of the heritability of intelligence. Let us assume the mid-point heritability figure of .50 which is well validated in the literature. Let us now take the extreme but mainstream media view that none of the black/white intelligence difference is due to genetics, but that it must be due to bad environments. Look at the next column and you will see that the black environment must be 1.414 worse than the white environment to account for the 15-point deficit. That is a very big environmental difference.
If the heritability of intelligence is in fact .60 (probably the case in adult life) then the environment would have to be 1.581 worse to account for the difference. Micro-aggressions are not going to do the job. There will have to be some macro, glaring, major, persistent environmental disadvantages, easily visible to the naked eye. The national standards of health, education, and housing will have to be at less than the 6th percentile rank, in order for the environment to drive down African American intelligence to its measured position at the 16th percentile rank. Remember, in these calculations we must use external causal variables, almost like confinement in ghettos or reservations, not measures which could themselves be a consequence of lower intelligence, such as less-well-paid jobs.
Environmentalists have often made the weak but persistent argument that there must be something, of some undefined sort, a Factor X, which causes intellectual damage. The above calculation shows that these factors, whatever they are, will have to amount to a big effect, to account for the observed 15-point gap.
Chapter 29 is a good, balanced, and brief summary of the position on the genetics of racial differences in genetics, and it is worth buying the book for that chapter alone.
In subsequent chapters Warne disposes of hypotheses about the X factor that may be causing African American low scores, and about stereotype threat, which appears to be a casualty of the replication crisis.
The last section looks at societal and ethical issues, such as whether research on racial differences should be banned, or held to higher standards. At this point, I should admit I don’t like “debunking of myths” as a format. “Myths” begs the point, as does “debunking”. I know it is a popular format, and my alternative “35 arguments evaluated” is hardly going to pull in the readers, but I think it is more restrained in its claims, and more in the spirit of discovery.
This section is very good, in that it examines difficult debates carefully. The tone is far more philosophical, and more based on competing values than competing facts. I agree with Warne that studies of group differences should be held to the same standards as all other research, not special additional ones; that the dangers of intelligence research have been over-done; that past political controversies and errors should not be used to condemn and prevent new inquiries; that compulsory eugenic legislation was usually promoted on medical and psychiatric grounds, and less frequently on psychometric ones; that intelligence research need not lead to negative social policies (on the contrary, better policies might be more likely); that society does not have to become more unequal (on the contrary, ability differences might be seen as requiring social adjustments and assistance) and indeed that heritability is higher in uniformly good and equal environments.
In Warne’s capable hands, each mistaken popular assertion is examined empirically and then logically. He gives short shrift to hypotheses which turn out to be unsupported, or formulated so vaguely as to be untestable. His process is entirely polite, but lethal. Proponents are treated kindly, false theories dispatched promptly and humanely. He asserts that our view of the world should not be obscured by proven errors, however much we may want to believe them.
Warne is judicious in his use of examples, and often at his best in his footnotes. Example, foot of page 71:
Some people prefer the term “IQ tests,” but I do not. As Geisinger (2019) stated, the tests do not measure IQ. Rather, they measure intelligence and produce a score called the IQ. A test should be named after the trait it is designed to measure – not the score it produces.
Faced with a wall of confident ignorance about intelligence, most researchers are tempted to shake their heads and go back to their lonely studies. Warne has kept his temper and thought carefully about his readers. He has bound together this disparate collection of misunderstandings (often the consequence of misinformation) into a common theme: hypotheses must be testable; proofs must cover the skills under discussion, and not be anecdotal or cherry-picked; where statistics are used to buttress a theory they must be fairly used and presented; and once some general results have been replicated, then then can be called into play to explain phenomena.
The amusing fact about intelligence is that it is very easy to measure. Indeed, many tests which purport to be tests of something else turn out to be intelligence tests of varying scope and power, but intelligence tests nonetheless. Brain power will out. It is a disgrace that so many canards about our mental abilities have been promulgated and accepted.
This book will go a long way to countering misinformed opinions. If it is widely read it will improve people’s general understanding of mental ability. I hope that happens, particularly for the sake of education, which seems particularly prone to distorted accounts and unsubstantiated claims. Will this book enter the mainstream of teacher training as an essential text? I fear it might not be read, because views have become so entrenched that empiricism on this topic is not welcomed.
This book deserves to be read. If not, at least it will provide intelligence researchers with a readily available text, and when critics refer once again to an unsubstantiated misconception, they can be answered with a chapter number, and a reference to this book.