Part One
States of Mind
Indicators of Intellectual and Cognitive Decline
1
The Troubling Trend of Cultural IQ
Mark Bauerlein
ANY BROAD DISCUSSION of the American Mind has to take into account an extraordinary finding first noticed in the middle of the twentieth century and confirmed and publicized widely in the 1980s and beyond. From World War I to World War II, cognitive psychologists discovered, the average score of U.S. soldiers on mental tests rose significantly, and in the decades after that, scores on IQ tests continued the trend. It came to be known as the âFlynn Effect,â named for New Zealand social scientist James R. Flynn, whose studies brought to widespread attention the intelligence gains by adults and children in the United States and elsewhere.
IQ tests always set the mean at 100, adjusting the scale accordingly over time so that it appears that test takers from year to year achieve the same cumulative score (a 100 places one at the fiftieth percentile). But if we were to maintain the same scale over the decades, without recalibrating it to keep the mean the same, we would find it rising, as follows for the Weschler Intelligence Scale for Children (WISC):
100.00 (1947â48) 107.63 (1972) 113.00 (1989) 117.63 (2002)
And here are the scores for the Weschler Adult Intelligence Scale (WAIS):
100.00 (1953â54) 107.50 (1978) 111.70 (1995) 115.07 (2006)
An average child in 2002, that is, would have scored seventeen points above average in the late 1940s, while an average adult in 2006 would beat the average in the early 1950s by fifteen points. In other words, an average person in 2002 (the fiftieth percentile) would fall into the eighty-fifth percentile in the 1947â48 group. This advance is remarkable, one that few scientists would have thought possible. As Flynnâs studies circulated in the 1980s and after, psychologists analyzed the data and pondered their causes and implications at length. What made American intelligence climb so rapidly, and will the gains continue? Do the numbers indicate that Americans are, in fact, getting smarterâmore knowledgeable, critical, ingenious, and reflective?
Explanations for the rising scores have varied, with researchers advancing several factors in different proportions, including improvements in early childhood nutrition and education, more people going to college, reduced exposure to toxins (particularly lead), a more demanding cognitive environment in the workplace and in leisure activities (created by technology and media), and the expansion of practices of hypothesizing and abstract reasoning. Flynn favors the latter cause, noting that IQ test questions mostly ask people to detach facts from concrete situations and apply logic to them, identifying patterns, similarities, and analogies, not invoking real-world knowledge. They require subjects to perform certain mental operationsâfor instance, examining numbers in sequence and predicting the next one to come, or observing objects in two pictures and determining relationships between them, then looking at a third picture and selecting a fourth picture (out of five provided) that forms the same relationship to it. Knowledge helps, but mainly knowledge that supports abstract inferences (such as familiarity with classification schemes and mathematical processes), not factual knowledge such as what happened on December 7, 1941. The actual objects in the pictures themselves donât matter, only the abstract relations one can draw. Consider the question, âWhat do dogs and rabbits have in common?â A concrete answer might be, âWell, dogs help you hunt rabbits,â while an abstract one would go, âBoth are mammals.â As Flynn notes in discussing this example in Are We Getting Smarter? Rising IQ in the Twenty-First Century (2012), IQ exams such as Ravenâs Progressive Matrices test for the abstract choice. He terms it a âhabit of mind,â a readiness to treat objects in a conditional, scientific way. âIf you are unaccustomed to using logic for anything but to deal with the concrete world,â he writes, âand indeed distrust reasoning that is not grounded in the concrete, you are not amenable to the change of gears that Ravenâs requiresâ (13). If a test question began, âSuppose two friends start from Point A and reach Point B five minutes apart . . . ,â a hyperconcrete mind might think, âWhy donât they drive the same speed?â while an abstract-habit mind would comfortably accept the opening as a working premise. Rising IQ scores, then, measure a particular kind of cognitive disposition that, in fact, primary and secondary education in the twentieth century had increasingly stressed. The gains are real, and they mean we are getting smarterâif, that is, we define intelligence as this capacity for abstraction.
Indeed, if we take the numbers at face value, we have to believe that American adults and children are much smarter than they were a half century ago, as if evolution accelerated so fast that it concentrated thousands of years of development into several dozen. The average child in 2010 would have been exceptional in 1950, the ordinary twenty-first-century adult able to run circles around a mid-twentieth-century adult whenever an abstract intellectual problem arose. Thatâs what fifteen points representâclear and distinct superiority, a marked gap in mental talent easy to demonstrate (hand both of them a Sudoku puzzle). A score of 117 puts someone in the top 15 percent of the population.
And yet, given this dramatic advance, those of us in the worlds of education and employment are stuck with a discordant fact. Why, we wonder, do so many high school students, college students, and younger workers seem so terribly deficient in basic knowledge and skills? Their reasoning abilities may have jumped forward, their ability to learn may make the 1950s mind appear lethargic, but their reading comprehension hasnât improved at all. One source of long-term data on reading comes from the National Assessment of Educational Progress (NAEPâthe âNationâs Report Cardâ), which shows scores for seventeen-year-olds flat since the test began in the early 1970s (National Center for Education Statistics). We have average scores on the SAT verbal exam going back even farther, and they underwent a shocking decline of fifty-four points from 1962 to 1980. The score has wavered up and down by a few points ever since (Adams). More youths are going to college, yes, but the numbers of them requiring remedial coursework in math and writing keep climbing, the rate of those in need standing at 10 percent for selective schools, 30 percent at less selective colleges, and 60 percent at two-year institutions, according to the National Center for Public Policy and Higher Education. Employers, too, constantly grumble about the lack of readiness of college graduates for their workplaces. A fall 2013 report by Chegg queried one thousand hiring managers on the skills of recent college graduates, and the rate of those who judged them âVery/Completely Preparedâ for the following tasks came up as follows:
âWorking on tasks independentlyâ: 58 percent
âWriting to sum up resultsâ: 51 percent
âSolving problems through experimentationâ: 49 percent
âMaking a persuasive argumentâ: 43 percent
At the same time, Chegg asked college students the same question, and they judged themselves âVery/Completely Preparedâ at much higher rates, for instance, 70 percent on âWriting to sum up resultsâ (Korn). More substandard results come from a 2013 ACT report on the entire adult population, not just recent graduates, that identifies three âcognitive assessmentsâ (âReading for Information,â âApplied Mathematics,â and âLocating Informationâ) that tally more closely the mental aptitudes tested by IQ questions. âLess than half (45%) of the examinees with a high level of education met the Locating Information skill requirements for 4 of the 5 occupations with a large number of openings and for 3 of the 5 highest-paying occupations requiring a high level of education,â the ACT concluded. âOnly 19% of the examinees with a low level of education met or exceeded the Locating Information skill requirement.â Overall, âSignificant foundational skills gaps exist between the skills of examinees with either a low or high level of education and the skills needed for jobs requiring a low or high level of education.â Clearly, the gains reported on IQ tests havenât spread to school and work.
If we drill down into the IQ data, however, the mismatch starts to make sense. The WISC (child) and WAIS (adult) IQ tests have several parts, each one testing a different mental function. The Digit Span subtest has subjects repeat a list of numbers that have been presented orally (testing memory and attention). The Block Design on the childrenâs test asks subjects to organize blocks according to a given model (testing spatial reasoning and motor skills), while the Vocabulary exam for adults tests how many words they know and how well they can deploy them. Other subtests include Visual Puzzles, Similarities, Arithmetic, Coding, Matrix Reasoning, Symbol Search, and Information (a straightforward measure of general knowledge derived from culture). If we look at child scores, we see that the rates of improvement over the decades have varied for each subtest, sometimes sharply, the difference masked when we combine them into an average score. At the high end, we find Similarities climbing 23.85 points from 1950 to 2004, while Coding jumped 18 points and Block Design 15.9 points. At the low end, the Information sub-test showed an improvement of only 2.15 points, while Arithmetic inched up only 2.3 points and Vocabulary results added 4.4 points. The lesser subtest outcomes explain why academics have stalled for U.S. schoolchildren. Flynn sets NAEP reading scores alongside Information and Vocabulary data for the years 1972 to 2002, when general-information knowledge was flat and vocabulary moved minimally. On the NAEP reading exams, elementary and middle-school students showed some improvements, but twelfth-graders none at all, leading Flynn to conclude, âTodayâs children may learn to master basic reading skills at a younger age, but are no better prepared for reading more demanding adult literatureâ (18). The small vocabulary IQ gain is reflected in better NAEP reading scores by nine-year-olds, but it doesnât help seventeen-year-olds whose reading tests contain passages with diction exceeding that gain.
The Information subtest factors into reading scores as well, a connection that laypersons, educators, and even professors of education often fail to appreciate. The significance of general information begins with the fact that texts always have unstated meaningsâcontent that is implicit in the sentences and taken for granted as known by readers. The 2009 NAEP reading assessment for twelfth-graders, for instance, chose as one passage a rental agreement and asked test takers to respond to prompts such as, âThe name of the tenant must be filled in on the rental agreement in two places. Identify the two places and explain why the name of the tenant needs to appear in each of them.â The passage didnât explain what a rental agreement is, define key terms (âtenant,â âdisclosureâ), or introduce the diction and style of legal language. Successful test takers had to know something about property rights and contracts in order to read quickly and answer the questions confidently. Those who didnât stumbled as they read, laboring to discern meanings that others apprehended automatically. Examine any newspaper story, press release, scientific report, contemporary novel, or other commonplace text and you find the same thing: buried knowledge that sets readers on a comprehension continuum, those at the high end having the knowledge that lets them proceed fluently, and those at the low end not having it and faltering. If you donât know anything about rental situations and have never seen a legal contract, critical thinking skills wonât much help. Those abstract mental abilities measured by other subtests, at best, help readers to make imperfect inferences from whatever aspects of the text are familiar. But thatâs a time-consuming and chancy process. Relevant prior knowledge, in contrast, allows readers to recognize implicit references and situations in a snap. The contrast has become clear in eye-tracking studies. Readers with some knowledge about the topic of the text move their eyes swiftly and linearly through the text, whereas low-knowledge readersâ eyes move haltingly, doubling back and slowing down. Their ensuing scores reflect their ease or difficulty.
The more background knowledge one has, then, the more agile and capacious a reader can be, which is why the Information and Vocabulary IQ trends are so important to the state of the American Mind. Other IQ subtest scores give ample reason for optimism, but the disappointing stagnation of general knowledge and vocabulary among the young in the last sixty years means that the benefits of the Flynn Effect may be limited to bounded spheres in which general knowledge and word breadth donât count so much, such as undemanding entertainments and unskilled labor, but not in the civic sphere, traditional cultural settings such as performance halls, or any place related to a marketplace of ideas, not to mention professional and managerial workplaces. Better spatial reasoning may help people read a map and navigate a video game, but it wonât help them choose better schools for their children, catch biblical references in President Obamaâs speeches, decide which mayoral candidate to vote for, or evaluate the trustworthiness of various news programs.
Adults have produced steady IQ gains on knowledge-driven sub-tests, rising 17.8 points in Vocabulary and a not-so-impressive 8.4 points in Information since 1950, the latter likely a result of many more people going to college and picking up general knowledge by enrolling in basic courses in U.S. history, economics, psychology, English, and other core subjects. (The virtually flat Information scores for youths indicate that the boost must happen after age seventeen.) Given the explosion of college attendance in the second half of the twentieth century, not to mention the proliferation of information about the world through the advent of television, we have to judge the adult improvement in Information a disappointment. Set alongside Vocabulary, Coding (16.15 pointsâa typical coding task is to copy a series of numbers and symbols in sequence within a time limit), and Similarities (19.55 points), Information looks like less of a positive development than it does a result casting doubt upon the capacity of abstract cognitive skills to incorporate concrete facts about history, politics, arts, and culture and make people curious about them. In contrast to Arithmetic, though, where adults gained only 3.5 points, Information looks gigantic (Flynn judges the minimal adult gain on Arithmetic âunexpected and shockingâ [22]). Still, if adults and children raised their abstract, fluid intelligence so much, why couldnât they have raised their general knowledge at parallel rates? What would have happened in college if students had entered with a significant knowledge gain already under way?
We can sharpen the question by focusing on Vocabulary and identifying an empirical puzzle. As adults have raised their general knowledge IQ by 8 points, presumably they have retained the knowledge long after they left college, got married, and had children. Greater knowledge certainly affects interests and conversation in the home, and we would expect a more knowledge-rich environment to have influenced the children, producing a parallel rise. But that hasnât happened, as the 6-point difference between the two groups indicates. As for Vocabulary, it should have shown an even firmer correlation between adult and child IQ, because vocabulary learning for small children transpires mainly through the environment that adults provide, including their conversation and reading material. Here, however, we find the most conspicuous gap of all between older and younger Americans. As noted above, from 1950 to 2004, WISC scores for Vocabulary rose 4.4 points, while WA...