Measuring intelligence: tests, norms and nature-nurture

Transcript

RyanFrance, the early 1900s. The French government brings a researcher named Alfred Binet a very practical question. Which children might have difficulty in school? They want a test to find out.

SoheeAnd from that one request comes more than a century of intelligence testing, with real benefits, real harms, and a few court cases along the way.

RyanWelcome to EPPP Prep, the podcast. I'm Ryan.

SoheeI'm Sohee. Last episode was theories of intelligence. Today is measurement: how tests are built, what the numbers mean, and what they can and can't tell us.

RyanFirst, the term. I Q stands for intelligence quotient, a score earned on a test designed to measure intelligence.

SoheeA bit of history. In the late 1800s, Sir Francis Galton developed the first broad test of intelligence, though he wasn't a psychologist. Then Binet's test, in the early 1900s, for the French government. It included many verbally based tasks.

RyanThen Lewis Terman, a Stanford professor, modified Binet's work. He standardized how the test was given and tested thousands of children of different ages to establish an average score for each age. The result, in 1916, was the Stanford-Binet Intelligence Scale. An updated version is still widely used.

SoheeTwo words to keep apart. Standardization means the administration, scoring and interpretation are consistent. Norming means giving the test to a large population, so groups like age groups can be compared. That produces norms, reference scores for interpreting future scores.

RyanAnd a lovely nuance in the textbook. Norms aren't expectations of what a group should know. They show what that group does know.

SoheeNext, David Wechsler, in 1939. He combined subtests tapping verbal and nonverbal skills, because he believed intelligence was, in his words, the global capacity of a person to act purposefully, to think rationally, and to deal effectively with his environment.

RyanThat became the Wechsler-Bellevue, later renamed the Wechsler Adult Intelligence Scale. The textbook names three Wechsler tests: the adult scale in its fourth edition, the Wechsler Intelligence Scale for Children, fifth edition, and the preschool and primary scale, fourth edition. Those are the editions the textbook lists, and the tests do get revised over time.

SoheeAnd they're periodically renormed. The children's test, the W I S C five, was given to thousands of children across the country, so a child taking it today is compared with same-age peers.

RyanThe W I S C five has fourteen subtests, making up five indices: Verbal Comprehension, Visual Spatial, Fluid Reasoning, Working Memory and Processing Speed. Each child gets a score on each index, plus a Full Scale I Q.

SoheeAnd notice how that reflects the theories from last time. The scoring reflects the view that intelligence is made of multiple abilities across several cognitive areas.

RyanThose periodic renormings revealed something surprising, the Flynn effect, named for James Flynn. Each generation has a significantly higher I Q than the last.

SoheeBut here's the trap. Flynn himself argues that rising scores don't necessarily mean younger generations are more intelligent. So if an answer choice says the Flynn effect proves people are getting smarter, be careful.

RyanNow the numbers. Intelligence test results follow the bell curve, a normal distribution. On most I Q tests the mean is one hundred, and in modern testing one standard deviation is fifteen points.

SoheeScores within one standard deviation of the mean, between eighty-five and one hundred fifteen, are considered average, and sixty-eight percent of the population falls in that range. A score of one hundred thirty or above is considered superior.

RyanPause and think. A child scores seventy on a test with a mean of one hundred and a standard deviation of fifteen. How would you describe that score?

SoheeThirty points below the mean, so two standard deviations below. And a score of one hundred fifteen would be one standard deviation above the mean, still within the average range.

RyanAt the high end, a popular belief once held that very intelligent people were maladjusted. Lewis Terman tested that. In 1921 he began a longitudinal study of more than fifteen hundred children with I Qs over one hundred thirty-five.

SoheeAnd they became well-educated, successful adults who were in fact well-adjusted. They were even above average in physical build and attractiveness, dispelling the idea that highly intelligent people were weaklings.

RyanSo why test at all? The textbook says the value is clearest in educational and clinical settings. Children with learning difficulties or severe behavior problems can be tested to see whether an I Q well below the mean for their age is part of the picture. Without testing, people who need extra support might be missed.

SoheeTesting is also used in courts, and people use the results to seek Social Security disability benefits.

RyanBut testing has a dark history, and the textbook names it. I Q tests have been used to support the eugenics movement. According to the textbook, the Supreme Court case Buck versus Bell legalized forced sterilization of some people deemed feeble-minded through this kind of testing, resulting in about sixty-five thousand sterilizations.

SoheeAnd in 2002, in Atkins versus Virginia, the Supreme Court ruled that executing people with intellectual disabilities is cruel and unusual punishment, prohibited by the Eighth Amendment. It left the definition of intellectual disability to the states.

RyanThen in 2014, in Hall versus Florida, the Court ruled that I Q scores cannot be used as the final determination of a prisoner's eligibility for the death penalty.

SoheeNow, where does intelligence come from? Twin studies are the usual tool. In the Minnesota Study of Twins Reared Apart, identical twins raised together, and identical twins raised apart, had more highly correlated I Q scores than siblings or fraternal twins raised together.

RyanThat reveals a genetic component. But the textbook says aspects of both nature and nurture are probably correct. One study suggests genetics seem to control the level of intelligence, while the environment provides both stability and change in how abilities show up.

SoheeWhich leads to range of reaction: the idea that each person responds to the environment uniquely based on their genes. Your genetic potential is a fixed quantity, but whether you reach it depends on environmental stimulation, especially in childhood.

RyanPause and think. A couple adopts a child with average genetic intellectual potential and raises her in an extremely stimulating environment. What does range of reaction predict?

SoheeThe stimulating environment will likely improve her intellectual outcomes over her life, within the range her genes allow. Genes set the range. Environment decides where in it she lands.

RyanThen group differences, where the textbook is direct. Some ethnic groups score higher on I Q tests than others, and it's likely the results don't have much to do with the quality of each group's intellect. The same is true for socioeconomic status.

SoheeIn 1969, Arthur Jensen argued I Q is determined by genetics and proposed two levels of ability: Level One for rote memorization and Level Two for conceptual and analytical abilities. His claims about racial differences in Level Two drew an intense response, and Robert Williams was among those who called out racial bias in his results.

RyanAnd in 1979, in Larry P. versus Riles, parents of African American students argued California's testing method for identifying learning disabilities was culturally unfair, because the tests were normed and standardized on White children.

SoheeThe method disproportionately classified African American children as mentally retarded, the term used then, and many students were incorrectly classified. The case summary found the tests racially and culturally biased, with a discriminatory impact on Black children.

RyanSo connect bias to norms and validation. The parents argued the tests were normed and standardized on White children, and the case summary says the tests had not been validated for the placements they were used to make.

SoheeOne last distinction. Learning disabilities are not the same as intellectual disabilities. Learning disabilities are specific neurological impairments, not global intellectual or developmental disabilities, and they most often affect children with average to above-average intelligence.

RyanDyslexia, trouble processing letters and sound-letter correspondence, is the most common learning disability in children. Dysgraphia affects writing, and dyscalculia affects learning arithmetic.

SoheeLet's recap. Binet built a test to find children who'd struggle in school. Terman standardized and normed it into the Stanford-Binet. Wechsler built the adult, child and preschool scales. The W I S C five has five indices and a Full Scale I Q.

RyanMean one hundred, standard deviation fifteen, eighty-five to one hundred fifteen is average. The Flynn effect shows rising scores, not necessarily rising intelligence.

SoheeTwins show a genetic component. Range of reaction says genes set limits and environment determines where you land. And biased norms, as in Larry P., can do real harm.

RyanHere's what I hope stays with you. A test score is a snapshot under particular conditions. It's never the whole of a person, and that includes you.

SoheeTry a few practice questions, and we'll see you next time on EPPP Prep.

Loading the read-along timings…