WASHINGTON — If judged solely by their IQ scores, today’s kids are smarter than any generation since testing began — so smart that many of their great-grandparents would have been found mentally deficient by today’s standards.
The gain in IQ scores averages about 3 points per decade. And the increases are beyond debate. Wherever IQ test trends have been studied — in the United States and 23 other industrialized countries, plus Kenya — average scores rise over time.
Experts in intelligence measurement are still debating what’s behind the surge. Among the explanations they offer are:
* A richer intellectual environment. In 1895, 5 percent of Americans graduated from high school, noted psychologist Douglas Detterman, the editor of the scholarly journal Intelligence. A century later, 5 percent of Americans earned advanced degrees. Better-educated home environments, the theory goes, may make modern offspring more adept at answering the kinds of abstract questions that stud IQ tests.
* Smaller families. In 1900, the average American woman had four children. Today, the average is 1.9. That means a higher percentage of high-striving firstborns who probably also are getting more parental attention. This should raise IQs.
* Testing dexterity. The main gains turn out to be in parts of IQ tests that measure visual and abstract thinking rather than the 3 R’s and rote knowledge, in which gains are nil. Today’s kids, analysts say, bring rich experience with visual puzzles, such as mazes, Rubik’s cubes and computer games, as well as more experience in solving abstract problems such as programming computers and cell phones.
* Genes. The changes have come too fast for genes alone to be a big factor. Moreover, high-IQ families have fewer children than lower-IQ families. But a small difference in gene-based intelligence these days can be magnified, experts think, by doting parents, demanding preschools and other boosters of a child’s environment.
James Flynn, the New Zealand researcher who discovered the global upward creep in IQ scores, thinks that all those factors help to explain the rise. But he doubts that the average American kid today is a near genius compared with the World War I recruits who took the first U.S. IQ tests and earned scores averaging 28 points lower.
Flynn thinks that Americans in the World War I era were “less literate and numerate” than Americans today and less apt as abstract thinkers. But it wasn’t that they were dumber, Flynn said in an interview. Rather, it was that their intellects developed only to the degree that the times demanded.
John Laurence Miller, a psychology professor at New York University who specializes in learning and intelligence, agreed. “Kids are better at the things that matter today,” he said. Their scores are up because their skills are closer to the skills that IQ tests measure.
Flynn, a U.S.-trained political scientist who teaches at the University of Otego in Dunedin, New Zealand, discovered the so-called “Flynn effect” in 1984 when he reported that almost everyone who took them both had done better on the 1949 version of the Wechsler IQ test than the 1974 version. This tendency proved true for every IQ test, including nonverbal ones.
Publishers of the principal U.S. IQ tests, the Wechsler series and the Stanford-Binet, know that as time passes, test-takers get a few more questions right and the average score inches up. So every decade or two they revise some questions to make the tests harder and test the new version on thousands of kids, typically aged 10 or so. The purpose is to craft tests on which the median score is back down to 100. That’s an average IQ by definition.
Even with their upward creep, IQ scores remain what they’ve always been: reasonably accurate instruments for predicting success in school, college and work for most people. And the rise in scores can make a big difference when IQ cutoffs are invoked as a standard or part of an admission process, according to Stephen Ceci, a professor of developmental psychology at Cornell University in Ithaca, N.Y.
When American IQ tests aged without revision between 1976 and 1989, and in effect got easier, Ceci found that the number of children placed in special-education classes fell from 960,000 to 520,000 nationwide. Enrollment in such classes has followed a saw-toothed trend line since, he added, rising sharply with new tests and falling gradually as they age.
Ceci and others find that for borderline applicants, rising IQ averages also sometimes skew:
* Numbers and placements in gifted classes.
* Eligibility for Social Security based on mental disability.
* Military recruiting and assignments.
* Appeals of defendants and death-row inmates whose claims for clemency depend on measured IQs below 70, the standard for mental deficiency.
For researchers who attribute the increasing scores largely to rising U.S. education levels, there’s one problem: The biggest IQ gains over the years turn out to be on questions that measure pure intelligence, unaffected by schooling. (Example: How are sunrise and sunset similar?) And the smallest gains turn out to be in areas that schooling boosts. (Example: In what country are the Pyramids found?)
Flynn, after studying IQ test scores and achievement test results, found a solution to that paradox. He concluded that modern American kids, while relatively skilled in abstract thinking, lack a lot of the basic three R’s and rote knowledge that their forbears possessed. At age 10, when IQ tests are generally given, stronger abstract-thinking skills give them an edge. But by the time today’s students exit high school, their achievement levels in reading and math — and rote knowledge such as the location of the Pyramids — are much closer to their great-grandparents’, Flynn said.
A Q&A on IQ tests
WASHINGTON — Many school districts don’t immediately buy the latest versions of IQ tests when they’re published. That’s because they’re expensive, the old tests seem to work just fine and their school psychologists are used to administering and scoring the old ones.
That can work out to your child’s advantage — or disadvantage, according to Stephen Ceci, a professor of developmental psychology at Cornell University who’s studied the consequences of using old IQ tests. He explains why in the questions and answers that follow.
Q. Why does the test’s age matter?
A. New tests are always harder than older tests. For every year after a test is updated, a child’s IQ score is inflated by three-tenths of a point. If the test is 10 years old, the inflation is about 3 points. If it’s 20 years old, it’s 6 points.
Q. What’s wrong with that?
A. An older test can prevent kids who need special educational services from qualifying for them if it lifts their scores above the cutoff for eligibility for special education classes. A new test can make IQ scores appear to plummet by comparison with prior years’ results. It can even influence admission into gifted and talented programs.
Q. How so?
A. Suppose your child was eligible for a gifted/talented program some years ago when she was last tested but recently was retested and told she’d scored a little too low. Perhaps her performance was unchanged but her score seemed to drop because a new test was used for the second administration. Whenever a new test is used, there’s a significant drop in IQ score, often 5 or 6 points. In fact, the same child taking two versions of the test on the same day will generally score 5 or 6 points higher on the older version.
Q. How might an old test be disadvantageous in a special-education placement?
A. Suppose a child has serious learning problems and would benefit from small classes and intensive instruction. Often, those classes are available only to students whose IQ scores are below 70. If a test is old, a child with a true IQ of, say 67, may appear to have an IQ of, say, 73.
Q. What can a parent do?
A. Ask which test battery was used or will be used. Ask what the “norming date” is, meaning the year when the test was last adjusted to account for changes in average IQ. If it’s not the latest version of the test, ask the administrator why. Then ask whether they’ll be adjusting the scores in any way.
Q. How do I know what the latest norming dates are?
A. For the Wechsler Intelligence Scale for Children, the latest test is the WISC-IV, released in 2003. The WISC-III came out in 1991. For the Stanford-Binet Fifth Edition the date is 2003. For the Stanford-Binet Fourth, it’s 1986.